Alpha Centauri 2

Sid Meier's Alpha Centauri & Alien Crossfire => Modding => Bug/Patch Discussion => Topic started by: kyrub on January 10, 2013, 02:42:53 PM

Title: SE choices for AI - suggestions
Post by: kyrub on January 10, 2013, 02:42:53 PM
This thread is created for help in an area that is severely under developped in the original game.


What should change for different factions?
What  is a typical SE mistake for AI?
How and when should they change to specific SE combinations?
What to do with AIs and the bloody Free Market option?

I'll be happy to hear your concrete suggestions.
Title: Re: SE choices for AI - suggestions
Post by: Kilkakon on January 10, 2013, 03:38:26 PM
The only thing I'll add from the outset--I would suggest to make it based on the specific stats that each SE choice gives, rather than just the slot, as many big mods and balance mods change what each individual thing does.
Title: Re: SE choices for AI - suggestions
Post by: Kirov on January 11, 2013, 02:58:10 PM
I haven’t followed AI habits for a long time, but I’ll try to recall as much as I can and I have looked into my single player save games.

First and foremost, I need to say that I find SE unbalanced and I’m biased towards some. I’m most happy when I can run Demo/FM/Wealth and shun war-oriented SE even when at war.

On the Politics line – I find Demo very good, Fundie poor, PS sucky

The Economics are most balanced, where even experienced players can have different approaches. FM is very good, Green and Planned are OK. I don’t really use Planned anymore, but others could disagree.

Values – Wealth is the best SE choice, as it’s very useful and comes with very low penalty, Knowledge is so-so, Power is heinously, hilariously bad.

All in all, PS, Fundie and Power suck. Now, I’m not saying I’d never switch to them, but I’d never make it a general rule (“use Power at wars” is very bad advice in my opinion). If I was to give a general tip, it would go “avoid those SE settings unless your very precise calculations say otherwise”.

With that in mind, my thoughts on the AI SE:

First a special case: the Hive. He’s the simplest. As I learnt from good Hive players, you simply switch to PS/Planned/Wealth and stick to it until the end of time. I see no reason why AI shouldn’t do so 100% times. Hive AI does in fact use PS and Planned, but prefers Power to Wealth, which brings me to:

1st sin: they don’t use Wealth; even those who don’t hate it and can pick it, still avoid it at all costs. Instead, they love them some Power. I have a savegame where everybody along with Morgan and Dee have Power, and another game where Aki runs Power even without any war!

- in my opinion; everybody who can, should run Wealth; if they can’t, run Knowledge. Seriously, Power is bad for you

- Pirates need Wealth very desperately, but there is a problem here: they can pick it, but they hate Wealthers.

- Santiago should be simply immune or half-immune to penalties of Power. Barring that, I’d still limit her using Power only to the most heated wars


2nd sin: no SE at all, even with the proper tech. This takes place in several situations. One was mentioned already: Miriam has IA and could pick Wealth, but no way, ‘she’s a warrior not a builder and ain’t be usin’ no goddam Wealth.’ So she runs nothing until the advent of AMA. I think this is related to their ‘agendas’ which you set in the scenario. If it’s default ‘conquer’, not ‘build’, Miriam/Hive will avoid Wealth no matter what, even if it’s good for their conquest.

Another situation is that they’re bound on their ‘preferences’. This is a bit silly because for example Miriam can’t run Knowledge, but will hate your guts for running PS or Demo. Of course, a Miriam human player can use both to his great advantage, so AI is very stupid here. I believe that AI is sometimes too much stuck to their preferences. I understand this as a part of flavour and roleplay, but still… A human Morgan can run Green, let’s allow it to AI.

However, this raises the problem when they use an SE they “hate” you for. I think Miriam should run Demo sometimes, Pirates run Wealth always, Morgan – Green sometimes, but what if they hate you for doing the same at the same time? Well, if it possible, either we should switch off their stupid rant when they run the same SE as you, or simply live with this absurd – I’d rather Pirates on Wealth who hate me for Wealth than Pirates who run Power and are completely useless.

As a general rule, it’s usually better to run some SE than not run anything through the midgame.

3rd sin: Using SE the AI can’t handle. This is particularly true for FM, as the AI can have 20 pacifism drones and still not see any problem here. AI should know to run 20% PSYCH under FM. Not that it helps much with p-drones. I think it should check if and how many wars it has. Other such examples are energy or industry-killing SE. So it’s 2200, Miriam runs Fundie/nothing/nothing and she has like five techs under her belt… Sad. In a savegame of mine I see Cha Dawn running the same, this is pathetic. Also Power, PS and Planned can kill the AI’s energy or mineral output.

One important note: I accept the AI running Power+Fundie if it can check that the Spoils of war are on and it’s possible to steal tech this way. Otherwise, AI simply must have some moderate energy income.

I don’t know how the underlying code looks like, but I was thinking if AI SE choices might be related to other game elements in certain ways:

1) positive triggers – this could work both ways. Either “switch to FM when you have a pactmate” or “look for a pactmate when you have FM”. Seriously, I think AI is too aggressive, also towards each other, and I don’t see any pattern in AI warmongering; they don’t fight for land or ideals, they just fight and make peace almost a random (how easy you can convince AI to call of vendetta for free and then you read that they signed treaty; really guys, that’s all it takes?)
Other sample triggers (I can pay more thought to that if needed): “use early Planned to churn out colony pods” “switch to Fundie if you lost 2 bases to mind control” “switch to Green if your total ecodamage averages 20 per base”,

2) negative triggers – this is about the costs. “give up Fundie and switch to Demo if you have like 3.2 lab points per turn” (AI can seriously have this much in MY 2200). “give up FM if you have on average more than 2 p-drones per base”, but also “give up Demo/FM/Wealth if you’re losing a war”

3) I wonder if the AI can check other players, AI and humans, to decided what to do. “if you’re the last one with tech, consider Demo/FM/Wealth” “if you’re the last one with tech, consider Fundie and hordes of probes” “consider Knowledge if there’s not so many probe teams on the map”.

Also, some other general notes I took down:

- the pacifists and builders should run Demo most of the time, warriors - sometimes

- as I said, sometimes they shouldn’t run their preferences (Miriam, Santiago)

- early Planned would be good for AI, but must combine this with building colony pods and formers (and it should be set that AI starts every game with researching CE). Then builders can go FM, while warriors can stay Planned, but only if it doesn’t kill their tech (even warriors need some basic conquer techs)

- Santiago can go early FM and hunt for bugs (she still has better odds), but her units shouldn’t leave her territory.

- Fundie might be interesting in AI’s hands, but it takes smarts to successfully probe; maybe AI could probe other AI, humans aren't easy prey


This is my general overview. I’ll try to play some SP to find more food for thought, but am a bit busy these days, so it may not happen in January.
Title: Re: SE choices for AI - suggestions
Post by: kyrub on January 11, 2013, 08:54:33 PM
Great post, Kirov, interesting suggestions. It made me fire up IDA... only to remind myself how complex the damned SMAC code is. I don't understand my own comments. Incredible.

What I understand right from beginning:
a) I can free the AIs from using only their preferred setting.
b) I can free the AIs from not using their hated setting.
c) I have also something I have called "AI_max_pref_socio_var_unused". Any idea what it could be (these things most probably coincide with some faction setting in alpha)? Maybe that skipped Miriam option or what?

Maybe just freeing the AI from its own limits would give us some interesting fruits. (Modifying alpha afterwards to make Green undesirable for Morgan, so that his diplomatic agenda fits, is probably not hard.)


-------


General idea:
How the system works - it seems the program runs over all socio variants and gives them a rating. It ommits. At the end it probably chooses the best rated one. Really not sure, it's a bit wild guess.


other observations (take with a large handful of salt, I understand about 66% of the ultra complicated code here)

a) ECONOMY and EFFICENCY are considered together. If the AI player is not anti aggressive (-1) and not (some being_attacked flag) or interested in wealth (+1) or tech (+1), their basing factor is (times) 1. Any of above cases double/triple the effect. Interest in tech has double effect itself. (and other factors enter in afterwards, like energy increase times number of bases etc.). There is further 200% boost for factions that
have neither growth nor power setting +1. No extra FM considerations here, I am afraid. 

b) Support POLICE- very bad understading, since the program uses some general player variable (nr of supported units??? or drone supported units???) I cannot decipher. Let's name it X. X seems to be negative factor. If support = -4, X is trippeed, if the player in war flag = TRUE, and support is -3, X is doubled. If war = TRUE and support = -1 or -2 AND the FM is on, X is 150%. (It seems FM is considered only if -1 or -2 support, strange!!!). If aggressive_inclination +1, then X = 200%, if growth_incli, then X =150%, power_incli, 150%, wealth_incli, 50%, tech_incli, 75%. Basically, Morgan is reluctant to take support problems in consideration. Hmmm. This is very rough, it could really get some improvement, although expereimenting is necessary.

c) MORALE... complicated. Factors are number of bases, some attacking units variable, (being_at_war_flag+1), (aggressive_incli+2). This is multiplied together and by morale setting (so larger map with more bases and enemies and more developped empires give you geomertically more ++ morale settings for AIs). Here is the "power love", I think. It's divided by 8, then.
Further movement: If AI_aggressivity >1, 200% (seems like bug, since no AIs have >1?). Power_incli +1 and Growth and Wealth_incli = 0 gives you 200%. Anti agressive, not power_incli, tech_inclined and not_in_war - all in the same time gives you 50%.

... to be continued

Title: Re: SE choices for AI - suggestions
Post by: Kilkakon on January 12, 2013, 12:06:13 AM
Remember that the player can't choose the hated SE setting either, so letting the AI do that would be cheating!

Will wait to see what else you add. :D
Title: Re: SE choices for AI - suggestions
Post by: Lord Avalon on January 12, 2013, 12:36:01 AM
Just a thought about AIs running an SE they hate: it's "do as I say, not as I do."

AI:  You greedy [progeny of unmarried parents], you're running Wealth, I despise you.

Player:  But you're running Wealth, too.

AI:  Shut up!  We're talking about you.  I hate you.
Title: Re: SE choices for AI - suggestions
Post by: kyrub on January 12, 2013, 01:29:13 AM
Incredible.
They royally messed the whole thing, beyond imagination.

They swapped POLICE and SUPPORT setting.
- So, the AI first checks its support rating instead of the police one, it uses it to count how much drones are induced in its cities (the actual drone count may be quite different, but AI does not notice).
- Then it considers it checks its police rating instead of the support one -  and again, it uses the police rating to count the number of minerals spent on support of military units.

This may explain why it does not work with Free Market: the AI does not notice that the bad police rating is destroying its bases. Also AI bases are being ran into ground by costly supports, especially if AI fails to notice, having nice ++POLICE, that support is apalling. Maybe that explains why Hive AI does so well: Police state improves both police and support rating and it covers the problem.
Title: Re: SE choices for AI - suggestions
Post by: Kilkakon on January 12, 2013, 04:46:41 AM
That might just the fix we need then! :O

Will you and Yitzi work together on one combined patch?
Title: Re: SE choices for AI - suggestions
Post by: Green1 on January 12, 2013, 04:58:22 AM
There also seems to be some kind of bug related to Free Market and spore lanchers. When an AI like Morgan or Aki run free market and a spore launcher is bombarding outside a city, the AI just sits building units instead of going out to kill the thing when it has units that could easily do so. A few games, including one of GrimthR's LPs (youtube) using your patch has pointed that out. It cripples an entire AI faction. I have noticed it myself.

EDIT: By the way... welcome to AC2. We are HUGE fans of your work here as well as Scient's. Your patch is more popular than you think. Many AC folks are lurkers, but AC does have a decent fan base. We are not bad people to boot.
Title: Re: SE choices for AI - suggestions
Post by: kyrub on January 12, 2013, 11:57:22 AM
Will you and Yitzi work together on one combined patch?
I have no time to cooperate or coordinate or whatever. But it should be very easy to combine our efforts in one patch, if Yitzi does not mind. I'm fine and as far as he lets me do my job (AI).

On the other hand, I still doubt (with past experience) everybody will want to "be forced" to play with AI changes. The changes have almost always some minor side effects, it's hard to avoid it in a game with such a complex AI code as SMAC. And I cannot make AI changes optional. So I think the patch will end with having two variants, (Yitzi + bugfixes) and (Yitzi + bugfixes + AI).

@Green
In this thread, we tackle just the SE settings. The problem you mentioined is AI units movement (and there are many, problems, I mean). So I 'd leave it elsewhere. Don't let your hopes raise, though, the movement function is a headache, in itself. Very hard to correct anything, I think I spent 3 days to locate the colony pod decision and I still could not force the game to do exactly what I wanted.
Title: Re: SE choices for AI - suggestions
Post by: Kirov on January 12, 2013, 03:51:05 PM
c) I have also something I have called "AI_max_pref_socio_var_unused". Any idea what it could be (these things most probably coincide with some faction setting in alpha)? Maybe that skipped Miriam option or what?

a) ECONOMY and EFFICENCY are considered together. If the AI player is not anti aggressive (-1) and not (some being_attacked flag) or interested in wealth (+1) or tech (+1), their basing factor is (times) 1. Any of above cases double/triple the effect. Interest in tech has double effect itself. (and other factors enter in afterwards, like energy increase times number of bases etc.). There is further 200% boost for factions that
have neither growth nor power setting +1. No extra FM considerations here, I am afraid. 

Thanks for your work kyrub, although I must say I barely follow you here (Ogg be no computer whiz). It took me a longer moment to figure out what you meant. I even downloaded an IDA tool and fired up terranx.exe, but let's not kid ourselves, I don't even know where to look. I have a question here - can you modify the strings or just the values? Can you flat out remove some consideration from the above list? I understand that you can't shoehorn the FM consideration here?

This SUPPORT and POLICE mix-up is a big thing, obviously. But there is something I don't get - so AI running FM 'thinks' it has -5 SUPPORT? What does it do with such a revelation? I still see it cranking units like there's no tomorrow.

I think that support stuff needs separate consideration. Like for example "don't clog your base minerals with more than (10-20%) maintenance unless" and 'unless' means here either 'you're winning a war' (took +2 bases), so it's a good idea to push on, or 'you're losing a war' (lost one base) so you must defend.

I'm quite busy this month, but for a longer while I wanted to start a thread on AI behaviour, especially in terms of diplomacy. I've always wondered if there is a slider like the one you see in Civ4, when you get 'points' for behaviour. I want to check the comparability of things, like for example 'running your hated SE is cancelled out by trading two techs'. I had this impression that sometimes it's better to refuse an outright demand from the AI and then gift them the very same tech. I could never prove it, though. Is it possible to see it with this tool you use?

And again, wow, this SUPPORT/POLICE thing is a crazy mistake. How could that fly under their radar?
Title: Re: SE choices for AI - suggestions
Post by: Kirov on January 12, 2013, 04:07:05 PM
Remember that the player can't choose the hated SE setting either, so letting the AI do that would be cheating!

I'm afraid there is a difference between "SE aversion" (you can't run it) and "SE hatred" (AI scolds you for it and will never use it itself). This is particularly conspicuous for Zak and Miriam, as their aversions and hatreds are not even in the same SE line. Either human or AI Zakharov can't run Fundie, but the AI Zak will additionaly never run Power and Wealth, something which is available to humans (and Zak should switch to Wealth in my opinion in the first possible moment). The same goes for Miriam, who won't use Knowledge (inherently), but also PS and Demo, and she could use Demo early on at least to grab basic conquest techs like NLM, Doc:Flex or maybe even SFF and D:AP.

Others are less obvious, but still - the AI will avoid 2 or 3 SE choices where the human has only one option off. This drives me crazy particularly for Wealth, which is 'averted' by Santiago and Cha Dawn, but also 'hated' by Zak and Pirates.

I'd like to remove this thing that Zak runs Wealth and hates you for running Wealth, because I find it absurd, but it's still better than what we have right now (especially the Pirates could use them some Wealth).
Title: Re: SE choices for AI - suggestions
Post by: kyrub on January 12, 2013, 08:52:13 PM
I have a question here - can you modify the strings or just the values? Can you flat out remove some consideration from the above list? I understand that you can't shoehorn the FM consideration here?
I can modify whatever I want as long as I find some space. People who are able to create DLL injections have no problems with space, sadly it's not my case - I am not a programmer, just a self learned  matemathician-amateur.

I can remove, move or replace a lot of AI SE choice. Don't worry. I cannot inflate the code with massive new tasks, basically, I have to work with the variables the AI already uses. Example: we have Nr_of_my_land_units_on_continent_X variable. We have Nr_of_my_naval_units as well. But we don't have Nr_of_my_air_units, so we cannot check that against running Free market - pity.

Quote
This SUPPORT and POLICE mix-up is a big thing, obviously. But there is something I don't get - so AI running FM 'thinks' it has -5 SUPPORT? What does it do with such a revelation? I still see it cranking units like there's no tomorrow.
Looking at the social agendas' table, perhaps the AI tries to offset -5 SUPPORT with POWER setting. It could be a factor.

Quote
I think that support stuff needs separate consideration. Like for example "don't clog your base minerals with more than (10-20%) maintenance unless" and 'unless' means here either 'you're winning a war' (took +2 bases), so it's a good idea to push on, or 'you're losing a war' (lost one base) so you must defend.
The game has it already, when choosing stuff to build in a base, the surplus is taken in consideration! The game AI has a lot of things inside, in fact. It is really complex, most complex AI from all games I have ever seen, I suppose it's a heritage from civ2 game. Just not working very well. (That's not a bad thing, since it's easier for me to make existing structures smoother, than creating AI code from the scratch.) 
Quote
I'm quite busy this month, but for a longer while I wanted to start a thread on AI behaviour, especially in terms of diplomacy. I've always wondered if there is a slider like the one you see in Civ4, when you get 'points' for behaviour.

I have not looked into diplomacy but I remember that all the main functions were identified. We can defeinitely look into AI diplomacy in not-very-near-future.

Quote
And again, wow, this SUPPORT/POLICE thing is a crazy mistake. How could that fly under their radar?
I actually think that there may be bigger problem in the Morale decision. The way it is counted, +Morale becomes a holy grail, just as you noticed in your analysis. The AI spawns many units, the more unit it spawns, the more continent it settles, the bigger the desire for +Morale. On bigger map, it's worse. I believe we should tone it down.

Quote
I'm afraid there is a difference between "SE aversion" (you can't run it) and "SE hatred" (AI scolds you for it and will never use it itself). This is particularly conspicuous for Zak and Miriam, as their aversions and hatreds are not even in the same SE line. Either human or AI Zakharov can't run Fundie, but the AI Zak will additionaly never run Power and Wealth, something which is available to humans.
I can switch off this behaviour easily, if you wish. If I'm right, it works like this: Zak has knowledge set as his preferred agenda, so he programmatically avoid the other options. That's in the code.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on January 13, 2013, 02:27:40 AM
Will you and Yitzi work together on one combined patch?
I have no time to cooperate or coordinate or whatever. But it should be very easy to combine our efforts in one patch, if Yitzi does not mind. I'm fine and as far as he lets me do my job (AI).

I'm fine with you doing AI; how should we coordinate so that one patch ends up with both sets of changes?  Generally actually changing the code is the last step for me (and I suspect for you), so we might just be able to have one person hold off any changes while the other is at that last stage, and then they can download that and work from there.

Also, in terms of AI changes, maybe it would be better just to list them, and if any seem likely to be controversial I can show you how to set a flag or set of flags based on alphax.  (If 0 is the current method, then it will be completely save-game compatible with my current patch.)
Title: Re: SE choices for AI - suggestions
Post by: kyrub on January 13, 2013, 02:56:38 AM
so we might just be able to have one person hold off any changes while the other is at that last stage, and then they can download that and work from there.
The nature of AI changes requires different approach, I am afraid. I need to create a lot of test builds, to see how AI reacts to different approaches. Afterwards, I reverse some features, I take others in. At the end of the process, a few months on, we may have a better game. Or not, if I fail. In between, the changes may seem even detrimental to the cause. I learned to understand that is normal with AI development.

So I guess you do your work, and I come late and join in. Then people here can agree on any options and flags etc.
Title: Re: SE choices for AI - suggestions
Post by: kyrub on January 13, 2013, 02:59:57 AM
Since I am going away for a few weeks, I cooked a quick test build for SMAC AI (requires SMAC only for now, sorry). SMAC_444_l version in the Downloads.
The highlight is the correction of Support <=> Police mismatch in AI social engineering. You'll find the rest in the short notes to the patch.


It seems the patch is unavailable for download. Anybody?  :-\
Title: Re: SE choices for AI - suggestions
Post by: Buster's Uncle on January 13, 2013, 03:14:34 AM
I just approved the file. 

Nothing notifies us when a file's waiting for approval, so always a good thing to mention it in case I haven't checked Downloads since...
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on January 13, 2013, 04:01:53 AM
Since I am going away for a few weeks, I cooked a quick test build for SMAC AI (requires SMAC only for now, sorry). SMAC_444_l version in the Downloads.
The highlight is the correction of Support <=> Police mismatch in AI social engineering. You'll find the rest in the short notes to the patch.

Thanks.  When (if) you extend it to a SMAX patch, please build it off my latest version at the time, so our patches can be combined.

By the way: One of the things mentioned in your AI patch was beelining; will that make the AI play badly if various dependencies/tech-bonuses/etc. are moved around?  (i.e. will it make the AI beeline for things even if you change the tech tree so that there is no longer an advantage to such beelining?)  If so, that might need changing.
Title: Re: SE choices for AI - suggestions
Post by: Kilkakon on January 13, 2013, 05:16:28 AM
I will admit that I would prefer if AIs stuck to their agenda, as otherwise their personalties tend to go a bit awry and in come cases they miss out on good synergy (in my game for example, most religious factions are immune to the penalties of their own faith, so makes sense for them to go with what they are good at)
Title: Re: SE choices for AI - suggestions
Post by: Kirov on January 13, 2013, 09:38:53 PM
I can modify whatever I want as long as I find some space. People who are able to create DLL injections have no problems with space, sadly it's not my case - I am not a programmer, just a self learned  matemathician-amateur.

I can remove, move or replace a lot of AI SE choice. Don't worry. I cannot inflate the code with massive new tasks, basically, I have to work with the variables the AI already uses. Example: we have Nr_of_my_land_units_on_continent_X variable. We have Nr_of_my_naval_units as well. But we don't have Nr_of_my_air_units, so we cannot check that against running Free market - pity.

Is there a variable "number of pacifist drones in my bases"? Or just "number of drones in my bases"? This could be useful.

I have another question: it is known that in SMAC as in all other games I know, the AI "sees" the map. It doesn't need to explore, it knows if you have weak military without any espionage, etc. Is it possible for the AI to "forecast" things? It could be simple, like for example compare the total combat value of its and enemy's units to estimate chances. Or checking that 90% of improvements and all its bases are so high in elevation that it doesn't have to worry about sea-rising and is free to do eco-damage. Or will gain more than others under Global Trade Pact, so will vote YEA. Stuff like that could really affect SE choices.

If Nr of enemy probe teams on continent is more than X --> switch to Fundie
If Global Trade Pact yields more than X --> switch to FM.

In other words, it could be not related only to the general situation ("I'm a builder, I use Demo"), but also to more specific conditions ("sea rising takes place, I use Green").

Another thing: we're talking right now about what affects SE choices. What about stuff which results from SE? Are there conditions like "I use FM, so I rushbuy drone-related facs" "I use FM, so want to trade", "I use Green, I want to build Dream Twister"?

Yet another question: Are there conditions linked to specific techs? "I focus (Demo/FM/Wealth) on teching to D:AP, then I switch to my beloved Fundie and beat the living hell out of everyone".
There are several techs which the AI needs to research regardless how aggressive it is, before it kills its lab output with PS/Fundie/building just Laser Squads forever and ever.

Quote
The game has it already, when choosing stuff to build in a base, the surplus is taken in consideration! The game AI has a lot of things inside, in fact. It is really complex, most complex AI from all games I have ever seen, I suppose it's a heritage from civ2 game. Just not working very well. (That's not a bad thing, since it's easier for me to make existing structures smoother, than creating AI code from the scratch.) 

Who would think it's so complex! I wouldn't, for one.

Quote
I actually think that there may be bigger problem in the Morale decision. The way it is counted, +Morale becomes a holy grail, just as you noticed in your analysis. The AI spawns many units, the more unit it spawns, the more continent it settles, the bigger the desire for +Morale. On bigger map, it's worse. I believe we should tone it down.

Yes, focus on MORALE can go, it is neglectable compared to industrial output.

Quote
I can switch off this behaviour easily, if you wish. If I'm right, it works like this: Zak has knowledge set as his preferred agenda, so he programmatically avoid the other options. That's in the code.

Yes, I'm strongly in favour of this solution. No need to handicap AI in the way humans aren't. Besides, they still can't run their 'opposite' choices, I think it's in their respective faction files.
Title: Re: SE choices for AI - suggestions
Post by: Kirov on January 13, 2013, 09:44:33 PM
Thanks.  When (if) you extend it to a SMAX patch, please build it off my latest version at the time, so our patches can be combined.

By the way: One of the things mentioned in your AI patch was beelining; will that make the AI play badly if various dependencies/tech-bonuses/etc. are moved around?  (i.e. will it make the AI beeline for things even if you change the tech tree so that there is no longer an advantage to such beelining?)  If so, that might need changing.

Please remember also to have a copy of kyrub's changes but without modifications to the game rules, tech tree, ecodamage, etc. MP players are quite conservative and although they're happy with bug fixes and AI improvements, they're are very reluctant to changes which force a shift in strategies.
Title: Re: SE choices for AI - suggestions
Post by: Lord Avalon on January 13, 2013, 09:49:03 PM
Thanks.  When (if) you extend it to a SMAX patch, please build it off my latest version at the time, so our patches can be combined.

Hey, kyrub, Yitzi, thanks for your efforts.

Just a small suggestion:  when you release a new version, could you please go into properties\summary of your exe and add a comment on what version it is?  Thanks.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on January 13, 2013, 10:57:17 PM
Please remember also to have a copy of kyrub's changes but without modifications to the game rules, tech tree, ecodamage, etc. MP players are quite conservative and although they're happy with bug fixes and AI improvements, they're are very reluctant to changes which force a shift in strategies.

All my changes except the clear bugfixes are set up so that you can use the default rules if you want*.

That said, I do want to make a mod that will force a shift in strategies, but that's going to be purely with alphax changes, so nobody has to play it.  But all my mods until "Alien Crossfire 1.5" will use the same tech tree, rules, etc.

*Well, actually this is not technically true; for instance, for purposes of convenience I plan to remove the "10Xdifference in chassis cost" term in the upgrade cost, but only because I can't think of any circumstance where it would actually be relevant.

Just a small suggestion:  when you release a new version, could you please go into properties\summary of your exe and add a comment on what version it is?  Thanks.

Each of my versions will contain a readme listing all the changes.
Title: Re: SE choices for AI - suggestions
Post by: Kirov on January 14, 2013, 12:04:09 AM
All my changes except the clear bugfixes are set up so that you can use the default rules if you want*.

Oh, I see, that's cool. I thought that you're planning some really massive rework of eco-damage, etc. Is it possible to make that optional, just through the alphax? I thought you're changing the code.

Quote
*Well, actually this is not technically true; for instance, for purposes of convenience I plan to remove the "10Xdifference in chassis cost" term in the upgrade cost, but only because I can't think of any circumstance where it would actually be relevant.

What do you mean here? I don't know the formulas off the top of my head. Is it a big deal? It is a valid strategy to build 'shell' units (trained 1-1-chassis) and upgrade from there, some players will not want to part with it (or pay more)... :)
Title: Re: SE choices for AI - suggestions
Post by: Lord Avalon on January 14, 2013, 12:31:35 AM
Just a small suggestion:  when you release a new version, could you please go into properties\summary of your exe and add a comment on what version it is?  Thanks.

Each of my versions will contain a readme listing all the changes.
I didn't mean list all the changes - just add "Yitzi patch v2" (or whatever it is) to the terranx.exe file.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on January 14, 2013, 12:38:30 AM
Oh, I see, that's cool. I thought that you're planning some really massive rework of eco-damage, etc. Is it possible to make that optional, just through the alphax? I thought you're changing the code.

I changed the code (that part is done) to make it optional just through alphax.  i.e. before my patch it's not possible at all, with my patch it is possible but not mandatory (or you can do all sorts of other fun stuff.)

Quote
What do you mean here? I don't know the formulas off the top of my head. Is it a big deal?

It will make upgrades that change the chassis slightly cheaper.  As far as I know, there is no way to get such upgrades anyway.

Quote
It is a valid strategy to build 'shell' units (trained 1-1-chassis) and upgrade from there

I'd disagree, and consider that strategy to be horribly imbalancing (as it throws off the mineral/energy balance), but my patch will (as usual) make it possible to change the rules to scuttle that strategy, but anyone who wants to play with the current rules can.

I didn't mean list all the changes - just add "Yitzi patch v2" (or whatever it is) to the terranx.exe file.

Add it how?
Title: Re: SE choices for AI - suggestions
Post by: Lord Avalon on January 14, 2013, 12:43:56 AM
Quote
Add it how?
Right click on the terran(x).exe file, go to Properties, go to the Summary tab, add comment.
Title: Re: SE choices for AI - suggestions
Post by: Kirov on January 14, 2013, 01:59:55 AM
I'd disagree, and consider that strategy to be horribly imbalancing (as it throws off the mineral/energy balance), but my patch will (as usual) make it possible to change the rules to scuttle that strategy, but anyone who wants to play with the current rules can.

Yes, I agree with you, I've never liked this particular strategy, although I would use it every now and then. It's simply not the way the game is meant to be played. However, each such change would make a few MP players drop out and insist on playing the vanilla version. Also, it'd be hard to replace it with a new rule that doesn't have weird side effects. If you cut out upgrading 1-1-chassis, people would simply start to make 2-1-chassis types.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on January 14, 2013, 02:18:39 AM
Yes, I agree with you, I've never liked this particular strategy, although I would use it every now and then. It's simply not the way the game is meant to be played. However, each such change would make a few MP players drop out and insist on playing the vanilla version.

I think that by targeting mainly the strategies that most people dislike (ICS, heavy energy focus, and air power will probably be the biggest targets) and those that clearly unbalance the game (e.g. those that make the race for midgame projects into purely a tech race, rather than a combination of techs and production), that should cut down on the dropouts.  And if you essentially get two different games, that's ok too.

Quote
Also, it'd be hard to replace it with a new rule that doesn't have weird side effects. If you cut out upgrading 1-1-chassis, people would simply start to make 2-1-chassis types.

That's why I wouldn't go that route, but instead change the upgrade cost to be more comparable to hurry costs.  That way, anyone who wants could upgrade, but it won't be a powerful strategy most of the time because you get more bang for your buck by just using minerals to build the desired unit normally.
Title: Re: SE choices for AI - suggestions
Post by: Kilkakon on January 14, 2013, 02:37:51 AM
While this is tangenting a little bit, as a person who's made an epic mod that's largely ignored by everyone, I'd advise to keep it to "tweak" levels rather than new things entirely. Unless you get a lot of positive feedback of course. :)
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on January 14, 2013, 03:23:05 AM
While this is tangenting a little bit, as a person who's made an epic mod that's largely ignored by everyone, I'd advise to keep it to "tweak" levels rather than new things entirely. Unless you get a lot of positive feedback of course. :)

I'm definitely planning to keep it to "tweak" levels.  The only really new thing (rather than formula or techtree tweaks, and a few house rules and a new mechanic just to tweak a few problem points) will be a ranking system for multiplayer which is designed to have all sorts of desirable properties (from encouraging attempts at solo victory to allowing handicaps via difficulty level/random faction selection to being able to tolerate a game where luck has a substantial role to play.)
Title: Re: SE choices for AI - suggestions
Post by: Kirov on January 15, 2013, 02:43:49 AM
I think that by targeting mainly the strategies that most people dislike (ICS, heavy energy focus, and air power will probably be the biggest targets) and those that clearly unbalance the game (e.g. those that make the race for midgame projects into purely a tech race, rather than a combination of techs and production), that should cut down on the dropouts.  And if you essentially get two different games, that's ok too.

Well, I think it's more like Kilkakon says. :) The first time I came here, I couldn't even find players for a game with crawlers and choppers banned, and if this thingies are not OP, then truly nothing is. So it's thoughtful to keep things optional. I imagine you'd lose a player or two if you changed as much as the period between council proposals by three turns. People play vanilla, they got used to exploit stuff like crawlers, air, ICS or shell upgrade and I'm afraid not much can be done about it. Hell, even I shed a tear for that bugged Children Creche, because I really relied on it, and I'm a guy who really understands that this bug must go. :)

This became a vicious circle in itself - personally I have nothing against mods, but I don't explore this topic too much for I know I won't find co-players to it anyway.

I think I managed to play once the Smaniac mod with other players, but it was a couple of years ago.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on January 15, 2013, 03:26:27 AM
Well, I think it's more like Kilkakon says. :) The first time I came here, I couldn't even find players for a game with crawlers and choppers banned, and if this thingies are not OP, then truly nothing is.

Admittedly, there are probably better fixes than banning them, and when I finish there almost certainly will be.

Quote
So it's thoughtful to keep things optional. I imagine you'd lose a player or two if you changed as much as the period between council proposals by three turns. People play vanilla, they got used to exploit stuff like crawlers, air, ICS or shell upgrade and I'm afraid not much can be done about it. Hell, even I shed a tear for that bugged Children Creche, because I really relied on it, and I'm a guy who really understands that this bug must go. :)

It actually works the same as it always did if you have negative MORALE; I think it's only the penalty with positive MORALE that Kyrub removed.
But clearly people got used to exploiting those things...but if they want a challenge, maybe they can try single-player Transcend with my planned mod (once it's ready).  As for multiplayer, I see it essentially being two different games, one full of exploits and the other more balanced.

But creating options rather than imposing your way on everyone is often a good idea.  Not always (in particular, anything with a strong ethical component does poorly if it considers itself merely an option), but for something like a game mod, it's generally a good idea (though not worth the effort in certain cases, most notably things like the Nessus Mining Station bug I fixed.)
Title: Re: SE choices for AI - suggestions
Post by: kyrub on January 16, 2013, 09:55:04 PM
Quote
People play vanilla, they got used to exploit stuff like crawlers, air, ICS or shell upgrade and I'm afraid not much can be done about it.
This became a vicious circle in itself - personally I have nothing against mods, but I don't explore this topic too much for I know I won't find co-players to it anyway.
So well said.

By the way: One of the things mentioned in your AI patch was beelining; will that make the AI play badly if various dependencies/tech-bonuses/etc. are moved around?   
Yes, if you change the position of some techs, the AI will end up beelining another tech. The beelining should probably be in alpha.txt, for modding purposes. (But in my modest opinion, modding for SMACX is dead, because nobody tends to play the mods that are created. Or am I wrong?)

   
Quote
I can switch off this behaviour easily, if you wish. If I'm right, it works like this: Zak has knowledge set as his preferred agenda, so he programmatically avoid the other options. That's in the code.

Yes, I'm strongly in favour of this solution. No need to handicap AI in the way humans aren't. Besides, they still can't run their 'opposite' choices, I think it's in their respective faction files.
I think I have an elegant solution here.
What about: If AI likes an option A, than options B or C are not totally excluded  (like now) but the whole SE "position" is worth 25% less. This will make AI
a) switch to the other possibilities when the A option is not yet researched
b) prefer A option to B and C massively
c) still use B or C when highly valuable

@Kirov
When I am saying the AI is complex it does not mean it uses its power well. But it has massive space in the code, which is good for me and potentially for the game improvement. The kind of stuff you ahve mentioned is absolutely normal in the code, BUT the AI tends to run along the five magical numbers in alpha.txt - e.g. aggressivity, power, growth, commerce, eco, instead of "I am running FM, so...". So you may mod it easily, but it's lot more stupid. As a result, the AI is fully moddable, but quite inept.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on January 16, 2013, 11:53:34 PM
Yes, if you change the position of some techs, the AI will end up beelining another tech. The beelining should probably be in alpha.txt, for modding purposes.

In that case:
1. What do you have it beeline?  Some of it might be adjustable to be based on the tech that gives X feature.
2. Would it be possible to write separate AIs for beelining and not-beelining in the same program?  If so, I can show you how to make it based on an alpha.txt variable.  (Note: For SMAC, you will need to reduce some limit (I used the landmark limit) to make room.)
3. Did you include the beelining in the SMAX mod as well, or just SMAC?
4. While we're on the subject, what terraforming did you teach it in the SMAX mod?  In particular, is it going to use boreholes and condensers like crazy?

Quote
But in my modest opinion, modding for SMACX is dead, because nobody tends to play the mods that are created. Or am I wrong?

If a good mod is created, interested players will be found.

   Yes, I'm strongly in favour of this solution. No need to handicap AI in the way humans aren't. Besides, they still can't run their 'opposite' choices, I think it's in their respective faction files.

That said, they still should focus on those somewhat, as they tend to work well with those factions.

Quote
I think I have an elegant solution here.
What about: If AI likes an option A, than options B or C are not totally excluded  (like now) but the whole SE "position" is worth 25% less. This will make AI
a) switch to the other possibilities when the A option is not yet researched
b) prefer A option to B and C massively
c) still use B or C when highly valuable

That sounds very good.
Title: Re: SE choices for AI - suggestions
Post by: Lord Avalon on January 17, 2013, 12:57:28 AM
(But in my modest opinion, modding for SMACX is dead, because nobody tends to play the mods that are created. Or am I wrong?)


Discussion/interest seems to have picked up, and maybe having a SMACentric site is part of that.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on January 17, 2013, 01:05:39 AM
One more thing, Kyrub: When you're almost ready to do (i.e. actually start changing the file, not just finding stuff and designing the code) a SMAX mod, please let me know so I can post my current progress so that you can work from there.
Title: Re: SE choices for AI - suggestions
Post by: Kilkakon on January 17, 2013, 08:15:52 AM
Yes, if you change the position of some techs, the AI will end up beelining another tech. The beelining should probably be in alpha.txt, for modding purposes. (But in my modest opinion, modding for SMACX is dead, because nobody tends to play the mods that are created. Or am I wrong?)
I can testify that modding is dead! At least for big things, it seems. I've spent two years of my life on a game nobody online has played and survived to tell the tale. Woo! I'm probably killing people without realising. XD

But that's not the point of this thread. I'd agree that having them work off the rules rather than hardcoded settings is better as otherwise they'll be really lolligagged/freaked out by some mods (e.g. crawlers are disabled in my game, and some more mods, not sure what would happen if the AI beelined to them).

I think I have an elegant solution here.
What about: If AI likes an option A, than options B or C are not totally excluded  (like now) but the whole SE "position" is worth 25% less. This will make AI
a) switch to the other possibilities when the A option is not yet researched
b) prefer A option to B and C massively
c) still use B or C when highly valuable
Can you make it take immunities to negative stats and civics into account? E.g. Hive's immunity to negative EFFIC, or one of the secret project's immunity to Cybernetic? Or is that automatic in the considerations?



Oh by the way! Some good news. I managed to get both Kyrub's and Yitzi's patches working on my laptop. The catch is that they MUST use DirectDraw=0 using that patch or it won't run. The other catch is that DirectDraw=0 doesn't like my laptop either, as even the basic game with only scient's patch will crash in a few turns after a bit of exploration. :(
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on January 17, 2013, 06:06:04 PM
I can testify that modding is dead! At least for big things, it seems. I've spent two years of my life on a game nobody online has played and survived to tell the tale. Woo! I'm probably killing people without realising. XD

There is a difference between total conversion mods and tweak mods.

Quote
But that's not the point of this thread. I'd agree that having them work off the rules rather than hardcoded settings is better as otherwise they'll be really lolligagged/freaked out by some mods (e.g. crawlers are disabled in my game, and some more mods, not sure what would happen if the AI beelined to them).

Well, at least that still gives hab complexes, which are also pretty important.  But what if mineral and energy lifting restrictions are moved, and it still beelines to Environmental Economics?

Quote
Oh by the way! Some good news. I managed to get both Kyrub's and Yitzi's patches working on my laptop. The catch is that they MUST use DirectDraw=0 using that patch or it won't run. The other catch is that DirectDraw=0 doesn't like my laptop either, as even the basic game with only scient's patch will crash in a few turns after a bit of exploration. :(

Sounds like a graphics issue with your laptop.
Title: Re: SE choices for AI - suggestions
Post by: Green1 on January 17, 2013, 07:04:38 PM
Some of these laptops have weird resolution. Top that off with a game designed for the drivers and cards of a decade ago, you have that problem. This even happens on modern games, but the modern games have far more flexible graphics settings than AC.

BTW.. I am loving all the discourse on AI. I have been staying silent to let the experts speak since I was a previous years long lurker and others can put what changes are good better than me.

I am sure there are many more like myself. In fact, I know there are.
Title: Re: SE choices for AI - suggestions
Post by: Kilkakon on January 18, 2013, 12:08:58 AM
There is a difference between total conversion mods and tweak mods.
Naturally, hence my line "at least for the big things".

Well, at least that still gives hab complexes, which are also pretty important.  But what if mineral and energy lifting restrictions are moved, and it still beelines to Environmental Economics?
That's what I mean--would be better to beeline to specific, valuable technologies, such the prereqs for certain facilities, tech req. for mineral lifting, etc. It's the perfect example as my mineral lifting tech is different from hab complexes and all the tech keys are changed.

Sounds like a graphics issue with your laptop.
Yeah definitely, it can't handle Baldur's Gate (even the GoG.com version) in full screen either, have to hard reset that (which is what happens with DirectDraw=0 off). So won't be able to use the nice patches for MP, but no biggie at this stage.

Agreed @ AI being better. :D
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on January 18, 2013, 01:35:41 AM
That's what I mean--would be better to beeline to specific, valuable technologies, such the prereqs for certain facilities, tech req. for mineral lifting, etc. It's the perfect example as my mineral lifting tech is different from hab complexes and all the tech keys are changed.

Actually, by default mineral lifting is different from hab complexes; mineral lifting goes with advanced terraforming (raise/lower/boreholes/mirrors/condensers), and hab complexes go with crawlers (and a commerce boost).

(In Alien Crossfire 1.5, however, I do plan to move mineral lifting to go with (greatly depowered) crawlers and hab complexes (but ditch the commerce boost), in order to have the three resource lifting techs be parallel (I plan to move energy lifting to Optical Computers, which is really looking lonely at the moment) rather than in succession, in order to reduce beelining (and make not everybody beeline for the same things).
Title: Re: SE choices for AI - suggestions
Post by: kyrub on January 19, 2013, 08:58:29 PM
@Yitzi
Please keep this thread free of discussion about modding, mod plans and about concrete existing mods. A mod changes game's rules, the AI patch, however, changes only the way how the game (with original rules) is being played by AI. These themes are quite different, I think. (I realise I made the mistake myself, while reacting to Kirov post, sorry.)

Other notes on AI or technical issues are most welcome.

Quote
BTW.. I am loving all the discourse on AI. I have been staying silent to let the experts speak since I was a previous years long lurker and others can put what changes are good better than me.

@Green
No expert stature is necessary. If you want to contribute, try to use the patch, write about your experience, what seemed weird/wrong. "Test" in the patch subtitle means it needs testing and I need feedback. I stopped developping previous AI patch versions because of week feedback.

Quote
In that case:
1. What do you have it beeline?  Some of it might be adjustable to be based on the tech that gives X feature.
2. Would it be possible to write separate AIs for beelining and not-beelining in the same program?  If so, I can show you how to make it based on an alpha.txt variable.  (Note: For SMAC, you will need to reduce some limit (I used the landmark limit) to make room.)
3. Did you include the beelining in the SMAX mod as well, or just SMAC?
4. While we're on the subject, what terraforming did you teach it in the SMAX mod?  In particular, is it going to use boreholes and condensers like crazy?


1. Not in this thread, please, start another one (like AI - research choices) and I will try to answer there.
2. This would be quite helpful. There should be an option to switch AI beelining off completely. BTW, if you put in the "Random research" option for yourself, it's not completely random, you are actually using the AI beelining... You should be able to choose "Completely random research" to increase the challenge.
3. No. SMAX patch has only the "basic terraforming lesson" included + the bug fixes. All other stuff is not there, even the extra care for formers and recycling tanks is gone...
4. It's basic terraforming, not the top stuff. Look there for details: http://apolyton.net/showthread.php/195007-SMAC-444-%28AI-experiment%29?p=5933522&viewfull=1#post5933522. (http://apolyton.net/showthread.php/195007-SMAC-444-%28AI-experiment%29?p=5933522&viewfull=1#post5933522.) If you want to respond, create a new thread on AI terraforming, please. Let's leave this thread focused on AI social engineering and its context.

Quote
Can you make it take immunities to negative stats and civics into account? E.g. Hive's immunity to negative EFFIC, or one of the secret project's immunity to Cybernetic? Or is that automatic in the considerations?

I think so. I'll give it a second look.

Title: Re: SE choices for AI - suggestions
Post by: Yitzi on January 19, 2013, 11:59:01 PM
1. Not in this thread, please, start another one (like AI - research choices) and I will try to answer there.
2. This would be quite helpful. There should be an option to switch AI beelining off completely. BTW, if you put in the "Random research" option for yourself, it's not completely random, you are actually using the AI beelining... You should be able to choose "Completely random research" to increase the challenge.

Send me a message when you're ready to hear the details of how to do it.
Title: Re: SE choices for AI - suggestions
Post by: Green1 on January 20, 2013, 03:07:51 AM
@Kyrub and Yitzi.

In an ancient post on CFC, there was a guy that figured out how to get all AIs to play themselves. It had something to do with going into world builder, setting the player faction to AI, then pressing END TURN till your finger fell off.

It has me wondering if there is a short hack to automate that process.

It would help with testing how they are doing with different SE choices.


Title: Re: SE choices for AI - suggestions
Post by: Petek on January 20, 2013, 10:43:16 AM
If by "automate that process" you mean how to avoid pressing END TURN repeatedly, you can set a weighted object on the ENTER key. I've set up simulations that ran overnight using this technique.
Title: Re: SE choices for AI - suggestions
Post by: kyrub on January 21, 2013, 12:29:20 AM
No need to wait all night.
The best way is to choose a faction you don't want to watch.

Start a game.
Go to Menu->Scenario->Eliminate
Choose your faction.
Switch off all notifications you don't want to see.
Now use "Y" key to hide everything.
Keep clicking Enter or End of turn.
If you want to have a look, press "Y", then "E" to study AI social engineering etc.

In 5 minutes, the game's done.
Title: Re: SE choices for AI - suggestions
Post by: kraze on January 21, 2013, 01:15:34 AM
There's an issue with AI about as major and game-affecting as crawlers but I think there are more chances of you fixing it.

Basically when you gather your forces next to an AI base to attack it and end your turn - the AI never counter-attacks (unless it has those weak artillery units and even then it attacks only with them) so next turn you are free to attack the base with all your troops/get additional troops in there.

This makes fighting AI too easy once you have strong enough forces - since when it gathers its units near your base you don't give AI any chance.
Even if AI has units with 13 attack and your units have 5 or less defense - it never attacks with them from inside the base, considering that often they have like 1-5 defense themselves so they aren't good defenders.

Is it possible for you to make AI attack gathered units from inside a base when it thinks it has enough power to do so (like some kind of city defender power X vs. enemy power condition Y where X and Y are  balanced with taking into account enemy striking back next turn)?



Also with the current patch AI seems to ignore building space units and also doesn't seem to use transports to cross the sea at all. Granted it's been a while since I've played the game last time but I'm fairly certain AI colonized overseas locations at least even if it didn't do invasions.
Title: Re: SE choices for AI - suggestions
Post by: Petek on January 21, 2013, 01:31:12 AM
No need to wait all night.
The best way is to choose a faction you don't want to watch.

Start a game.
Go to Menu->Scenario->Eliminate
Choose your faction.
Switch off all notifications you don't want to see.
Now use "Y" key to hide everything.
Keep clicking Enter or End of turn.
If you want to have a look, press "Y", then "E" to study AI social engineering etc.

In 5 minutes, the game's done.

I'll have to try that. Thanks!
Title: Re: SE choices for AI - suggestions
Post by: Buster's Uncle on January 21, 2013, 05:02:49 PM
Welcome to AC2, krase!  How did you find us?
Title: Re: SE choices for AI - suggestions
Post by: kyrub on January 21, 2013, 08:34:36 PM
Is it possible for you to make AI attack gathered units from inside a base when it thinks it has enough power to do so (like some kind of city defender power X vs. enemy power condition Y where X and Y are  balanced with taking into account enemy striking back next turn)?
Good idea, Kraze.
Don't know when it will be possible to improve unit movement. This is very far from us now. Not impossible, I'm rather optimistic, but far.

Quote
Also with the current patch AI seems to ignore building space units and also doesn't seem to use transports to cross the sea at all. Granted it's been a while since I've played the game last time but I'm fairly certain AI colonized overseas locations at least even if it didn't do invasions.
AI colonizes overseas even with the patch. I have witnessed it myself. Please, beware from making quick judgments based upon one game or one segment of a game. We need more evidence, more refined notions. You may have found a flaw in the patch, but we need more details, better understanding of the conditions under which something works / does not work. - Thanks!
Title: Re: SE choices for AI - suggestions
Post by: Kilkakon on January 22, 2013, 02:47:25 PM
I can confirm that the unkyrubed (scient) AI can colonise and invade. I've seen both in LE.
Title: Re: SE choices for AI - suggestions
Post by: kraze on January 23, 2013, 05:28:20 AM
Let me clarify on my observations.

I did a full runthrough of SMAC at a Librarian difficulty with the current patch and the map was generated in such way that me and 2 other factions ended up being divided from the other 4 by the sea (think 2 big continents, 1st had 3 factions, 2nd had 4). For the following 200 years no faction from another continent managed to come into a contact with any of "home" factions up until Zakharov built an Empath Guild.

I also was the only faction that was building/launching space stuff. And when I revealed the map with the satellite (300 years into the game) there were a lot of big uncolonized islands and zero overseas colonies.

I also had wars with overseas factions but they never tried to invade me.

It's only by the very end (~2450) that I saw factions building sea colonies and one even captured a base overseas.

I certainly do not remember them ignoring sea stuff so hard in vanilla version. And I'm also certain they were building satellites too.


It's just what I experienced.
Title: Re: SE choices for AI - suggestions
Post by: kyrub on January 23, 2013, 09:31:52 AM
Thanks for the info, kraze.
I am seeing a lot of naval bases, orbital stuff, overseas expansion in my testgames.  That is why I was surprised.
The thing done to AI in naval section are:
- serious increase of interest for naval formers, even if no naval bases are built
- increse in early naval base expansion (contrary to your findings)
- and lately, on request of one player, decrease of plus factor for naval transports per colony

As far as I see, only the last one could really have done some harm to AI in your game (but I may be mistaken).
Could you please post your basic save (2101)? I would like to rerun the game. Also: did you use "small lands, huge world" setting?
Title: Re: SE choices for AI - suggestions
Post by: kraze on January 23, 2013, 11:23:29 AM
I have only 2109 year save from that game unfortunately
But I also included the final save for you to compare

It also appears that AI was building bases on other continents only because factions made sea levels drop through voting enough to make that sea passable.

http://www.mediafire.com/?nzv15y8ox8axgaz (http://www.mediafire.com/?nzv15y8ox8axgaz)

Check the satellite count too.

I've used GoG's version of SMAC + your jan 13 patch.

Quote
Also: did you use "small lands, huge world" setting?


Nope. Huge world, minimum sea coverage.
Title: Re: SE choices for AI - suggestions
Post by: kyrub on January 23, 2013, 11:42:30 AM
I have only 2109 year save from that game unfortunately
But I also included the final save for you to compare
Got it. Will run a few watch-games on it tonight, with Vanilla and the patch.
Thanks a lot.
Title: Re: SE choices for AI - suggestions
Post by: Kirov on January 23, 2013, 12:24:26 PM
Since I am going away for a few weeks, I cooked a quick test build for SMAC AI (requires SMAC only for now, sorry). SMAC_444_l version in the Downloads.
The highlight is the correction of Support <=> Police mismatch in AI social engineering. You'll find the rest in the short notes to the patch.

I started one game some time ago with that new patch to check for SE behaviour, however I’m afraid I’m too busy these days to play it off, so I switched to watch AI as Petek prescribed. I started with time warp 2161-2185 then handed over to AI until 2220.

First let me ask about the things I don't understand. What does the following mean in your patch description?

d  close base and fungus love from 70
e   unit_abilities_8
g   formers chance

Ok, now the game itself:

I started in 2161 as PK. The starting SE for the AI was:

Deirdre:    Demo/-/-           (although she had access to Green, and also PS and Fundie)
Hive:       -/Planned/-         (access to PS, Wealth, FM)
Zak:         Demo/FM/-        (access to Planned, Power)
Morgan:   Demo/-/Wealth (access to FM)
Sparta:    -/FM/-                 (access to Planned)
Miriam: -/Planned/-           (access to FM, Wealth)

There were no Vendettas at that moment. All in all, the AI was quite peaceful.

In 2185 SE was:

Deirdre:    Demo/-/Wealth      (access to PS, Fundie, Planned, Green, Power)
Hive:       PS/Planned/-           (access to FM, Wealth) 1 Vendetta
Zak:         Demo/FM/-             (access to PS, Planned, Power, Wealth)
Morgan:   Demo/FM/Wealth   (access to PS, Power)
Sparta:    PS/FM/-                   (access to Planned)
Miriam: -/Planned/-                (access to Demo, FM, Wealth) 1 Vendetta.

Then I watched AI playing until 2220:
Deirdre:    Demo/-/Power             (access to PS, Fundie, Planned, Green, Know, Wealth) 1 Vendetta
Hive:       PS/Planned/Wealth       (access to FM, Know) 4 Vendettas
Zak:         Demo/Planned/Know    (access to PS, FM, Power, Wealth) 2 V
Morgan:   Demo/-/Wealth            (access to PS, FM, Power, Know) 2 V
Sparta:    Fundie/Planned/-          (access to PS, Demo, FM, Power, Know) 1 V
Miriam: -/Planned/-                       (access to PS, Demo, FM, Wealth) 2 V
Lal: Demo/Planned/Power            (access to Fundie, FM, Know, Wealth) 4 V

All in all, I see big improvement in:

1) not sticking to preferred SE at all costs
2) use of Wealth by Morgan and Hive, really congrats guys and good for you
3) smarter use of FM (i.e. AI uses it but switches off under Vendetta)

However, I still think that Wealth is under- and Power is overused. Can Power be made as something of last resort? Or barring that, something which is used only if >95% bases build combat units?

Also, Deirdre doesn’t use any economics now at all (I say both Planned and Green are better than nothing in the long run), Miriam could use Wealth and maybe Demo if not war.

Is there any way we can relate it to those five magical factors you mentioned? Wealth is just as good for aggressive factions, I don't know how to explain it to the AI...

Which reminds me - what's the difference between aggressivity and power? (you mentioned them along with growth, eco and commerce).

Title: Re: SE choices for AI - suggestions
Post by: Kilkakon on January 23, 2013, 12:30:44 PM
Based off those findings, I would prefer them having their preferred "agenda" SE choice over the basic one, if the others weren't an improvement.

So Best > Agenda > Simple?

Seems not bad though, just the lack of agenda of Deirdre/Miriam/Morgan seems a little off, although Morgan does have 2 wars going so fair enough there.
Title: Re: SE choices for AI - suggestions
Post by: Kirov on January 23, 2013, 12:41:34 PM
So Best > Agenda > Simple?

This seems like a good idea to me, I think it can be related somehow to what kyrub mentioned earlier, with the agenda SE being preferred, but other options are still used.

I agree Deirdre should run Green over nothing, but I'm not so convinced that I want to watch Miriam trying to bash coconuts with sticks as she runs Fundie. I've seen that a lot in my life and I just can't help but think of her as a complete retard. As far as Morgan and FM is concerned - if AI can't manage pacifism drones, then the simplest rule is to give up FM when at war, otherwise the FM-er in question is bound to suffer a lot.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on January 23, 2013, 01:29:43 PM
I agree Deirdre should run Green over nothing, but I'm not so convinced that I want to watch Miriam trying to bash coconuts with sticks as she runs Fundie.

Maybe instead it would make more sense to teach her to probe tech.
Title: Re: SE choices for AI - suggestions
Post by: kyrub on January 25, 2013, 11:38:25 AM
First let me ask about the things I don't understand. What does the following mean in your patch description?
d  close base and fungus love from 70
e   unit_abilities_8
g   formers chance
These are my internal notes, they give little sense in themselves, sorry.
d - close base, I forced AI to place bases not as close (I dislike the ICS). I reversed the change afterwards, because it changes game's rules. It works normally now.
d- fungus love: when AI places new bases, it has a strong dislike for fungus. As a result, it failed to populate whole zones on its continents. I put a time cap on the dislike, to free the AI colonization, from 70 turns it ignores the fungus threat.
e - I made a whole new AI design set (8 means it is the 8th generation of the set in my book). The AI now designs new units (like AAA/ECM defenders, or clean garrison, submarine ships, SAM rovers etc.). You may verify that by clicking on AI cities in test mode, just take a look at its build choices. The trouble is, even if I designed them, the AI still fails to build them most of the time. -- I need to make another step and improve the unit building routine now.
g- AI builds a LOT more formers. This i quite evident from the game, I think.

Quote
All in all, I see big improvement in:

1) not sticking to preferred SE at all costs
2) use of Wealth by Morgan and Hive, really congrats guys and good for you
3) smarter use of FM (i.e. AI uses it but switches off under Vendetta)
Thanks a lot for some positive feedback, we achieved something without actually changing a lot.

Quote
However, I still think that Wealth is under- and Power is overused. Can Power be made as something of last resort? Or barring that, something which is used only if >95% bases build combat units?
If we make INDUSTRY setting more pertinent, the AI will prefer Wealth and use Power only as a last ressort. This should do the trick, easy.[/quote]

Quote
Which reminds me - what's the difference between aggressivity and power? (you mentioned them along with growth, eco and commerce).
It's "ai-fight", the first number of the five. Willigness to use army (the number can range from -1 to 1.). I say aggressivity, because it seems to fit better to the fact. This number is quite powerful throughout the code. Of all numbers, most used are (sorted): ai_fight, ai_power... ai_tech, ai_wealth. Ai_growth is severely underused, it's strong only in the research choice, IIRC. Surprisingly, AI growth does not lead to ICS.

As far as those numbers go, we could MAYBE tie them to specific situations, without nullifying them. Like, when AI is the last in power chart AND not in war, behave as if ai_tech = 1, ai_wealth = 1 (runs FM/DEMO, tries to catch with tech). Or when last in power chart and in war, behave as if ai_fight=1 and ai_tech=1, to try to probe enemies.

This, instead of rewriting the whole AI, we could use those numbers for "behaviour under special conditions". Normally, AI would still use their own traits, just in special circumstances it could change them.
Title: Re: SE choices for AI - suggestions
Post by: Kirov on February 01, 2013, 10:00:23 PM
d - close base, I forced AI to place bases not as close (I dislike the ICS). I reversed the change afterwards, because it changes game's rules. It works normally now.

This could also discourage AI from building any further bases at all. I don't know if you changed anything, but in this test I played (it's 2220), I noticed all AI have 1-3 active colony pods and look like they are not sure what to do about it (the numbers are 1, 3, 1, 1, 2, 1).

I checked my old SMAX game (MY 2261) and the AI have 0-2 colony pods. So maybe the difference is not that big. Still, it's better that AI dump colony pods anywhere than just roam around for no apparent reason.

Quote
g- AI builds a LOT more formers. This i quite evident from the game, I think.

The numbers are 4, 13, 3, 3, 8, 1, so it's seem fine. I wonder why the difference between 1 and 13? Is there any way to introduce a lower limit ('thou shalt not have less than 2/3 X # of bases')?

Meanwhile, for comparison purposes I checked my other game and funny thing - 3 AI have no formers and 3 other haven't even prototyped the basic one! Some of them still have fungicidal formers, to be fair. And Miriam has 24 regular formers. These are some huge differences, only partially accountable by wars waged. Do they come from ai_growth or map factors?

Quote
It's "ai-fight", the first number of the five. Willigness to use army (the number can range from -1 to 1.). I say aggressivity, because it seems to fit better to the fact. This number is quite powerful throughout the code. Of all numbers, most used are (sorted): ai_fight, ai_power... ai_tech, ai_wealth. Ai_growth is severely underused, it's strong only in the research choice, IIRC. Surprisingly, AI growth does not lead to ICS.

We are talking about how these numbers affect the AI behaviour. Can you say a few words how they can in turn be affected by external factors (specific faction choices, map, game settings, situation on the map)? Maybe we could manipulate it in this way. 'If 1st in military - increase ai_fight.' 'If alone on a continent - decrease ai_fight'. Something along these lines.

Also, can they be interrelated or are they uber-factors? 'If ai_tech=1, then increase ai_growth'.
Title: Re: SE choices for AI - suggestions
Post by: kyrub on February 02, 2013, 12:17:30 AM
I am quite busy, working on a few things. I get some better results on advanced terraforming (boreholes and condensers) now and I have got clear image of AI unit's design and building decision.  I found some of the AI to AI diplomacy code, which could really help to make AI's mood a bit more stable. Trouble is, it will affect the relations to human player as well... SMACX is a sincere game, everybody gets the same diplomatic treatment!

This could also discourage AI from building any further bases at all. I don't know if you changed anything, but in this test I played (it's 2220), I noticed all AI have 1-3 active colony pods and look like they are not sure what to do about it (the numbers are 1, 3, 1, 1, 2, 1).
There's no change, I just pushed some factions to be more pro-active in spreading (Morgans, University). It's quite visible. AI has a general problem to place its colonies, the pods go round in circles. The bug source is not known, yet.
 
Quote
The numbers are 4, 13, 3, 3, 8, 1, so it's seem fine. I wonder why the difference between 1 and 13? Is there any way to introduce a lower limit ('thou shalt not have less than 2/3 X # of bases')?
Room for improvement? Surely. I am currently working on better AI building, so Formers can fall in. I am pretty positive about the potential improvement here, it will make AI better.

Quote
We are talking about how these numbers affect the AI behaviour. Can you say a few words how they can in turn be affected by external factors
Oh no, these are constants from the faction txt files. Open Morgan.txt and the 5 numbers you see in the header are the constants that influence 60% of AI behaviour in the game. That's why AI Hive is usually the best, it has the most succesful number combination.

What I suggested was to make them not_constant. Switch ai_fight on +1 even for "peaceful" factions (-1), under some defined circumstances.


EDIT: I've found the answer on the new wiki (thanks ete).
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on February 03, 2013, 12:27:01 AM
What I suggested was to make them not_constant. Switch ai_fight on +1 even for "peaceful" factions (-1), under some defined circumstances.

It might not be a good idea to mess with ai_fight, as think that really determines more what circumstances it prefers (i.e. aim for peace or not) rather than how it acts in a particular circumstance.  If it's in a major conflict, I'd think you'd want more focus on ai_conquer and leave ai_fight alone.
Title: Re: SE choices for AI - suggestions
Post by: kyrub on February 04, 2013, 02:49:13 AM
It might not be a good idea to mess with ai_fight, as think that really determines more what circumstances it prefers (i.e. aim for peace or not) rather than how it acts in a particular circumstance.  If it's in a major conflict, I'd think you'd want more focus on ai_conquer and leave ai_fight alone.

Try to read this thread: http://apolyton.net/showthread.php/19938-Re-Balancing-SMAX-Factions?highlight=Switching%20Sides (http://apolyton.net/showthread.php/19938-Re-Balancing-SMAX-Factions?highlight=Switching%20Sides)
They think ai_fight has massive influence on how much army is built, how offensive / radical the AI behaves... What I see, seems to confirm that. AI_conquer is less relevant.
Title: Re: SE choices for AI - suggestions
Post by: BFG on February 04, 2013, 04:50:48 AM
Kyrub, if you modify the AI so that it knows how to use gravship formers/supplies and aerial colonies properly, then I will be forever indebted to you!
I think I mentioned this to you on another board, but just having an AI that's capable of aerial colonization would, in my opinion, vastly strengthen it.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on February 04, 2013, 06:08:54 AM
Try to read this thread: http://apolyton.net/showthread.php/19938-Re-Balancing-SMAX-Factions?highlight=Switching%20Sides (http://apolyton.net/showthread.php/19938-Re-Balancing-SMAX-Factions?highlight=Switching%20Sides)
They think ai_fight has massive influence on how much army is built, how offensive / radical the AI behaves... What I see, seems to confirm that. AI_conquer is less relevant.


Even so, I'm pretty sure that it also affects how much the AI tries to seek or avoid war in the first place, and that probably should not be situation-dependent (at least not for the same situations).
Title: Re: SE choices for AI - suggestions
Post by: ete on February 04, 2013, 11:10:26 AM
Kyrub, if you modify the AI so that it knows how to use gravship formers/supplies and aerial colonies properly, then I will be forever indebted to you!
I think I mentioned this to you on another board, but just having an AI that's capable of aerial colonization would, in my opinion, vastly strengthen it.
By the stage of the game when air is available, I'm generally not needing to build many of bases, and though getting there a few turns earlier is cool.. it seems to be a pretty minor thing compared to many of the AI's other weaknesses. And Gravship formers? Why not just make tank/speeder and sea ones, on land you'll have a magtube network by then anyway so they can get many places faster than a gravship. Plus by Gravships the AI is effectively always doomed, due to other issues.
Title: Re: SE choices for AI - suggestions
Post by: BFG on February 04, 2013, 05:25:49 PM
And Gravship formers? Why not just make tank/speeder and sea ones, on land you'll have a magtube network by then anyway so they can get many places faster than a gravship. Plus by Gravships the AI is effectively always doomed, due to other issues.
Frankly, I disagree.  Aerial colonies - as early as Needlejets - would open up new possibilities for the AI: rapid expansion to deal with drone problems, open up new vectors rapidly in Vendettas, settlement of far-off islands that would otherwise require a Transport + Colony pod (i.e. more turns building, and more support resources) and 5-10x as many turns, etc.  That said, I do agree that the AI has more glaring flaws than this one that need dealt with.

And Gravship formers? Why not just make tank/speeder and sea ones, on land you'll have a magtube network by then anyway so they can get many places faster than a gravship. Plus by Gravships the AI is effectively always doomed, due to other issues.
This is really just a personal annoyance of mine - I like replacing land+sea formers with Gravship ones, as they're more efficient support-wise and I think they look cool.  Plus they get additional protection due to being aerial units.  And I've got to believe it's a minor bug that causes Gravship formers to automate properly on land, but not on sea.  But yes, this has little if any priority in AI fixing :)
Title: Re: SE choices for AI - suggestions
Post by: ete on February 04, 2013, 05:41:44 PM
Frankly, I disagree.  Aerial colonies - as early as Needlejets - would open up new possibilities for the AI: rapid expansion to deal with drone problems, open up new vectors rapidly in Vendettas, settlement of far-off islands that would otherwise require a Transport + Colony pod (i.e. more turns building, and more support resources) and 5-10x as many turns, etc.  That said, I do agree that the AI has more glaring flaws than this one that need dealt with.
Rapid expansion tends to lead to more drone problems due to BDrones, even with high efficiency. Having a base as a recharge point in an attack is interesting, but unless the AI was very smart about it may just give away a base. The newly founded base is rarely going to contribute much to an attack other than healing which will be further away from the action than a captured base, and bringing a speeder colony pod along with an attack would do almost the same thing.

Expanding to far away islands in a way the AI is better at may be worthwhile, but ultimately bases nearish to the HQ are going to be a lot more valuable to an empire due to inefficiency issues and the fact that they can be defended much more easily (plus bases founded nearby can be improved by your main former pack, faraway bases are on their own).

Actually, I think perhaps the biggest advantage to flying colony pods would be bypassing all the fungus and the worms hiding in it, which may well pick off normal colony pods otherwise. This is somewhat notable, but not huge compared to many other AI problems.

I do think it would be kinda cool for AI to expand like this, but I'm not sure how much more of a challenge it would provide.

This is really just a personal annoyance of mine - I like replacing land+sea formers with Gravship ones, as they're more efficient support-wise and I think they look cool.  Plus they get additional protection due to being aerial units.  And I've got to believe it's a minor bug that causes Gravship formers to automate properly on land, but not on sea.  But yes, this has little if any priority in AI fixing :)
Long before Gravships all your formers should be Clean, negating support entirely, and you should have a tube network, making speeder formers faster in many situations. I could see gravship formers having a very few niche uses, and they do look cool, but teaching the AI when they're worthwhile and how to use them seems entirely not worth it. Maybe making automation work for them (at least a copy of Land automation) would be much simpler and so worthwhile though.
Title: Re: SE choices for AI - suggestions
Post by: kyrub on February 08, 2013, 02:31:29 AM
I have some bad news and a few good ones.


First, the bad and my sincere apologies. I thought I found a bug, the AI mixing up POLICE setting with SUPPORT. After hours of checking, I found I was wrong. Sorry.

Second, based on this disapointment, I decided to overview the whole SE code, in detail. I have found 3 bugs (this time I hope they are real!) and a number of underachievments.

The bugs are placed:
a) in POLICE setting code. The program carefully counts a large, negative Drone_factor, but in next step, it is multiplied by Police_Setting and added to the score. Now, let's imagine what happens for negative police* negative drone factor... positive result is added! --- Consequence: The AI with large unrest thinks that having POLICE -3 is simply great.
b) in PLANET setting code. Again, substracting what should be added. Consequence: The AI with Green predispositions (those who have ai_Growth) will try to avoid Green more than any other player. That's why Deirdre seems so reluctant to use her favourite Green, as Kirov noticed!
c) in EFFICIENCY setting code. Efficiency is wrongly factored in, so that AI does not recognize the effect it could have on improving economy. I have noticed this strange ignorance to efficiency in a test game with Spartans.

Other flaws
d) in GROWTH, AI tends to overestimate the growth potential even for bases that lack nutrient surplus. A large empire in mid-late game may thus go for large GROWTH, without significant effect.
e) AI seems to use precalculated Tech_deficit_to_HP in a questionable way. It never tries to catch up with favouring Energy, Efficiency or Tech, when low on Tech scale. Instead it goes immediately for power, ignoring the fact that a player who is 15-20 techs ahead won't probably crumble. If the AI has tech supremacy, it refuses to capitalise in power, on the other hand. Further refining of these choices is probably necessary.
f) sum_of_units variable is used a lot, which IMO makes AI vulnerable to its own strategy to create and store large hosts of soldiers. Morale is increased, if army is large, but ++Morale is a positive factor for building more soldiers as well. One can easily see, where this leads in the long run.

(.. and there is more.)
Title: Re: SE choices for AI - suggestions
Post by: ete on February 08, 2013, 11:22:35 AM
Interesting finds! Especially the Police/Support multiplier, that could well be very significant. Would it be practical to post/share your disassembled version with functions you've identified, so that perhaps others would be encouraged to join in the search for improvements, or at least gain insight into the workings of the game?
Title: Re: SE choices for AI - suggestions
Post by: kyrub on February 08, 2013, 10:32:35 PM
Interesting finds! Especially the Police/Support multiplier, that could well be very significant. Would it be practical to post/share your disassembled version with functions you've identified, so that perhaps others would be encouraged to join in the search for improvements, or at least gain insight into the workings of the game?
It's police/police multiplier, I made a typo. Sorry.


As far as sharing of disassembled database, I understand your interest in testing and probing AI. I love studying current AI algorithms and finding new ones and better, same as you, ete.
But the short answer to your question is: not now.

The longer answer: I will most probably disclose my IDA database later, in months to come, to wider community for any use. I did so in all my other projects so far (bar one that was first and is still unfinished). So you may more or less rely on the fact.
Right now, I am doing the AI business myself, in the pace I set to myself, without any pressure and enjoying all the surprise and fun it offers. It's a bit egoistic but it's well deserved at the same time (see below). When I am done with my effort, I'll open the database to the community and you may improve it or build a new concurential model. You may try to create a scripted AI as well, if you wish and have abilities to do so.


To explain my stance a bit further: the ground disassembling is the really hard part, the rest are the fruits. When you disassemble a game, the first stage is "simple findings", very fruitful (like combat bonuses or POP limits locations). This is short. The second stage is very hard and can take a hundred of hours time, because you are trying to understand the non-obvious in the code. It also brings few fruits. After that, you may try to understand the AI. This is the hardest, because you must penetrate something that is not visible in the game AT ALL: how the AI uses its own data, what analysis it does for its goals, how it communicates in AI-AI mode. Also, without understanding the AI more than 80%, you fail to improve it. That's another hundreds hours thing (for SMAC I mean, other games have much simpler AI).

I have put this effort in the game, and that's why I tend to think I can enjoy a small time for tinkering with it, before I feed to the player's community. I am trying to do my work to help the players, but, for now, I do it privately, myself. - Hopefully that does make some sense to you and others here.
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on February 09, 2013, 01:59:26 AM
Sorry to be late chiming in here, but here are my two cents about SE:

Line 1: Unless you are Hive, all faction run Democratic.  Period.  No calculation needed.  Change as soon as you research it.  Hive will run Police State.  Immediately, as soon as you get the energy to change.  And do not waste the energy on anything else.

Line 2: ALL of the options are good in one situation or another.  But, other than Morgan maybe, I would not allow any other AI faction to run FM.  It is just too complicated to manage pacificism.  I don't see how an AI could be programmed to do it.  I would start all factions who can with Planned, and only move to Green based on: A. I am exploring fungus a lot and would like to attempt mindworm capture, or B. I am losing too much energy to inefficiency, and so I will give up the Planned bonuses to get a net +4 improvement in Efficiency.

Line 3: I would agree with Kirov that Wealth is the answer, but that is only taking the posititives into account.  If you are attacking, you may find having -2 Probe to be less troublesome than -2 Morale, so you might choose Knowledge.  But unless you have something that gets rid of the Power negatives, do not choose Power.

Line 4: Most AI would probably most benefit from the energy boost, so go with that, and combine with Wealth.  Yes, that is -4 morale, so maybe not if you are hot and heavy at war, but maybe it is worth it anyway.

Hope this helps.
Title: Re: SE choices for AI - suggestions
Post by: kyrub on February 09, 2013, 02:47:38 AM
Thanks a lot for important insight. There's a problem, that AI does not think in SE choices (e.g. "I like Democracy", or "Bleh, Free Market"), but rather in how good/important/harmful is settingX=Y. This should be conserved, as players are often modding the SE choices. Looking at your suggestions, I could strengthen Growth factor, and AI will go for the choices you emphasized more. But it will be hardly as straightforward.


That aside, I have a few important questions about pacifism drones: what's the problem with them? Why AIs cannot manage, in your opinion? Another question (trivial), is POLICE = -3 supposed to create more drones than 1 in every base (because that seems so in the code, as if "For every unit over 1 outside of borders, there's a PD" - while the help.txt suggests that only 1 pacifism drone is created in every base)?

Also, is it true that PDs cannot be stopped by classic facilities, like hologram theatre etc.?
Title: Re: SE choices for AI - suggestions
Post by: Lord Avalon on February 09, 2013, 03:03:32 AM
Are you saying the code has 1 pac drone for each 2-n military units away from base at Police -3?  Per manual it seems like there should be only 1 pac drone.

At Police -4 it goes up to extra pac drone for each military unit away from base.
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on February 09, 2013, 04:21:42 AM
Lord Avalon is correct below, and I modified my post to try to give the correct info.  Sorry for the initial mistake.

Police -4 gives one drone for each unit out of the area AND one drone for each aircraft built that is not an intercepter (regardless of whether the aircraft is in your area or not).  There are no "free" units; every single unit out of the area adds a drone.  Police -3 says it give one free unit, but I have never tested it. 

Furthermore, these pacifism drones are added AFTER most other drone handling facilities have had their effect, so they are not stopped by these facilities.  Which makes pacifism drones particularly hard to deal with; the only way I know to deal with it is to create enough talents or specialists to offset the drones.  But I typically use the "tainted city" trick I discuss below to avoid having this problem except in a single city.

Police -5 DOUBLES this to two drones per unit and per aircraft.  This is usually where I am at when playing FM.

As a human, I take care not to move any units out of my area except the independent units (typically these are only units at the start of the game, or from pods popped far enough away from a city).  I use some tricks to help deal with pacifism drones.  First, it is best to have one city to be set as the home for any offensive military unit.  But that city will need a lot of crawler support, since this city can support no workers at all.  You may need a crawler or two for food so that you can grow your base to size 5 to get your choice of specialists, but most crawlers will be mineral.  Since each non-clean unit homed to the city requires one mineral, you will need to crawl minerals to support every offensive unit, including every aircraft.  In theory you do not have to home intercepters to the city, but in practice sometimes an intercepter will move out of your area and create drones anyway, so it is best to home all aircraft.  Then you can have a problem with ecodamage if you are crawling a lot of minerals into the city.  At some point, if you have enough offensive units, you may need a second city due to ecodamage of supporting them all in one city.

Once you get the tech you can build a punishment dome so that any population can work the squares, but this does not make that much difference, since the minerals gained this way will probably not help all that much in terms of reducing crawler support.

I am not sure you can very easily program this kind of strategy into an AI, which is why I said it was probably best for the AI to avoid FM.
Title: Re: SE choices for AI - suggestions
Post by: Lord Avalon on February 09, 2013, 05:08:17 AM
Hmm, another that's-not-what-the-manual-says moment.  Supposedly it's Police -5 that gives 2 drones per external military unit.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on February 10, 2013, 01:24:50 AM
By the way, this is vaguely off-topic, but once you've solved the big AI problems I'm wondering if you might be interested in making a version of your patch without the AI beelines?  The beelines are going to throw things off horribly if the tech tree is changed in a mod, and I get the idea they aren't really that difficult to do anyway in alphax (just increase the value of the relevant techs for the relevant foci.)
Title: Re: SE choices for AI - suggestions
Post by: kyrub on February 10, 2013, 10:19:26 PM
I checked again and this is how Pacifism drones work:

1) POLICE = -3
Every unit after the first one outside of your territory causes +1 pacifism drone in your base.
(This should really be corrected in the new Datalinks)

2) Pacifism drones are added in the "police phase", the one before the last phase. That's why they are not nullified by classic facilities. For the rank of phases, you may look in to your "detailed" view in the base unrest window.

3) The last phase is the Special project phase. This means that Human Genome Project, The Longevity Vaccine and the Clinical Immortality SPs can stop pacifism drones even after they are created.

4) Punishment sphere is included in Special project phase.

5) AI does not check for pacifism drones when trying to make SE choice.


---------- Beelines -----------

Quote
By the way, this is vaguely off-topic, but once you've solved the big AI problems I'm wondering if you might be interested in making a version of your patch without the AI beelines?
I have no problem with it and I find it desirable.

As far as numbers go, I actually stripped off the original AI from about 1/5 of beelines. (I am not a big fan of AI beelines, unless they are really sophisticated, I dislike formulaic AIs.) The only thing I really changed (IIRC), was a very strong push for Centauri Ecology. I think I included Industrial Automation beeline, as well as the tech lifting mineral/nutrient limits. Reactors, as well (not sure). These were important, I thought.

You may look yourself. The tech_value AI fonction starts at 1d3f90h.
The IsInBeeline fonction is 1d3f10h. Arg4 is tech (number according to alpha.txt). Arg8 is beeline depth.
If you put these in alpha.txt and make them customizable, I'll be grateful, in fact. At least people get some material to tinker and mod AI.

! - Please remember, that "Blind research" option now uses these AI beeline routines, in fact. It's not very blind at all. The modders should know that as well, it should be mentioned in your new alpha.txt file.
Title: Re: SE choices for AI - suggestions
Post by: ete on February 10, 2013, 11:09:32 PM
As far as sharing of disassembled database, I understand your interest in testing and probing AI. I love studying current AI algorithms and finding new ones and better, same as you, ete.
But the short answer to your question is: not now.

The longer answer: I will most probably disclose my IDA database later, in months to come, to wider community for any use. I did so in all my other projects so far (bar one that was first and is still unfinished). So you may more or less rely on the fact.
Right now, I am doing the AI business myself, in the pace I set to myself, without any pressure and enjoying all the surprise and fun it offers. It's a bit egoistic but it's well deserved at the same time (see below). When I am done with my effort, I'll open the database to the community and you may improve it or build a new concurential model. You may try to create a scripted AI as well, if you wish and have abilities to do so.


To explain my stance a bit further: the ground disassembling is the really hard part, the rest are the fruits. When you disassemble a game, the first stage is "simple findings", very fruitful (like combat bonuses or POP limits locations). This is short. The second stage is very hard and can take a hundred of hours time, because you are trying to understand the non-obvious in the code. It also brings few fruits. After that, you may try to understand the AI. This is the hardest, because you must penetrate something that is not visible in the game AT ALL: how the AI uses its own data, what analysis it does for its goals, how it communicates in AI-AI mode. Also, without understanding the AI more than 80%, you fail to improve it. That's another hundreds hours thing (for SMAC I mean, other games have much simpler AI).

I have put this effort in the game, and that's why I tend to think I can enjoy a small time for tinkering with it, before I feed to the player's community. I am trying to do my work to help the players, but, for now, I do it privately, myself. - Hopefully that does make some sense to you and others here.

Wanting to play around with code yourself given the effort required to get that far is is entirely understandable. I look forward to being able to take a look, and see what other programmers may find and improve so I'm very happy that you plan to release it at some point, but you seem to be doing great things and if having the code with your notes to yourself improves your motivation and enjoyment then by all means keep it to yourself until you're happy to share.

Also, could you clarify your work on SMAX/SMAC? How much of your research can be applied directly/almost directly to SMAX, and are you currently still working almost entirely on SMAC?


On topic: "1) POLICE = -3
Every unit after the first one outside of your territory causes +1 pacifism drone in your base.
(This should really be corrected in the new Datalinks)"
Checked in SMAX, and corrected (http://alphacentauri2.info/mediawiki/index.php?title=Helpx.txt&action=historysubmit&diff=895&oldid=366).

Quote
5) AI does not check for pacifism drones when trying to make SE choice.

Giving them this check properly could well be the key to making them handle FM drones properly, or learn to avoid FM unless they can. I'd discourage entirely dissuading AI from FM unless there's no reasonable way to prevent them incurring far too many drones, since FM is extremely helpful to a faction if used correctly. If EM's methods are too complex to program in as they may be, perhaps adjusting build priorities so no non-Air Superiority fighters are made if police is low and encourage the AI to keep all units in borders? Perhaps that too would be extremely complex.
Title: Re: SE choices for AI - suggestions
Post by: kyrub on February 10, 2013, 11:45:40 PM
Wanting to play around with code yourself given the effort required to get that far is is entirely understandable.
Thanks for that, ete.

Quote
Also, could you clarify your work on SMAX/SMAC? How much of your research can be applied directly/almost directly to SMAX, and are you currently still working almost entirely on SMAC?
I am using SMAC for changes, since my knowledge was gathered in SMAC exe database (the postSMAX version 4.0). To project the changes into SMAX requires some time and energy, the SMAX exe, while being extremely similar, has different addresses. - So I avoid it until I am sure the changes are tested and valuable. Once we have a whole AI module (like: SE choices, or building decisions, or terraforming) ready, tested, approved , it can be transferred to SMAX.

Quote
Quote
5) AI does not check for pacifism drones when trying to make SE choice.
Giving them this check properly could well be the key to making them handle FM drones properly, or learn to avoid FM unless they can. I'd discourage entirely dissuading AI from FM unless there's no reasonable way to prevent them incurring far too many drones, since FM is extremely helpful to a faction if used correctly.
I'd try the same thing. I have seen AI Morgan keep his units inside his territory for whole periods of his game, so there is some space for cautious optimism.
Title: Re: SE choices for AI - suggestions
Post by: Lord Avalon on February 11, 2013, 03:11:13 AM
Should pacifism drones come after regular facilities?  It's a pretty strong penalty if (nearly) every military unit is causing a drone, and very strong for two.  If you have unused calming potential, why shouldn't that be applied?

Don't you hate it when you get drones because a unit that was inside your borders, but due to a border change, now is not, or you pop a unit that gets assigned a home base - and on top of that, it takes a few turns to get back to your territory?
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on February 11, 2013, 04:35:29 AM
Actually, Lord Avalon, what you mean to say is that Pacifism drones should be assessed first, then all drone management facilities, projects, etc. should apply.

I agree, but I am not sure this is something that can be modded.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on February 11, 2013, 04:50:55 AM
Actually, Lord Avalon, what you mean to say is that Pacifism drones should be assessed first, then all drone management facilities, projects, etc. should apply.

I agree, but I am not sure this is something that can be modded.

It probably can be, but would be fairly tough to give people the choice.  It usually won't make that much difference in terms of drone riots (for a large base, if you've got enough psych that the facilities get rid of all your drones and have some left over, you've probably got plenty of spare talents as well), so it really just makes it very hard to get a golden age with pacifism drones...and I'm not so sure that's a bad thing.
Title: Re: SE choices for AI - suggestions
Post by: ete on February 11, 2013, 12:56:08 PM
Should pacifism drones come after regular facilities?  It's a pretty strong penalty if (nearly) every military unit is causing a drone, and very strong for two.  If you have unused calming potential, why shouldn't that be applied?
If they were applied before facilities, pacifism drones would be almost no penalty, and Free Market (already extremely useful) would lose its largest downside: Not being able to fight wars. I often get to the stage where all or almost all of my citizens start off as drones, and still suppress them with facilities/SPs/Psych. It'd be a significant gameplay change improving the power of already the best social engineering choice.. I'm not keen on it.

Plus, changing it does not really make sense thematically. People generally bored and unhappy (normal drones) can be distracted easily enough by Rec commons or hologram theaters, but angry anti-war protesters (pacifism drones) are much harder to deal with.
Title: Re: SE choices for AI - suggestions
Post by: Lord Avalon on February 11, 2013, 04:55:30 PM

Actually, Lord Avalon, what you mean to say is that Pacifism drones should be assessed first, then all drone management facilities, projects, etc. should apply.

I agree, but I am not sure this is something that can be modded.

No, I meant what I asked, which was to open debate about changing the game to what you said vs status quo.

Would it really be "almost no penalty," like ete says?  You can get Free Market pretty early, when bases aren't that large and have Rec Commons, or maybe RC + Hologram Theater.  How well can they deal with extra drones, even if pacifism drones are assessed with other drones?  (Off the top of my head, I don't know when the 1st pop drone turns up, and the 2nd.)

What if all drones were assessed first, but pacifism drones are super drones?
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on February 11, 2013, 05:12:31 PM
No, I meant what I asked, which was to open debate about changing the game to what you said vs status quo.

Would it really be "almost no penalty," like ete says?  You can get Free Market pretty early, when bases aren't that large and have Rec Commons, or maybe RC + Hologram Theater.  How well can they deal with extra drones, even if pacifism drones are assessed with other drones?  (Off the top of my head, I don't know when the 1st pop drone turns up, and the 2nd.)

The problem is that total drones before psych and facilities (except from bureaucracy) are capped at the number of nonspecialists at your base.  So on Transcend, 10 pacifism drones and playing as University is only 1 drone more than with no pacifism and not as the University.  (And if you have a specialist, there's no difference at all.)

Quote
What if all drones were assessed first, but pacifism drones are super drones?

Now that seems it'd work better (even if pacifism only makes superdrones after there are no more non-drone non-specialists left); I'm planning part of my next patch to make it possible to do that sort of thing in terms of the superdrones, so it might be possible to also include an order switch...once you do that, though, what difference does it make what order it's in?  The only difference I can see is that pacifism makes it much harder to get a golden age...which still seems like a good thing.
Title: Re: SE choices for AI - suggestions
Post by: ete on February 11, 2013, 05:21:40 PM
How many drones you start off with depends mostly on your number of bases (more=more starting drones) and efficiency rating (higher=less starting drones) via this: http://alphacentauri2.info/mediawiki/Bureaucracy (http://alphacentauri2.info/mediawiki/Bureaucracy) but there are other factors: http://alphacentauri2.info/mediawiki/Drones_%28Advanced%29 (http://alphacentauri2.info/mediawiki/Drones_%28Advanced%29) (i love having a wiki, will make it better though, plenty to fix)

As I understand it, superdrones are only harder to suppress via psych, facilities turn them into citizens as easily as normal drones. Base when "bases aren't that large" with the anti-drone facilities should be able to handle an all citizens start as drones scenario with moderate psych investment, especially if you get an anti drone SP or two. It does mean you've gotta build a couple of facilities and direct a bit into psych, but compared to the current "base is going to massively riot if you try to fight/explore under FM" it seems likely to be very minor.
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on February 11, 2013, 08:28:13 PM
I think one of the problems with pacifism is that Aircraft that are not intercepters cause the penalty if if they have never moved a single square from the city where they were built.  So that makes it very difficult to create aircraft even for purely defensive purposes.

With pacificsm drones coming as late as they do, almost nothing can help against them.  So people are forced into the tainted city strategy, even if they would rather avoid it.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on February 11, 2013, 08:35:37 PM
As I understand it, superdrones are only harder to suppress via psych, facilities turn them into citizens as easily as normal drones.

Even so, when everybody's a drone a large base will depend heavily on psych, so it still definitely matters.  If you have drones and superdrones, psych will turn superdrones into drones before turning drones into workers and then talents, so unless you can deal with your drone problems using only facilities and projects (no psych support at all), you're going to have to fully suppress your superdrones.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on February 11, 2013, 08:47:58 PM
Interesting note I just found: Pacifism drones are capped at 5 drones per base.
Title: Re: SE choices for AI - suggestions
Post by: ete on February 11, 2013, 08:58:06 PM
Earthmichel: Yes, limiting aircraft is one of the biggest disadvantages of FM. However: Do you find FM to be underpowered? Do you find it to be the SE setting you spend much of the game in? If your answers were no and yes respectively, then why would you want to remove its biggest penalty?

Players have found tricks to bypass the penalty, and if the game was to be changed (which is probably a bad idea for this mod even if kryub was not focusing on AI, because forking an already small playerbase..), I'd argue in favor of either limiting or removing those tricks rather than just removing the penalty.

Yitzi: It may force you to up your psych in some situations if done that way. Still a vastly smaller penalty than the current and entirely justifiable system. But again: Why would we want to alter the game to buff arguably the best Economic setting, and one of the best SE settings overall?
Title: Re: SE choices for AI - suggestions
Post by: kyrub on February 11, 2013, 09:00:16 PM
I agree with ete, it's fair that PDs are so hard to control, because FMs is too good.

Although I would love it if there was a special way to go round it. Something like a secret forces rule:
"A cloaked unit is not counted as "a unit outside of territory" and it does not create pacifism drones".
Or " [... same as above ...] unless it attacks".

That would be splendid, thematically.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on February 11, 2013, 09:43:25 PM
Yitzi: It may force you to up your psych in some situations if done that way. Still a vastly smaller penalty than the current and entirely justifiable system. But again: Why would we want to alter the game to buff arguably the best Economic setting, and one of the best SE settings overall?

More than just some situations; the net result would be that any base that can control its drones without using psych (in the midgame, this is (leaving out University and bureaucracy) up to size 5 plus the minimum of either {number of specialists} or {difficulty level from the bottom}.)  So unless you're using a lot of specialists to cut down on drones or dealing with small bases, superdrones are effectively twice the significance of the regular sort.

That said, you are right about not needing to buff FM; if anything, it needs a nerf (probably of the ecological variety).
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on February 11, 2013, 10:01:51 PM
I think that people are grossly underestimating the difficulty in playing FM.  The penalties are very difficult to manage, especially at the early stages and when at war.

Just the fact that Police cannot be used at all for drone control hurts the early game, and the fact that the penalty is so severe that even using Police State to offset accomplishes very little.  Exploration is much more difficult, due to pacificsm penalties.

The highly negative Planet rating means that after the initial "half-strength" period, mindworms are a major threat!  While a non-FM faction can literally farm mind worms, given the 1.5x attacker rating, a FM faction has less than a 50% chance of success, especially if Wealth is also used.  So a FM faction has to research Empath capability pretty early to have a chance to fend off mindworms.

Furthermore, the negative Planet rating means that any overproduction of minerals at a base does FAR MORE ECODAMAGE, and thus the chance of mind worm pops are far more likely.  (Coupled with the fact that mindworms are far harder to deal with.)

Think about it:  such highly negative penalities on Police and Planet, just to gain a +2 bonus!  At first glance, one would wonder why anyone would possibly make such a tradeoff.  Furthermore, if you take FM, you have to give up Planned, which is the most amazing government for the early game.

Anyone who thinks FM is easy, just play a few games yourself on Transcendent on the normal map of Planet, and see how you deal with it.

One Dilbert comics strip showed the pointy-haired boss thinking that anything he does not understand must be easy, so he assign tasks ard priorities accordingly, giving Dilbert one week to upgrade the entire corporate PC base to a new OS.  It makes me wonder if the people who think FM penalties are easy to deal with have ever played a transcendant game with FM themselves?
Title: Re: SE choices for AI - suggestions
Post by: ete on February 11, 2013, 10:38:54 PM
I think that people are grossly underestimating the difficulty in playing FM.  The penalties are very difficult to manage, especially at the early stages and when at war.

Just the fact that Police cannot be used at all for drone control hurts the early game, and the fact that the penalty is so severe that even using Police State to offset accomplishes very little.  Exploration is much more difficult, due to pacificsm penalties.
Yes, FM makes it extremely hard to go to war, which is pretty much the only reason you're ever going to leave FM. And very early before your facilities are up too, though that's fairly short and Planned's Growth/Industry bonuses are very appealing then.

The highly negative Planet rating means that after the initial "half-strength" period, mindworms are a major threat!  While a non-FM faction can literally farm mind worms, given the 1.5x attacker rating, a FM faction has less than a 50% chance of success, especially if Wealth is also used.  So a FM faction has to research Empath capability pretty early to have a chance to fend off mindworms.

Furthermore, the negative Planet rating means that any overproduction of minerals at a base does FAR MORE ECODAMAGE, and thus the chance of mind worm pops are far more likely.  (Coupled with the fact that mindworms are far harder to deal with.)
Yes, if badly managed or fighting a native faction, FM can give problems. However, with care and adequate management, worms are entirely dealable with, you've just got to remain vigilant and accept that exploring fungus is going to be very difficult.

Think about it:  such highly negative penalities on Police and Planet, just to gain a +2 bonus!  At first glance, one would wonder why anyone would possibly make such a tradeoff.  Furthermore, if you take FM, you have to give up Planned, which is the most amazing government for the early game.

Anyone who thinks FM is easy, just play a few games yourself on Transcendent on the normal map of Planet, and see how you deal with it.

One Dilbert comics strip showed the pointy-haired boss thinking that anything he does not understand must be easy, so he assign tasks ard priorities accordingly, giving Dilbert one week to upgrade the entire corporate PC base to a new OS.  It makes me wonder if the people who think FM penalties are easy to deal with have ever played a transcendant game with FM themselves?
I'm not talking from theory. I may not MP, but all my games have been Transcend for many years, and recently the AIs have mostly been given extremely overpowered factions just to try and make them feel like some kind of threat for part of the game. Once I discovered the power of FM (it was admittedly some time before I gave it a good try because at a glance it looks really bad), I've used FM almost exclusively in my games (with exceptions: very early, when I'm fighting a war without using rehoming tricks which seem cheap vs AI, when popbooming, when I'm fighting a heavy native faction like the bards, and very late game when I feel like reducing micromanagement of anti drone stuff). You know full well that FM is extremely worthwhile in its current form, and that its penalties, while severe, are more than counteracted by the massive boost in energy and research. Not easy to use by any means, but very far from needing a buff.

Edit: To clarify. My original statement:
Quote
If they were applied before facilities, pacifism drones would be almost no penalty
I should have said that this was only compared to the current system, but I did not say FM as a whole would have almost no penalty. -3 Planet is significant, though no where near as significant as the magic +1 energy per square.

My view distilled: Switching the order drones are applied in would dramatically reduce the costs associated with FM, and the costs are already not enough to outweigh the benefits for large stages of most games, so attempting to reduce those costs seems strange.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on February 11, 2013, 11:30:02 PM
I think that people are grossly underestimating the difficulty in playing FM.  The penalties are very difficult to manage, especially at the early stages and when at war.

I see how the drone-related FM penalties are very difficult to manage until you're ready (and able) to just keep to your territory and tech, but how is the PLANET penalty that big a problem?

Quote
The highly negative Planet rating means that after the initial "half-strength" period, mindworms are a major threat!

Not if you use empath units.

Quote
While a non-FM faction can literally farm mind worms, given the 1.5x attacker rating, a FM faction has less than a 50% chance of success, especially if Wealth is also used.

Unless you're using command centers, the downside of Wealth is completely negated by having a creche in the unit's home base.  (This is probably a bug).  With equal morale on both sides and no other modifiers, anyone except the Believers actually has a slight advantage when attacking mind worms under FM; 3X.7 is still more than 2.

Quote
So a FM faction has to research Empath capability pretty early to have a chance to fend off mindworms.

True.  However, it's not such a big extra step; if you're getting Secrets of the Human Brain anyway it's only 1 extra tech, otherwise it's 2.

Quote
Furthermore, the negative Planet rating means that any overproduction of minerals at a base does FAR MORE ECODAMAGE

Sort of.  Let's take a "typical" early midgame faction, say 20 techs without a centauri preserve or 40 techs with one.  For simplicity, we'll say that his terraforming is not relevant here; say he has mostly forests so they cancel out his other stuff.  Also assume you're on Transcend, with normal native life and not at perihilion.  Let's also assume he's already had 2 pops and built 10 relevant facilities after the first pop.  Now let's look at some possibilities:
-One guy is producing 28 or fewer minerals.  No ecodamage under any economic choice.
-One guy is producing 30 minerals.  He produces 2X20X5X2/300=1.33 rounded to 2 eco-damage under Green, 4 eco-damage under Simple or Planned, and 8 eco-damage under Free Market.
-Another guy is producing 35 minerals.  He produces 5 eco-damage under Green, 14 eco-damage under Simple or Planned, and 28 eco-damage under Free Market.
In other words, the difference between 30 and 35 minerals at that point is more than the difference between Free Market and Planned; the difference in minerals between "Green has no more than 5% ecodamage chance per base" and "Free Market has no more than 5% ecodamage chance per base" is around 6.  Not such a huge difference.  Between Free Market and Planned is even less.

So yes, Free Market forces you to keep your minerals down to avoid worm rape.  But not running Free Market isn't much better.  Changing that is really the only nerf that Free Market needs.

Quote
Think about it:  such highly negative penalities on Police and Planet, just to gain a +2 bonus!  At first glance, one would wonder why anyone would possibly make such a tradeoff.

Because +1 energy per square is a huge boost, and if you run wealth as well you get a big boost to commerce.

Quote
Furthermore, if you take FM, you have to give up Planned, which is the most amazing government for the early game.

For the early game, that does make sense (though if you're going full-on energy focus, FM with occasional use of Planned to pop-boom is actually better), but what about after that?

Quote
Anyone who thinks FM is easy, just play a few games yourself on Transcendent on the normal map of Planet, and see how you deal with it.

Oh, it's not easy, no question.  But the ecodamage effect, which is supposed to be one of its downsides, really isn't because everybody has to carefully watch their minerals as compared to the cap.  That's the only change that it needs.

In games when you're at peace (which will mainly be many-player multiplayer games or lower-difficulty vs. AI; HtH and Transcend are not peace-conducive) and have collected all the pods you're interested in, it seems to me that FM is clearly the way to go; the extra energy will boost your tech more than enough to justify having to grab a couple of Centauri techs, and rushing-and-selling Centauri preserves will cancel out the ecodamage chance penalty fairly quickly.

[quoteIt makes me wonder if the people who think FM penalties are easy to deal with have ever played a transcendant game with FM themselves?[/quote]

Transcend is unusual due to the whole homicidal-AI thing; a game more conducive to peace (say at Talent but with souped-up AI factions for the difficulty boost) would be a better comparison.  It'd still be difficult to do early, but would become substantially easier later on once you can use a punishment sphere base and rush-and-sell centauri preserves to cut ecodamage.

So no, I don't think Free Market is easy (I have tried it myself, and it's hard to do early), but it still needs a nerf in terms of making the alternatives not almost as bad as it is in terms of having to limit minerals to control ecodamage.

I'm not talking from theory. I may not MP, but all my games have been Transcend for many years, and recently the AIs have mostly been given extremely overpowered factions just to try and make them feel like some kind of threat for part of the game. Once I discovered the power of FM (it was admittedly some time before I gave it a good try because at a glance it looks really bad), I've used FM almost exclusively in my games (with exceptions: very early, when I'm fighting a war without using rehoming tricks which seem cheap vs AI, when popbooming, when I'm fighting a heavy native faction like the bards, and very late game when I feel like reducing micromanagement of anti drone stuff). You know full well that FM is extremely worthwhile in its current form, and that its penalties, while severe, are more than counteracted by the massive boost in energy and research. Not easy to use by any means, but very far from needing a buff.

Indeed; I'd say it needs no buffs and needs no nerfs except in the difference between controlling ecodamage under FM and controlling ecodamage under other economies.  I think the best approach is to make it easy to control ecodamage under Green, moderately difficult under Planned or Simple, and under Free Market, not exactly hard, more like requiring abandoning any hope of high-production bases.  (As opposed to now, where it's moderately difficult under Green, a tiny bit harder under Planned or Simple, and a tiny bit harder under Free Market, unless you build-and-sell Centauri Preserves in which case it's easiest under Free Market once you hit the midgame because you can rush buy easily.)
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on February 12, 2013, 02:11:57 AM
Actually, the thing I would like to change is that Centauri Preserves, Tree Farms, etc. keep their industry immunity after they have been built and sold.  If you sell it, the bonus should go away.  This would perhaps be the penalty you are looking for with Free Market, if you could not build and sell Centauri Preserves to build mineral immunity.

I wonder if it is possible to mod that, so that if you sell the building, the mineral immunity bonus goes away?
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on February 12, 2013, 02:45:40 AM
Actually, the thing I would like to change is that Centauri Preserves, Tree Farms, etc. keep their industry immunity after they have been built and sold.  If you sell it, the bonus should go away.  This would perhaps be the penalty you are looking for with Free Market, if you could not build and sell Centauri Preserves to build mineral immunity.

I wonder if it is possible to mod that, so that if you sell the building, the mineral immunity bonus goes away?

It would be possible, but tricky.  Furthermore, even if selling it makes it go away, it's sort of absurd that those facilities, if built in a single base, increase the mineral cap for the entire faction.  So maybe it'd be better just to remove that increase, and provide an across-the-board decrease to ecodamage to compensate.
Even so, however, ICS remains a very good way to avoid the cap, since the free 16 minerals still per-base (and a mineral-focused base doesn't need all the multiplier facilities, since most of them are energy-centered).  So I figure that it would also be advisable to remove all the "clean minerals" (but let pops reduce the effect of Planet Busters and tectonic missiles*), and reduce ecodamage a lot (something like a 10X decrease) to compensate.  Maybe also cut the global warming rate some.

*While we're at it, we might as well get rid of the stupid rule where using nerve gas increases ecodamage after the whatever-number time.
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on February 12, 2013, 04:57:03 AM
I don't think it is at all silly for these enhancements to raise the clean mineral limit.  They each reduce pollution and industrial impact overall.  And they each have a maintence cost that has to be paid for the privilige (as long as you are not allowed to sell them and keep the benefit.

Actually, it makes less sense to me that a fungal pop would increase the clean mineral limit, but that is the way the game is designed.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on February 12, 2013, 03:44:20 PM
I don't think it is at all silly for these enhancements to raise the clean mineral limit.  They each reduce pollution and industrial impact overall.  And they each have a maintence cost that has to be paid for the privilige (as long as you are not allowed to sell them and keep the benefit.

That maintenance cost is well paid for by the Economy boost of tree farms and hybrid forests, and even for Centauri Preserves is fairly justified by the ability to effectively halve ecodamage in the base where they're built.  But to have a facility you can build in every base which gives faction-wide bonuses for each one built...that's just ridiculous.

Quote
Actually, it makes less sense to me that a fungal pop would increase the clean mineral limit

I think that's meant to make it so that going over by a bit won't lead to a lot of pops; even so, I feel there are better ways to do that.
Title: Re: SE choices for AI - suggestions
Post by: ete on February 12, 2013, 04:26:24 PM
Actually, the thing I would like to change is that Centauri Preserves, Tree Farms, etc. keep their industry immunity after they have been built and sold.  If you sell it, the bonus should go away.  This would perhaps be the penalty you are looking for with Free Market, if you could not build and sell Centauri Preserves to build mineral immunity.

I wonder if it is possible to mod that, so that if you sell the building, the mineral immunity bonus goes away?
Agreed. This would be my no. 1 change to eco damage, and a bugfix. Making the eco damage formula count the number of current eco facilities rather than the number you've build would make much more sense and close a major unintended loophole.

A major redesign of eco damage as yitzi suggests would be interesting, but should be separate from main SMAX imo, and ideally be a separate patch (i.e. yitzi's bugfixes integrated with kryub's AI for the unofficial standard patch, yitzi's redesign of eco damage and various other things in a separate package) for compatibility reasons unless the all old maps/saves/scenarios need hex editing to work issue can be overcome. With the facility sell/rebuild issue solved.. the current eco setup would be okay, though it does take too little account of Planet.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on February 12, 2013, 05:53:05 PM
I didn't think of the idea of having it just be based on facilities currently present but still giving 1 per facility (mainly because it still seems flawed to me); keep it in mind for when I start asking for ideas for more things to mod, which will be right after I remove the "needs hex editing to use old maps" issue (I checked it out, and it can be done, though it'll take quite a bit of work; I figure that will be what defines patch version 2.0.)
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on February 12, 2013, 08:29:57 PM
That was what I was proposing  (although ete said it more clearly).

To me, being able to build and destroy Centauri Preserves (and similiar facilities) to gain an extra clean mineral each time is clearly a bug, and a bug fix needs to be done.

Yitzi, I think you will find that if the player can't just create and destroy Centauri Preserves (the cheapest facility of this kind) over and over to increase the mineral limit, THEN the Planet penalty and ecodamage becomes much more significant. 

That is they way I play anyway; I never create and destroy these facilities, because I consider it exploiting a bug, even though most game rules allow it.  So perhaps I feel the pain more than most, because I guess most people have no qualms about exploiting this.

The player will be limited by the number of bases and the number of kinds of facilities that he can currently build at his tech level.  Being able to build a Temple of Planet becomes useful, instead of a big yawn to those who know they can achieve the same result much more cheaply by creating and destroying a few more Centauri Preserves.

If you don't think this is enough, then you can always do whatever you want with respect to modding to change this.  But I think the key distinction is the bug fix (so that players can no longer exploit destroyed facilities still having their effect) verses an actual change to the game.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on February 12, 2013, 09:47:53 PM
Yitzi, I think you will find that if the player can't just create and destroy Centauri Preserves (the cheapest facility of this kind) over and over to increase the mineral limit, THEN the Planet penalty and ecodamage becomes much more significant. 

Somewhat, but not that much; he can still build a bunch of ICS-style bases and stick Centauri Preserves (and probably tree farms as well) in them; while he does have to pay maintenance for the preserves, it's not all that much when divided among all his producing bases.

And even if not...if he's got 12 bases (a fair midgame number) and a tree farm and hybrid forest and centauri preserve at each one, he can produce 50 minerals per base with no ecological damage, no matter what economics he's running.

But I'm perfectly willing to include the other version once I reach the stage where I finish the current batch and then change it so the new variables can be applied to existing maps without a hex editor (as I will be making the patch such that you can use either version).
Title: Re: SE choices for AI - suggestions
Post by: ete on February 12, 2013, 11:41:46 PM
I think the clean minerals being globally raised to fairly high levels in mid/late game was a design feature to combat ICS, which would otherwise be strongly encouraged once bases were hitting their clean production limits. Unless either all bases produce eco damage (perhaps smaller ones more) or large ones can be made to not, it's going to encourage ICS. And doing the former.. well, it'll be interesting to see how it turns out. Perhaps it'll work very well.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on February 12, 2013, 11:55:14 PM
I think the clean minerals being globally raised to fairly high levels in mid/late game was a design feature to combat ICS, which would otherwise be strongly encouraged once bases were hitting their clean production limits.

The problem is that the global raising happens via methods that are most effective when using ICS.

Quote
Unless either all bases produce eco damage (perhaps smaller ones more) or large ones can be made to not, it's going to encourage ICS. And doing the former.. well, it'll be interesting to see how it turns out. Perhaps it'll work very well.

The former is possible (though not doing it based on size) with my patch.  However, here's another idea that would avoid encouraging ICS: Remove clean minerals entirely.  No clean minerals means no large freebie that's multiplied by your number of bases (you'd still want a small freebie in the form of rounding down or similar, just so that you're not facing ecodamage almost as soon as you start), so there's no ecodamage benefit to ICS.  200 minerals faction-wide would be the same whether it's spread among 10 bases or 50.  (You would of course have to cut ecodamage substantially to compensate, but that's not necessarily a bad thing.)
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 14, 2013, 05:09:32 PM
Quote
The former is possible (though not doing it based on size) with my patch.  However, here's another idea that would avoid encouraging ICS: Remove clean minerals entirely.  No clean minerals means no large freebie that's multiplied by your number of bases (you'd still want a small freebie in the form of rounding down or similar, just so that you're not facing ecodamage almost as soon as you start), so there's no ecodamage benefit to ICS.  200 minerals faction-wide would be the same whether it's spread among 10 bases or 50.  (You would of course have to cut ecodamage substantially to compensate, but that's not necessarily a bad thing.)

I would say that pollution should be on a per-base basis.  Logically only bases that pollute would have fungal pops.  Also, this adds some strategy in having mineral cities with eco facilities to offset.  The problem with eco facilities as they are is that they benefit all your bases by raising the clean mineral limit (a pro-ICS mechanic).

I would tie eco dmg to these factors only:
- Population size of the base
- Eco facilities (ideally, I think only the preserves.  Tree farm and Hybrid Forest are strong enough).
- Planet rating

How about this?

ECODMG =MAX(ECO_RATE*(MINERALS/BASESIZE^BASE_RATE)*(1-(PLANET*PLANET_RATE))-BASE_CLEAN_MINERALS),0)

MINERALS: minerals produced by the base
BASE_RATE: a small exponent to slightly benefit larger bases in terms of eco dmg.  I think this will be needed in the late game where you will have much higher mineral totals from Genejack, Replicator, etc.  May take some playing around but I would put this at a value of ~1.1.
PLANET: City owner's PLANET rating in SE
PLANET_RATE = 12.5% (for 12.5% more/less pollution per planet rating) seems reasonable.  FM would thus pollute quite a bit more than Green, as this is calculated before clean minerals are subtracted.
BASE_CLEAN_MINERALS: the base clean minerals for each individual base.  I think putting this around 8 would work, if you're looking to shut down size 2-3 bases each with a crawled borehole.  Though this also depends on if you keep the existing road/improvements as counting against it (I think).
ECO_RATE: a constant that can be also modified by difficulty rating.  Somewhere around 1 for lowest difficulty and 5 for Transcend.

Then you can have the preserves give +1 PLANET or some such, if that's possible.   Or they could increase BASE_CLEAN_MINERALS if that is not.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 14, 2013, 06:04:07 PM
I would say that pollution should be on a per-base basis.  Logically only bases that pollute would have fungal pops.  Also, this adds some strategy in having mineral cities with eco facilities to offset.  The problem with eco facilities as they are is that they benefit all your bases by raising the clean mineral limit (a pro-ICS mechanic).

Well, they also benefit the base in which they are built, but I'd agree with having each facility only affect the base where it's built (projects should affect the whole faction, though.)

Quote
I would tie eco dmg to these factors only:
- Population size of the base
- Eco facilities (ideally, I think only the preserves.  Tree farm and Hybrid Forest are strong enough).
- Planet rating

Having it based on base size rather than mineral production (and removing the effects of terraforming entirely) is an interesting idea; it's not to my taste, but once I start taking modding requests I should be able to add that option.

Making it not depend on techs is also interesting and would greatly change the game effects; not for the better, though, IMO.  I like the idea of ecodamage becoming more of a concern later in the game.

Quote
ECODMG =MAX(ECO_RATE*(MINERALS/BASESIZE^BASE_RATE)*(1-(PLANET*PLANET_RATE))-BASE_CLEAN_MINERALS),0)

MINERALS: minerals produced by the base
BASE_RATE: a small exponent to slightly benefit larger bases in terms of eco dmg.  I think this will be needed in the late game where you will have much higher mineral totals from Genejack, Replicator, etc.  May take some playing around but I would put this at a value of ~1.1.

Interesting idea, but fractional exponents are fairly hard to pull off, and as written larger bases would actually produce less ecodamage (as basesize will grow as fast as minerals).  As for being needed, I think that removing clean minerals would allow for an across-the-board 90% slash to ecodamage, which will make higher mineral totals a lot easier to handle, assuming you build stuff like preserves/temples/nanoreplicators and get a higher PLANET rating in the late game.

Actually, with baserate>1 and BASE_CLEAN_MINERALS=8, you'd almost never get any ecodamage, as you'd need to have more than 8 minerals per population, which is unlikely at any stage of the game.

Quote
PLANET: City owner's PLANET rating in SE
PLANET_RATE = 12.5% (for 12.5% more/less pollution per planet rating) seems reasonable.  FM would thus pollute quite a bit more than Green, as this is calculated before clean minerals are subtracted.

Could work, though I think a 20% would work better with no clean minerals.

Overall, your ideas are interesting, but I'm not sure that they'd be any better than the following (far easier to implement) formula:

ECODMG=MAX(0, [(MINERALS/(5X(1+FACILITIES))+(TERRAFORMINGX(1-(FORESTFACILITIES/2)))]XTECHSX(1-PLANET/5)/20)XMAPMODIFIERXDIFFMODIFIER

FACILITES=All facilities in the base, and projects in the faction, that are marked as "reduce the effects of industry on Planet's ecology)
TERRAFORMING would be calculated as before.
FORESTFACILITIES are tree farm and hybrid forest.
MAPMODIFIER is 1.5 for heavy native life, 1 for normal, 0.5 for light native life.
DIFFMODIFIER is 1 on Thinker and Transcend, 0.6 for Librarian and lower.

Thus, even in the lategame with 300 minerals in a base, you have at least 3 facilities (nanoreplicator, preserve, temple) cutting it down to 15XTECHSX(1-PLANET/5)/20)XMAPMODIFIERXDIFFMODIFIER.  So with 86 techs on a normal map on Transcend running Green, that's 38.7...uncomfortable*, but not unreasonable for an endgame high-mineral base.

*As it should be; as the game progresses, the native life should become a very significant factor.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 14, 2013, 08:25:48 PM
Quote

Actually, with baserate>1 and BASE_CLEAN_MINERALS=8, you'd almost never get any ecodamage, as you'd need to have more than 8 minerals per population, which is unlikely at any stage of the game.

Looking back, it's not really 'clean minerals per base' that I was subtracting.  It's a flat 8 off of eco damage at the end and is not multiplied by population.  In effect this works like a clean mineral limit but it really is not.

Your formula fits the game descriptions of how the base facilities work, and it's close to how it is now.  I do agree that eco damage should be tied to tech (as Planet approaches its Transcendence). 

The problem with eliminating the clean mineral limit entirely is that the game is balanced around running 0 eco damage for most players, most of the time.  Even tiny bases will be triggering fungal pops in the early game.  Now this is probably how it should be, but with how the boils progress they will quickly get out of hand.  Even running Green and all facilities will not really be enough.  As you showed, you will be creating 39 eco damage for every 300 minerals produced, and that's late game when all facilities are even available and built to cut down pollution.  If you're running FM and no facilities it's ~413 eco damage for the same 300 minerals.  Facilities should be strong, but I think that they have way too much an impact if clean minerals are taken out.  They'd be more or less a requirement and late game captured bases would be pretty much just a burden by the pollution they make.

For example, Green in early-mid game (25 techs) with no facilities (nano and temple are very late game, and preserve is mid-game), is 60*TECH/20*0.6, or 300 mins -> 45 eco damage (across any number of bases).  With FM on it's 120 eco per 300 minerals.  This is a lot higher than current eco damage numbers.  Add into this that the AI will now be polluting as well (assuming you don't code them to just cheat around it).

I think the real issue is that the boils progress very sharply.  They are balanced around a very all-or-nothing model of eco damage (which means surpassing clean minerals but not till the very late game typically).  What if the boil frequency was triggered by Eco damage, but the size/power was tied to tech rather than the number of pops?  And then getting rid of native mass captures and free energy for stack killing.  As it stands eco damage can really be exploited as a good thing, when it should only ever be bad.
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on April 14, 2013, 08:30:23 PM
Quote
I would say that pollution should be on a per-base basis.  Logically only bases that pollute would have fungal pops.  Also, this adds some strategy in having mineral cities with eco facilities to offset.  The problem with eco facilities as they are is that they benefit all your bases by raising the clean mineral limit (a pro-ICS mechanic).

This is not a pro-ICS mechanic, but just the opposite.

With ICS, any particular base produces only 1/4 - 1/6 or less of what would be produced in a non-ICS scenario.  When you do not allow eco facilities to have a general effect, then this actually promotes ICS, so that you can keep the production of any city smaller.

Non-ICS requires the current mechanism to be competitive with ICS.  If you get rid of the current mechanism, you force players away from large cities, and toward ICS to minimize the production at any particular city.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 14, 2013, 08:44:37 PM
Quote
I would say that pollution should be on a per-base basis.  Logically only bases that pollute would have fungal pops.  Also, this adds some strategy in having mineral cities with eco facilities to offset.  The problem with eco facilities as they are is that they benefit all your bases by raising the clean mineral limit (a pro-ICS mechanic).

This is not a pro-ICS mechanic, but just the opposite.

With ICS, any particular base produces only 1/4 - 1/6 or less of what would be produced in a non-ICS scenario.  When you do not allow eco facilities to have a general effect, then this actually promotes ICS, so that you can keep the production of any city smaller.

Non-ICS requires the current mechanism to be competitive with ICS.  If you get rid of the current mechanism, you force players away from large cities, and toward ICS to minimize the production at any particular city.

The eco damage subtractor is not needed.  You could get rid of it if desired, and just make the whole pollution linear to minerals produced.  Which is also logical.  The base exponent would also make larger bases pollute less than an equal number of smaller bases.  For example, a size 10 base making 64 minerals would pollute as though it was 10^1.2 -> ~10 size 1-bases each making 4 minerals.  Or in other words, the size 16 city would pollute at (10/10^1.2) or ~63% the rate of all the small cities combined.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 15, 2013, 02:08:03 AM
Looking back, it's not really 'clean minerals per base' that I was subtracting.  It's a flat 8 off of eco damage at the end and is not multiplied by population.  In effect this works like a clean mineral limit but it really is not.

I got that; what I missed is that it's subtracted after multiplying by ECO_RATE.  Although in that case, it still gets wonky in that the larger a base grows, the smaller its ecodamage is.

Quote
Your formula fits the game descriptions of how the base facilities work, and it's close to how it is now.

Which I think is an advantage, both in terms of coding and balancing.  The main differences are that it has no clean minerals, and that there's an across-the-board ecodamage cut.

Quote
The problem with eliminating the clean mineral limit entirely is that the game is balanced around running 0 eco damage for most players, most of the time.  Even tiny bases will be triggering fungal pops in the early game.

That is a concern if it happens with small enough bases, but subtracting even just 1 from the ecodamage chance, together with an across-the-board proportional cut will prevent ecodamage from showing up too early.

Quote
Now this is probably how it should be, but with how the boils progress they will quickly get out of hand.  Even running Green and all facilities will not really be enough.  As you showed, you will be creating 39 eco damage for every 300 minerals produced, and that's late game when all facilities are even available and built to cut down pollution.

Yeah...but in late game where you're producing 300 minerals in a base, you should be able to afford the empath units and super formers to keep it under control anyway.  And 300 minerals is a lot even in the late game unless a single base is crawling far more than its share.

Quote
If you're running FM and no facilities it's ~413 eco damage for the same 300 minerals.

Firstly, there's no real way to get that; no facilities means no nanoreplicator, and that cuts down your mineral potential.
Secondly, even with facilities, 300 minerals is assuming you've got a lot of forests and/or fungus; if you want to run FM in the late game, you're probably going to prefer farm/solar which is a lot lower in mineral production (especially when you consider that minerals directly from Nessus stations don't count).

Quote
Facilities should be strong, but I think that they have way too much an impact if clean minerals are taken out.  They'd be more or less a requirement and late game captured bases would be pretty much just a burden by the pollution they make.

Keep in mind that no facilities also means no mineral-multiplying facilities, which means you're looking at a maximum mineral production of 50 or less, which will compensate for the lack of ecological facilities.

Quote
For example, Green in early-mid game (25 techs) with no facilities (nano and temple are very late game, and preserve is mid-game), is 60*TECH/20*0.6, or 300 mins -> 45 eco damage (across any number of bases).  With FM on it's 120 eco per 300 minerals.  This is a lot higher than current eco damage numbers.

Actually, it depends how many bases you're looking at.  Under current rules with 25 techs and +2 PLANET, 45 eco-damage (assuming normal native life, transcend difficulty) is 45X300/2/25/5=54 minerals over clean minerals.  (That is, with 54 non-clean minerals you'll be looking at 45 eco-damage; 54X2(LIFE)X5(DIFF)X25(TECHS)/300=45.)  Thus, to get 300 minerals with only 45 eco-damage requires that 246 of them be clean minerals; assuming that there have been no fungal pops yet, that requires 16 bases, which isn't going to happen in the early-mid game without ICS.

So no, it really isn't higher than current eco-damage numbers; it's just that in the early-mid game your entire faction usually isn't producing anywhere near 300 minerals.

Quote
I think the real issue is that the boils progress very sharply.  They are balanced around a very all-or-nothing model of eco damage (which means surpassing clean minerals but not till the very late game typically).

You mean how a small increase in minerals produces a large increase in ecodamage?  Part of the point of the formula I described is that that doesn't happen.  Also, the boils don't seem to progress all that sharply (keep in mind, though, that without my patch building ecological facilities after the first pop causes boils to progress as though they were pops, which may be why they progress sharply in your experience.)

Quote
And then getting rid of native mass captures and free energy for stack killing.

I believe native mass captures are already impossible, and I know that ecodamage-caused worms can't be captured at all.  Planetpearls could do with some reduction, though.

This is not a pro-ICS mechanic, but just the opposite.

It really depends on what you compare it to.  Compared to "remove clean minerals for facilities, don't change anything else", the current system is more anti-ICS.  But compared to "remove clean minerals, decrease ecodamage across the board to compensate", the current system is more pro-ICS.

The base exponent would also make larger bases pollute less than an equal number of smaller bases.  For example, a size 10 base making 64 minerals would pollute as though it was 10^1.2 -> ~10 size 1-bases each making 4 minerals.

Your numbers are off.  10^1.2~15, so your 64-mineral base would produce as much ecodamage as one size-1 base with 4 minerals.  Which is pretty absurd.
Title: Re: SE choices for AI - suggestions
Post by: Buster's Uncle on April 15, 2013, 02:46:00 AM
I believe native mass captures are already impossible, and I know that ecodamage-caused worms can't be captured at all.
That turns out not to be the case - late-game, you start getting that "A more powerful mind seems to be controlling these worms", but up to then, I've captured upwards of 20 ecodamage-caused worms at a stroke; I did it just today.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 15, 2013, 03:14:31 AM
I believe native mass captures are already impossible, and I know that ecodamage-caused worms can't be captured at all.
That turns out not to be the case - late-game, you start getting that "A more powerful mind seems to be controlling these worms", but up to then, I've captured upwards of 20 ecodamage-caused worms at a stroke; I did it just today.

Can you link a save of the game, by any chance?
Title: Re: SE choices for AI - suggestions
Post by: Buster's Uncle on April 15, 2013, 03:23:35 AM
I can, but I already used up those units, so no point. 

I've got Voice of Planet now, so I'm getting no more ecodamge worms this game - I'll capture a stack for you next time, though that won't be for several days...

Oh wait!  I've actually got a freshly captured eco-stack, avec screenie, because there's something buggy going on - hold on a minute.
Title: Re: SE choices for AI - suggestions
Post by: Buster's Uncle on April 15, 2013, 03:53:02 AM
Okay, I was already working up screenies for something else, and having to use this old game - the one I used for testing the timewarp loading between games thing you had me do recently.  Coincidentally, in the process, I'd let everything go for several turns and gotten a number of fungus pops, and a few hours ago, captured a NL stack (the second time today, in different games).

I saved this because that city-thing kept popping up among the ecodamage native life - boy, are they tough, too.  It's generally a sign a game is bugging and about to crash, but I see these once in a long time - the last was a bad load of Kil's LE crossover GotM last year.

(http://alphacentauri2.info/MGalleryItem.php?id=554)

So in the save, you want to check 44,42 - not what I'm doing with the active formers on the other side of the map.  For once there's no custom factions to complicate things; it's a really old game.
Title: Re: SE choices for AI - suggestions
Post by: Buster's Uncle on April 15, 2013, 04:13:47 AM
Here's a better view of the city-looking thing - and an uncaptured one stacked with another unit.  Attack it with the active unit, and reload the attached save over until you capture those two.  It shouldn't take all night.  This is how the game normally behaves - it's not capture-able because the game's gone buggy.

(http://alphacentauri2.info/MGalleryItem.php?id=555)

scient wants me to get the game to crash, and I've still not finished the other thing I was working on, and it's late, so...
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 15, 2013, 04:36:32 AM
Re: Boils:

Yea maybe it's different in your patch.  I've seen games where boils progress to 20+ stacks of Elite Locusts.  These stacks could not be captured and took no collateral damage..  It seems like this was after only a dozen pops or so.  They seemed controllable at any amount of minerals because a base can only produce one unit/facility a turn (something else that imo should be changed, but would be very hard to code).

I know I have captured large (10+) stacks of Mindworms in the current 2.0 (GOG) patch.  I'd have to test this more in a game where I try to pollute.  They may have been randomly spawned stacks but I don't believe this was so.  Usually all you see spawned is an Isle of the Deep with 2 mindworms in it, and the occasional roaming mindworm/locust.

Re: Eco facilities:

300 faction minerals is mid game faction-wide production.  Certainly not early-mid.  That's 10-15 bases, around size 7 with 2 crawled boreholes each, and no mineral multipliers. 

I forsee two problems.  Nanoreplicator, Temple, and Pholus Mutagen are very late game.  In addition, facilities reduce eco damage way too much.  As they are, each eco facility reduces eco damage as follows:
#1: -50%
#2: -66%
#3: -75%
#4: -80% (Pholus Mutagen with all 3)

The first facility is like going from 0 to 2.5 Planet alone.  In fact, it's even stronger if you are running FM...taking you from an effective -3 Planet (1.6) up to +1 Planet (0.8).  Now you can argue the absolute benefit of the facilities is the same for all SEs, but this does show their extreme power even against a strong +/- 20% pollution per Planet modifier.  For example going from 0 to 4 facilities with FM on is equivalent to going from -3 Planet to +3.2 Planet.  While I don't think you can actually exceed 3 Planet, this shows that their effect is more than the difference between the most and least polluting SE settings.  In your late-game example (Green and 3 facilities), you're cutting eco damage down by 85%, and it's still quite a high amount to deal with.  While I think Planet rating and facilities should matter, the effects of running FM and/or no facilities will be too devastating to even contemplate these options. 

What is the cap on +Planet?  I assume it's +3, as there's no benefit past that on the SE page.  What if rather than being the divisor for minerals, facilities and Pholus Mutagen each increased Planet rating by +1, up to a max of +3 rating?

---

Yea I butchered the numbers really badly on the city size idea.  My head's been a bit off today.  What I meant was that a base would have its eco damage reduced slightly by a function of its size, which is anti-ICS.

The pollution in a base would scale down by a factor of [N/(N^1.2)], where N is the city size.
size 1 - 100% (1/1^1.2)
size 2 - 87% (2/2^1.2)
size 3 - 80%
...
size 10 - 63%
size 16 - 57%
size 25 - 53%
size 50 - 46%

Therefore a size 50 city making the same minerals as 50 size-1 cities, would pollute at 46% the rate.  Sorry, I made this really unclear.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 15, 2013, 12:17:27 PM
Here's a better view of the city-looking thing - and an uncaptured one stacked with another unit.  Attack it with the active unit, and reload the attached save over until you capture those two.  It shouldn't take all night.  This is how the game normally behaves - it's not capture-able because the game's gone buggy.

(http://alphacentauri2.info/MGalleryItem.php?id=555)

scient wants me to get the game to crash, and I've still not finished the other thing I was working on, and it's late, so...


Ok, thanks.  I'll try to take a look at it, and see if I can easily figure out what's going on or if it needs to wait for when I'm focusing on just that.

Re: Boils:

Yea maybe it's different in your patch.  I've seen games where boils progress to 20+ stacks of Elite Locusts.  These stacks could not be captured and took no collateral damage..  It seems like this was after only a dozen pops or so.


How many ecological facilities were built in that game?  There's a bug (fixed in my patch) where it treats facility-building as pops for certain purposes.

Quote
Re: Eco facilities:

300 faction minerals is mid game faction-wide production.  Certainly not early-mid.  That's 10-15 bases, around size 7 with 2 crawled boreholes each, and no mineral multipliers. 


Exactly; and by that point you should have Centauri Preserves to help keep the ecodamage down.  Or you could just deal with a pop every other turn.

Quote
I forsee two problems.  Nanoreplicator, Temple, and Pholus Mutagen are very late game.  In addition, facilities reduce eco damage way too much.  As they are, each eco facility reduces eco damage as follows:
#1: -50%
#2: -66%
#3: -75%
#4: -80% (Pholus Mutagen with all 3)


Yes, they're effectively essential for high-production bases; I don't see how that is a  problem, seeing as they're equally available to everyone.

Quote
While I think Planet rating and facilities should matter, the effects of running FM and/or no facilities will be too devastating to even contemplate these options. 


Only in the later game when going for high mineral production.  If you ditch the boreholes and most mineral-boosting facilities and build condenser/farm everywhere, you're looking at far less.

And in the later game, requiring ecological facilities if you're getting production facilities as well is IMO not such a big problem.

Quote
What is the cap on +Planet?  I assume it's +3, as there's no benefit past that on the SE page.


+2 for ecodamage actually; my proposal would increase it to +5.

---

Quote
Yea I butchered the numbers really badly on the city size idea.  My head's been a bit off today.  What I meant was that a base would have its eco damage reduced slightly by a function of its size, which is anti-ICS.

The pollution in a base would scale down by a factor of [N/(N^1.2)], where N is the city size.


Ah, so the actual exponent would be 0.2.  That makes more sense...but if so, your proposal would be even harsher than mine in the late game for mineral-focusing factions; a 300-mineral size-50 base would produce 5X(300/50^(0.2))*(3/8)-8=5X137.19X3/8-8=249.23, rounds up to 250, and that's before applying techs.
Title: Re: SE choices for AI - suggestions
Post by: Buster's Uncle on April 15, 2013, 01:52:27 PM
Actually, when I opened that game a tried the capture, NOW I was getting the "vastly more powerful mind controlling" message, at which point captures never seem to happen.  It might be possible to find a stack to capture in the first save I posted which was a turn earlier and when I captured that stack - there had been a lot of fungal pops, and I doubt I found anything like all of the ecodamage worms.

On reflection, I think this was originally a SMAC game, and that's why it's so buggy.  I did succeed in crashing it - by attacking a worm two turns later, with that captured city thing...  scient can tell you more, but it had a reactor strength of zero.  You CAN duplicate the crash effortlessly by moving it up and over from the right in the last save posted, and attack that worm pictured in place of the active unit.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 15, 2013, 02:22:59 PM
Actually, when I opened that game a tried the capture, NOW I was getting the "vastly more powerful mind controlling" message, at which point captures never seem to happen.  It might be possible to find a stack to capture in the first save I posted which was a turn earlier and when I captured that stack - there had been a lot of fungal pops, and I doubt I found anything like all of the ecodamage worms.

On reflection, I think this was originally a SMAC game, and that's why it's so buggy.  I did succeed in crashing it - by attacking a worm two turns later, with that captured city thing...  scient can tell you more, but it had a reactor strength of zero.  You CAN duplicate the crash effortlessly by moving it up and over from the right in the last save posted, and attack that worm pictured in place of the active unit.

So it seems the "capture ecodamage worms" is a bug that has since been fixed anyway, so it's not a concern.
Title: Re: SE choices for AI - suggestions
Post by: Buster's Uncle on April 15, 2013, 02:31:38 PM
I'm not aware of it being fixed, but then the save is a borked game only demonstrable of instability...
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 15, 2013, 06:34:35 PM
I'm not aware of it being fixed, but then the save is a borked game only demonstrable of instability...

What version are you guys playing?  I've seen mass captures happen in my games.

Quote
Ah, so the actual exponent would be 0.2.  That makes more sense...but if so, your proposal would be even harsher than mine in the late game for mineral-focusing factions; a 300-mineral size-50 base would produce 5X(300/50^(0.2))*(3/8)-8=5X137.19X3/8-8=249.23, rounds up to 250, and that's before applying techs.

Yea it's very off, I had an error in the formula.  It was more of an idea to reduce the pollution a bit for non-ICS strategy.  Though with more thought, I think the current formula is very good but just needs some logical tweaks in addition to removing clean minerals (a bad mechanic, as it is an invisible value).

Quote
+2 for ecodamage actually; my proposal would increase it to +5.

With a quick combat test it seems that there is no cap on +Planet.  So the game does not just set the user's Planet to +3 if you are over.

+5 Planet would be 0 pollution at 20% per Planet.  I'm not so sure that's balanced, Gaia and Cult can hit +5 and +6 respectively with Green+Cyber.  Green+Cyber will be pretty much a requirement to get +4 when they are already very strong SE settings due to +Efficiency.  As well Believers/Usurpers would only be able to hit +3, which is double the pollution of +4. 

I would put pollution effects at more like 10% per Planet or cap it at -60% pollution.  Consider that +Planet rating is already good where it gives +10% attack to and with native life, free native life, and easier fungus movement.  It's probably a known bug but the Planet combat modifier should be the differential between attacker and defender rather than just the attacker's +Planet.
 
Here's what I would want:
- Keep eco facilities as is, and a formula similar to the current where it scales by tech.
- Eco damage modified by Planet 
- Boils scaling by tech and/or faction size, rather than linearly out of control.  I will have to test this as well, to see how it scales in the patched version.
- No clean minerals.  Clean minerals just encourages smaller bases and thus ICS.
- Modified values for eco damage from terraforming.  How difficult would it be to change the terraforming portion of eco damage?

Currently terraforming is calculated as follows:
"Terraforming = [(2*# worked (not crawled) improvements other than kelp farms)+(# of unworked improvements) + 8*Boreholes + 6*Echelon Mirrors + 4*Condensors +1 if a Seabase -#of Forests]/8.
Divide by 2 for presence of a Tree Farm and reduce to 0 for presence of a Hybrid Forest."

It seems highly exploitable that crawled tiles do not count as terraforming eco damage.  For example I could create several boreholes outside a city radius and crawl them with 0 terraforming eco damage?  I will have to test this in my next game.

I would argue that all tiles crawled should add for the improvements on the tile.  Further, uncrawled improvements should add at half the rate to the nearest base for eco damage (likely much harder to implement).  For example, you crawl a mine with a road, that counts as both a mine and road.  An uncrawled borehole outside a city range adds eco damage to that city at half the rate.

I think Boreholes should be much more eco damaging and maybe a little quicker to terraform.  If a Condensor is 4 and a Mirror is 6, a Borehole should be like a 20.  Boreholes seem to be a staple of early game ICS.  What if Boreholes also only got -25% pollution from Tree / Hybrid?  Or if Tree / Hybrid only mostly reduced terraforming eco damage for all improvements.  My concern is that Tree Farm / Hybrid Forest will become even more 'must build first in every base' than they are now.

As for Echelon Mirrors and Condensors, I'd reduce their Eco impact.  As well as take out the negative to terraforming from Forests, they're plenty good enough as is.


Quote
Only in the later game when going for high mineral production.  If you ditch the boreholes and most mineral-boosting facilities and build condenser/farm everywhere, you're looking at far less.

Replacing forests only seems worth it after Hab domes, and even then I think it's debatable.  You're giving up a lot of minerals for extra specialists.  Right now there's not much downside to heavy mineral strategy even late game.  High minerals are needed to build a satellite every turn.  Perhaps this is a side issue where satellites are way too good/cheap relative to Formers, and tend to end the game very quickly.  Formers and crawlers are more viable with low mineral strategy and this is something to consider when balancing.  I've always felt the pacing of the game is a bit off in that the early game techs come a bit slow, and the late game ones come very, very fast. 

I was thinking late-game pollution could be a good incentive to get out of mid-game Forests.  The option of high minerals would still be there I suppose, and you could keep a few 'mineral bases' with eco facilities for high-mineral things like satellites and Secret Projects.  The low mineral, high-pop/energy bases would make 1 turn crawlers / formers / army units.  In turn, late-game techs would take more labs.  I'll have to do out the math to see what 1 late game energy = in terms of average energy credits, then convert it back against modified minerals.  Usually by this point the game is won and I pay little attention.

To me this seems like the most interesting end game, though a bit contrived if only due to the 1 unit/facility a turn limitation.  It does seem a little odd that Forests will also pollute a lot (2 mins each) when we they are supposed to be the 'greenest' land with thematic facilities like Tree Farm and Hybrid Forest.  I feel it's a bit strange that Nutrients take over in the late game, rather than Energy.  But I guess it works, and it seems to be the developer intent with facilities like Hab Dome allowing population to go to 99, Enrichers as the last terraforming upgrade at B7, and no real upgrade to Solar.  Maybe if Echelon mirror doubled Solar Panels in the late game, raised Farm/Solar would blow away Forest.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 15, 2013, 09:06:24 PM
What version are you guys playing?  I've seen mass captures happen in my games.

Apparently he was playing an unpatched game; if someone can provide a game with the latest update with mass captures or (more importantly) ecodamage worms being captured, that might help with tracking down and fixing the problem.

Quote
Yea it's very off, I had an error in the formula.  It was more of an idea to reduce the pollution a bit for non-ICS strategy.

As Earthmichael will tell you, ICS has enough weaknesses anyway; all that's needed is a method of reducing ecodamage that doesn't favor ICS.

Quote
Though with more thought, I think the current formula is very good but just needs some logical tweaks in addition to removing clean minerals (a bad mechanic, as it is an invisible value).

Actually, the 16+pops is noted; only the facility bonus is "extra".  But even the 16+pops is a problem in that it encourages ICS.

Quote
With a quick combat test it seems that there is no cap on +Planet.  So the game does not just set the user's Planet to +3 if you are over.

Yes; for captures it maxes out at +3, for ecodamage it maxes out at +2, and for combat it has no max.

Quote
+5 Planet would be 0 pollution at 20% per Planet.  I'm not so sure that's balanced, Gaia and Cult can hit +5 and +6 respectively with Green+Cyber.  Green+Cyber will be pretty much a requirement to get +4 when they are already very strong SE settings due to +Efficiency.  As well Believers/Usurpers would only be able to hit +3, which is double the pollution of +4.

Keep in mind, though, that:
1. Even +4 would reduce ecodamage to quite a manageable amount even in the late game.  Not 0, but fairly low.  Even +3 would cut ecodamage to a quite manageable level, especially if you're using lower-production terraforming methods.
2. The competitors of Green and Cybernetic also have advantages.  Planned and Eudaimonic give +Growth (quite important if you can keep satellites up), FM and Eudaimonic give +Economy (quite important as well, especially with a lot of trading partners), and Thought Control is great for the warpath.
3. The Cult and Gaians paid for it in the earlier part of the game, by being unable to run FM for +2 ECONOMY.
4. EFFIC does decrease fairly substantially in value once you get a lot of it; the difference between +4 and +6 is far smaller than between +2 and +4.

Even so, I think cutting Cybernetic to +1 PLANET would be a good change with that, as it's pretty strong anyway.  The real losers would be the Drones, who can't run Green; changing their aversion to Cybernetic might help fix that (as +2 PLANET would probably be enough to manage, especially since they can then get +5 INDUSTRY.)


Quote
I would put pollution effects at more like 10% per Planet or cap it at -60% pollution.

I'd say that at 10% it doesn't need a cap, as -60% would already be the best you can get.
However, 10% may be a bit low; perhaps 12.5%?

Quote
Consider that +Planet rating is already good where it gives +10% attack to and with native life, free native life, and easier fungus movement.

On the other hand, getting it isn't cheap; until the latae game, you have to either give up that +1 energy/square (or +2 GROWTH/+1 INDUSTRY if you'd go Planned) and take an additional -2 GROWTH, or play a faction that can't get +1 energy/square until the late game.

Quote
It's probably a known bug but the Planet combat modifier should be the differential between attacker and defender rather than just the attacker's +Planet.

Is this documented anywhere?
 
Quote
- Boils scaling by tech and/or faction size, rather than linearly out of control.  I will have to test this as well, to see how it scales in the patched version.

Use one of my patches.

Quote
- Modified values for eco damage from terraforming.  How difficult would it be to change the terraforming portion of eco damage?

My current patch already allows it to be changed somewhat, but I'm not sure that a patch is needed.

Quote
It seems highly exploitable that crawled tiles do not count as terraforming eco damage.  For example I could create several boreholes outside a city radius and crawl them with 0 terraforming eco damage?

That's what I thought, but then Earthmichael pointed out that those things take a whole lot of former-turns to create; since crawling only provides one resource, it generally won't be worth it.

Quote
I think Boreholes should be much more eco damaging and maybe a little quicker to terraform.  If a Condensor is 4 and a Mirror is 6, a Borehole should be like a 20.  Boreholes seem to be a staple of early game ICS.  What if Boreholes also only got -25% pollution from Tree / Hybrid?  Or if Tree / Hybrid only mostly reduced terraforming eco damage for all improvements.  My concern is that Tree Farm / Hybrid Forest will become even more 'must build first in every base' than they are now.

I've gone through these sorts of ideas too, but as Earthmichael pointed out, the lengthy time to terraform does a lot to balance things out.  Even without terraforming ecodamage, things tend to be pretty balanced out.

Quote
Replacing forests only seems worth it after Hab domes

Not necessarily, but even if that's true, 300 minerals to a base isn't going to happen until substantially after that point anyway.  At the "shortly before hab domes" point, you're probably looking at no more than 60 minerals per base if not crawling minerals, or 80 minerals if crawling.

Quote
Right now there's not much downside to heavy mineral strategy even late game.

Actually, right now there's a huge one, in the form of absurdly high ecodamage.

Quote
High minerals are needed to build a satellite every turn.  Perhaps this is a side issue where satellites are way too good/cheap relative to Formers, and tend to end the game very quickly.

Yes, satellites need reworking.

Quote
I was thinking late-game pollution could be a good incentive to get out of mid-game Forests.  The option of high minerals would still be there I suppose, and you could keep a few 'mineral bases' with eco facilities for high-mineral things like satellites and Secret Projects.  The low mineral, high-pop/energy bases would make 1 turn crawlers / formers / army units.  In turn, late-game techs would take more labs.  I'll have to do out the math to see what 1 late game energy = in terms of average energy credits, then convert it back against modified minerals.  Usually by this point the game is won and I pay little attention.

I'd be interested in seeing what you get.

Quote
Maybe if Echelon mirror doubled Solar Panels in the late game, raised Farm/Solar would blow away Forest.

Raised farm/solar already is competitive forest in the mid-late to late game; raised farm/solar with mirrors and condensers will tend to be something like 4/0/6 or 4/1/6, which is quite competitive with 3/2/2 even at a substantially higher terraforming time.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 16, 2013, 12:52:39 AM
Quote
As Earthmichael will tell you, ICS has enough weaknesses anyway; all that's needed is a method of reducing ecodamage that doesn't favor ICS.

True, and I'd agree.  In fact these changes hurt ICS even more as-is, because they make Eco facilities more valuable.  ICS tends not to run many facilities.

Quote
Keep in mind, though, that:
1. Even +4 would reduce ecodamage to quite a manageable amount even in the late game.  Not 0, but fairly low.  Even +3 would cut ecodamage to a quite manageable level, especially if you're using lower-production terraforming methods.
2. The competitors of Green and Cybernetic also have advantages.  Planned and Eudaimonic give +Growth (quite important if you can keep satellites up), FM and Eudaimonic give +Economy (quite important as well, especially with a lot of trading partners), and Thought Control is great for the warpath.
3. The Cult and Gaians paid for it in the earlier part of the game, by being unable to run FM for +2 ECONOMY.
4. EFFIC does decrease fairly substantially in value once you get a lot of it; the difference between +4 and +6 is far smaller than between +2 and +4.

Even so, I think cutting Cybernetic to +1 PLANET would be a good change with that, as it's pretty strong anyway.  The real losers would be the Drones, who can't run Green; changing their aversion to Cybernetic might help fix that (as +2 PLANET would probably be enough to manage, especially since they can then get +5 INDUSTRY.)

You could drop Cyber to +1 PLANET.  I'd argue that +PLANET doesn't even make sense for Cyber.  But it would be difficult for everyone to agree on what 'makes sense'.  Eudaimonia is much stronger for a builder until +2 ECON becomes diminished.  Thought control is decent but I consider it weak by the time you get it.  This is because PROBE, POLICE, SUPPORT are not all that important (other than perhaps -5 POLICE).  For the same reasons you rarely see Power, Police State (other than for Yang), or Fundamentalism picked.

Quote
I'd say that at 10% it doesn't need a cap, as -60% would already be the best you can get.
However, 10% may be a bit low; perhaps 12.5%?

12.5% to eco damage per +/- Planet level is reasonable.  FM would thus pollute about twice as much as Green (137.5% vs 75%). 
 
Quote
On the other hand, getting it isn't cheap; until the late game, you have to either give up that +1 energy/square (or +2 GROWTH/+1 INDUSTRY if you'd go Planned) and take an additional -2 GROWTH, or play a faction that can't get +1 energy/square until the late game.

This is true.  I don't feel that Gaia is too strong right now, probably about right.  And Cult is definately weak.  -1 IND is too much of a hit on top of not being able to hit +2 ECON. 

Quote
It's probably a known bug but the Planet combat modifier should be the differential between attacker and defender rather than just the attacker's +Planet.

Logically native life should scale in both attack and defense with +Planet rating.  Morale works this way.  But again, I think there's a treading into making FM useless and Green possibly too good.  I would argue that in the later game FM gets diminished as is, when more production is off base and +1 energy/sq is less of a % increase to total energy.  On the other hand,  there's also SPs and unit abilities that hard counter native life.

Quote

That's what I thought, but then Earthmichael pointed out that those things take a whole lot of former-turns to create; since crawling only provides one resource, it generally won't be worth it.

If clean minerals are removed, then this exploit becomes more worth it (to avoid early eco damage).  Losing even one crawler or other unit to eco damage would likely outweigh those former turns.  I'll try and do some analysis on this one to see what I come up with.  Boreholes are certainly slow at 24 former-turns.  And losing out on the energy is also a big drawback.

Quote
Actually, right now there's a huge one, in the form of absurdly high ecodamage.

Well you can exploit around eco damage as is (the Tree farm trick).  If you're at 300 minerals a base, yes you'll have absurdly high eco damage.  But by the time you reach that, the game is over.  Usually you only need around ~60-80 clean mineral limit to pretty much have 0 eco damage in the late game.

Quote
Raised farm/solar already is competitive forest in the mid-late to late game; raised farm/solar with mirrors and condensers will tend to be something like 4/0/6 or 4/1/6, which is quite competitive with 3/2/2 even at a substantially higher terraforming time.

Yea I'll try to run some analysis to prove this.  A lot comes down to the 'value' of nutrients, minerals, and energy, which is the first step.  There are various terraforming layouts possible with raising.  It's fairly complex, because multiple mirrors affect the same solar panel and you need a certain amount of condensors to keep it rainy.  Further the condensors produce 6/0/0. 

My first thought is that 'all condensors and no solar' might be superior to farm/solar raising post Hab domes.  It only takes 12 turns to make a condensor.  It takes 12 turns to raise just once, and 12 for an Echelon mirror.  2 nutrients feeds a Transcendi specialist, which gives 4 labs and 2 econ (and 2 psych, as an added boost.  you can sell off pretty much all psych facilities and still pop boom, at a certain point).  As well each +1 population can give 1/1/1 if you build the satellites.  So a transcendi is essentially worth half another transcendi, and so on.  The effect is powerful - I think you can get to 99 pop without even crawling any nutrients (123 base nutrients, and 99 from satellites).  The only thing holding this strategy back is that you can only pop boom 1 citizen a turn.  The very late game always seems so quick that you'd be unlikely to actually reach 99 size on most bases.  Maybe on a tech stagnation game or something.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 16, 2013, 02:09:59 AM
You could drop Cyber to +1 PLANET.  I'd argue that +PLANET doesn't even make sense for Cyber.  But it would be difficult for everyone to agree on what 'makes sense'.  Eudaimonia is much stronger for a builder until +2 ECON becomes diminished.

Keep in mind that Eudaimonia is the latest tech available; by that point, ECON is pretty close to as diminished as you can get.

Quote
Thought control is decent but I consider it weak by the time you get it.  This is because PROBE, POLICE, SUPPORT are not all that important (other than perhaps -5 POLICE).  For the same reasons you rarely see Power, Police State (other than for Yang), or Fundamentalism picked.

The reason those are rarely picked is because people play maps that are meant for a builder's game.  With a more balanced structure, MORALE and PROBE become substantially more important (MORALE is the other strength of Thought Control; SUPPORT is its weakness); POLICE remains fairly weak in the late game for most purposes, but can help in keeping a newly captured base under control long enough to hurry back its psych facilities.

If one person is going Eudaimonic or Cybernetic, and the other is going Thought Control, and they start the war on an even footing, the side with Thought Control will probably win.

Quote
This is true.  I don't feel that Gaia is too strong right now, probably about right.

Yeah; I'd say that things that boost them should come with a small hit on top of it; how does increasing their fairly irrelevant -1 POLICE penalty to -2 sound?  (The Angels probably could also use the same treatment; while not the easiest faction to play, they seem they'd be very powerful in a 7-player game, since they can keep around the middle in tech even if they neglect it entirely, or become one of the best techers if they put stuff into tech and simply avoid the usual beelines.)

Quote
Logically native life should scale in both attack and defense with +Planet rating.  Morale works this way.

I don't really consider that enough to change the rules, though it could be made an option once I'm ready to add requests.

Quote
But again, I think there's a treading into making FM useless and Green possibly too good.

Considering that the consensus now seems to be that FM is the best for a skilled player, and Green never used until the late game unless you're Deidre, Dawn, or Morgan, I don't see that as such a concern.

Quote
I would argue that in the later game FM gets diminished as is, when more production is off base and +1 energy/sq is less of a % increase to total energy.

True...of course, if you're running Wealth (and maybe Eudaimonic) as well, it's worth a bonus to commerce, which can be quite the boost when you've got several pact brothers and a global trade pact.

Quote
If clean minerals are removed, then this exploit becomes more worth it (to avoid early eco damage).

Not really, as you can build a mine+road on a rocky square at half the cost (less if you were planning to make a road anyway) and crawl that instead.

Quote
Losing even one crawler or other unit to eco damage would likely outweigh those former turns.  I'll try and do some analysis on this one to see what I come up with.

I've found that a former-turn can be estimated at around 0.2 resources per turn (depending on various things such as tech level; this is fairly rough).  At 20 minerals/former, that means that a mineral-turn is worth around 0.01 minerals/turn, implying a per-turn return of 1/10, so a former-turn is worth about 2 minerals.  At 24 former-turns=48 minerals, it's not worth it for a single crawler, but would be worth it for 2.  Keep in mind, though, that a single borehole doesn't produce such extreme amounts of ecodamage; under standard rules it's a bit over 1 mineral's worth, and even with my proposal it's only 6.25 minerals' worth when worked.

Quote
But by the time you reach that, the game is over.

Yeah, that's another annoying thing: The game pretty much never plays out to the end.

Quote
Yea I'll try to run some analysis to prove this.  A lot comes down to the 'value' of nutrients, minerals, and energy, which is the first step.  There are various terraforming layouts possible with raising.  It's fairly complex, because multiple mirrors affect the same solar panel and you need a certain amount of condensors to keep it rainy.  Further the condensors produce 6/0/0. 

Good luck.  (Also, IMO 6/0/0 is a bit much; see what happens if it's 5/0/0 instead.)

Quote
My first thought is that 'all condensors and no solar' might be superior to farm/solar raising post Hab domes.

With transcendi and satellites, "all condensors and no solar, crawled" is probably superior to fungus with the Manifold Nexus and +3 PLANET.  It is seriously overpowered, as you say.  That's a large part of why I think cutting it to 5 nutrients would be better.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 16, 2013, 05:45:00 AM
Quote
If one person is going Eudaimonic or Cybernetic, and the other is going Thought Control, and they start the war on an even footing, the side with Thought Control will probably win.

Thought Control is decent, and probably balanced enough.  I'd want to see it at +3 POLICE, and/or some higher POLICE benefits.  I like the idea of 'conquer' SEs being able to more easily put down drones of captured bases (without nerve stapling).  I would say Thought Control is much better than say Power, Police State, or Fundamentalism.  Even in a small warring map, I couldn't see picking any of these 3.  The early techs are too much of a power increase.  Maybe if Fundamentalism was +2 MORALE, Police State had +3 POLICE (PS/FM combo would be more interesting at -2 POLICE), and Power was -1 IND, I might consider them.  I do feel they should be more viable options, right now it's all about staying in 'builder' SEs even when warring. 

Quote

yeah; I'd say that things that boost them should come with a small hit on top of it; how does increasing their fairly irrelevant -1 POLICE penalty to -2 sound?  (The Angels probably could also use the same treatment; while not the easiest faction to play, they seem they'd be very powerful in a 7-player game, since they can keep around the middle in tech even if they neglect it entirely, or become one of the best techers if they put stuff into tech and simply avoid the usual beelines.)

-2 POLICE is quite crippling at the very beginning.  Much worse than -2 ECON or -1 INDUSTRY.  The second citizen always has to be allocated to Specialist.  Police State comes too late.  Rec Commons are too much of an early investment.  I'm not so sure Gaia is stronger than say University or Morgan either, even if you fix eco damage.  Believers definately need some help, though.

[Considering that the consensus now seems to be that FM is the best for a skilled player, and Green never used until the late game unless you're Deidre, Dawn, or Morgan, I don't see that as such a concern.]

Agreed, FM is very strong.  Though once airpower comes out, it seems like you can't run it anymore.  To me the whole -POLICE of FM doesn't make a lot of sense.  That seems like something that should be tied to Democracy. 

Quote
[With transcendi and satellites, "all condensors and no solar, crawled" is probably superior to fungus with the Manifold Nexus and +3 PLANET.  It is seriously overpowered, as you say.  That's a large part of why I think cutting it to 5 nutrients would be better.]

Although one can argue it's the Transcendi specialists that are just as game-breaking.  2 nutrients into 8 energy, and a satellite cap increase.  What were they thinking? Lol

I'll get to it over the next while, its a decent sized problem.  I think it will come out that even 5 nutrients win out, crawled or not. 
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 16, 2013, 02:25:30 PM
Thought Control is decent, and probably balanced enough.  I'd want to see it at +3 POLICE, and/or some higher POLICE benefits.

Even +2 is pretty good, especially as there's a good choice you're going Police State and Power with it.

Quote
I would say Thought Control is much better than say Power, Police State, or Fundamentalism.

Future societies tend to be strong.

Quote
Even in a small warring map, I couldn't see picking any of these 3.  The early techs are too much of a power increase.

I think the idea is after you get the techs you need for your rush; you don't want to run fundie right away, but once you've got the ability to build those impact rovers, it's a nice boost (or for Santiago an extremely strong boost, as combined with command centers and monoliths you can hit elite and get +1 movement for all your units.)
Police State seems it should be actually not that bad for a builder in the early game; that early, efficiency isn't as important, and -2 drones per base is quite nice (especially on Transcend).
And as for Power...morale lets your forces win battles, so you can send them back (or into captured bases) to repair, vastly outweighing the industry effect if you're going for full-fledged war.  The ability to support more non-clean units is also fairly nice, as clean units are expensive.  (A more builder-ish game will probably find it not worth it, though.)

Quote
I do feel they should be more viable options, right now it's all about staying in 'builder' SEs even when warring. 

I suspect a lot of that is because people play on maps designed for builder games, with factions designed for builder games.  Although nerfing the advantage of early-game tech slightly might be a good idea, perhaps by nerfing beelines by making the three resource-lifting techs all be on different paths (rather than in a line one after the next), with none of them requiring Centauri Ecology.

Quote
-2 POLICE is quite crippling at the very beginning.  Much worse than -2 ECON or -1 INDUSTRY.  The second citizen always has to be allocated to Specialist.  Police State comes too late.  Rec Commons are too much of an early investment.

I do see the concern (though on the other hand, Transcend is supposed to be very hard.)  Any alternative ideas?

Quote
I'm not so sure Gaia is stronger than say University or Morgan either, even if you fix eco damage.

True; those are quite strong.  Although there are other ideas that I've been considering that would weaken them (fixing drones so that University's drone penalty actually means something, and nerfing Democracy to make pop booms difficult* for everyone and impossible for Morgan and Aki before Eudaimonia)

*Usually around Hive level.

Quote
Believers definately need some help, though.

I suspect the help they need is simply not playing on maps designed to hurt them, plus nerfs to a few of the strongest beelines.

Quote
Agreed, FM is very strong.  Though once airpower comes out, it seems like you can't run it anymore.  To me the whole -POLICE of FM doesn't make a lot of sense.  That seems like something that should be tied to Democracy.
 

I think the idea is one of social unrest.  As for airpower, the usual trick is to base all the air units from a single base (or two if you really have a lot) which has a punishment sphere.

Quote
Although one can argue it's the Transcendi specialists that are just as game-breaking.  2 nutrients into 8 energy, and a satellite cap increase.  What were they thinking? Lol

On the other hand, the idea of a third tier of specialists does make sense as a late-game/endgame thing; I think a better approach would be to reduce the value of pure-condenser spaces, make it possible in the late-game/endgame to crawl all three resources in the base square (so that producing only one resource isn't an advantage that way), and make satellites a lot easier to shoot down than to build.

Quote
I'll get to it over the next while, its a decent sized problem.  I think it will come out that even 5 nutrients win out, crawled or not.

5 nutrients not crawled will not win out except maybe with satellites, which do need a nerf.
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on April 16, 2013, 06:29:32 PM
Police State is not a good builder choice.

1. If you are running Planned, Police State -2 EFF will cripple energy production.  Only Hive can do this, because of immunity to EFF.

2. If you are running FM, Police State's police bonus is not that useful.  And the -2 EFF still hurts a lot (but not crippling).  +2 Support can be worth up to 2 minerals per city, but I would usually prefer the growth (and EFF) bonus of Democracy.  The main problem with Democracy to me is losing the 10 minerals to start a new city (which allows me to rush build Recycling Tanks), so I usually pick nothing in the top row until I have finished my first wave of expansion.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 16, 2013, 06:37:20 PM
Police State is not a good builder choice.

1. If you are running Planned, Police State -2 EFF will cripple energy production.  Only Hive can do this, because of immunity to EFF.

2. If you are running FM, Police State's police bonus is not that useful.  And the -2 EFF still hurts a lot (but not crippling).  +2 Support can be worth up to 2 minerals per city, but I would usually prefer the growth (and EFF) bonus of Democracy.  The main problem with Democracy to me is losing the 10 minerals to start a new city (which allows me to rush build Recycling Tanks), so I usually pick nothing in the top row until I have finished my first wave of expansion.

Ah...and in a builder game you're getting one of those (probably both) before Doctrine: Loyalty, and usually it's more important to get one of them than to get the extra POLICE.  Basically, Police State isn't worth getting because Frontier/Planned is better than Police State/Simple, and Police State/Planned is horrible unless you're Yang (or maybe Deidre or Aki, but they tend not to work well with that combo anyway.)

Although Police State might still have niche uses in hybrid games (where you want a substantial standing army but don't want to sacrifice as much research as fundie would imply)...of course, that's not a very good approach when playing on a map like Veterans.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 16, 2013, 08:07:49 PM
Quote
Even +2 is pretty good, especially as there's a good choice you're going Police State and Power with it.

+2 POLICE from POLICE is not enough to offset -2 EFF.  As EM stated it's only ever worth it for Yang.  It should be +3 POLICE or have -1 EFF to make it a more interesting choice.  FM/Police would be -2 POLICE which I think is interesting.  Yang could run FM/Police but IMO Planned is still every bit as good for him (where +GROWTH and +IND get better the more you have). 

I feel that PS/TC is not worth going because there's no added benefit to +4/+5 POLICE.  Likely you'd see Demo/TC or Fundamentalism/TC.  POLICE needs a +4/+5 modifier or its rather useless late game. 

Something like:
+4, 3 units function as police, police effect tripled!
+5, 3 units function as police, police effect quadrupled! (only reachable with Ascetic Virtues or as Spartans)

By the time you hit Thought Control, the builder SEs don't seem to have much drone problems anyways.  Thoughts?

Quote
On the other hand, the idea of a third tier of specialists does make sense as a late-game/endgame thing; I think a better approach would be to reduce the value of pure-condenser spaces, make it possible in the late-game/endgame to crawl all three resources in the base square (so that producing only one resource isn't an advantage that way), and make satellites a lot easier to shoot down than to build.

I agree, specialists should get stronger late game.  I do like that they get more powerful...it's just that Transcendi are ~4x more powerful than previous Specialists.  It's really hard to balance around 2 N -> 8 E.  Triple crawling is an interesting idea.  If you really want to benefit farm/solar late though I would consider making it double crawling.  Perhaps to limit Forest and Borehole power late, have the double crawling work only for Nutrients+Energy. 

Quote
I suspect the help they need is simply not playing on maps designed to hurt them, plus nerfs to a few of the strongest beelines.

Believers get smashed even harder early game on small maps.  They take too long to get Impact Laser / Rovers.  The +2 SUP is decent for early large armies but not enough on its own.  At the least I think they should have IMPUNITY on Fundamentalism. 

Quote
[I think the idea is one of social unrest.  As for airpower, the usual trick is to base all the air units from a single base (or two if you really have a lot) which has a punishment sphere.

Interesting.  This really seems like an exploit to me.  SE Settings under 0 POLICE really should not be able to punishment sphere (it's pretty much like nerve stapling). Though p-drones don't work all that great either.  It seems they are calculated after facilities so that you pretty much always get riots from them.  IMO air units at home shouldn't cause p-drones.  I assume they made it this way to 'balance' choppers.

Quote
I think the idea is after you get the techs you need for your rush; you don't want to run fundie right away, but once you've got the ability to build those impact rovers, it's a nice boost (or for Santiago an extremely strong boost, as combined with command centers and monoliths you can hit elite and get +1 movement for all your units.)
Police State seems it should be actually not that bad for a builder in the early game; that early, efficiency isn't as important, and -2 drones per base is quite nice (especially on Transcend).
And as for Power...morale lets your forces win battles, so you can send them back (or into captured bases) to repair, vastly outweighing the industry effect if you're going for full-fledged war.  The ability to support more non-clean units is also fairly nice, as clean units are expensive.  (A more builder-ish game will probably find it not worth it, though.)

Yea I may be underestimating/misunderstanding Morale.  Hitting +3 SE morale seems to be the big breakpoint.  I will have to play more games aiming for high morale as I thought troop rank was determined at the time of troop production.  Therefore Power switching isn't that good.  -2 IND is a lot less troops.

Hitting Elite is definately huge against human players.  +1 move is absolutely insane, imo even better than the +% modifiers, as a lot of battles come down to positioning.  You can pretty much deadzone the enemy backwards a tile a turn and there's little they can do to counter it.

Quote
I do see the concern (though on the other hand, Transcend is supposed to be very hard.)  Any alternative ideas?

Put Gaia to +1 EFFIC for a small nerf.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 16, 2013, 08:49:02 PM
+2 POLICE from POLICE is not enough to offset -2 EFF

Why not?  If you were previously at +0 or better and get it in the early game, 2 extra non-drones could be worth 4 energy per base, which could easily be more than the EFFIC penalty is worth.  Later on, it becomes less useful, though SUPPORT can still help a lot in wartime.

Quote
Yang could run FM/Police but IMO Planned is still every bit as good for him (where +GROWTH and +IND get better the more you have). 

More importantly, FM only raises him from -2 ECONOMY to +0, which really isn't worth it.

Quote
I feel that PS/TC is not worth going because there's no added benefit to +4/+5 POLICE.

But there is to +3, especially in a newly captured base.  And if you're playing Believers or got the Living Refinery (but didn't manage to grab the Cloning Vats), the SUPPORT boost of PS can be quite nice as well to offset the penalty of TC and allow you to get a nice large number of troops without cutting into your production capacity or paying extra for clean. 

Quote
Likely you'd see Demo/TC or Fundamentalism/TC.

Fundamentalism/TC might happen if you're fighting the Data Angels or something (and aren't playing Believers), but usually you're going to want to combine TC with Power for the morale boost, which makes Fundamentalism not so useful.  Demo/TC seems unlikely, as TC usually means you're planning to go for the big offensive soon, which makes the long-term benefits of Demo less useful.

Quote
By the time you hit Thought Control, the builder SEs don't seem to have much drone problems anyways.  Thoughts?

True usually due to all the psych-boosting facilities, but newly captured bases could be different.

Quote
I agree, specialists should get stronger late game.  I do like that they get more powerful...it's just that Transcendi are ~4x more powerful than previous Specialists.

No, just 1.6-2X more powerful than the previous, which were ~1.6-2X more powerful than the first tier.

Quote
It's really hard to balance around 2 N -> 8 E.

Not really; just accept that nutrients increase in power as the game progresses.

Quote
Triple crawling is an interesting idea.  If you really want to benefit farm/solar late though I would consider making it double crawling.  Perhaps to limit Forest and Borehole power late, have the double crawling work only for Nutrients+Energy.

Farm/solar is already competitive with forest (it'd be stronger except that forest is a lot easier to terraform), and boreholes have been left far in the dust by this point. 

Quote
Believers get smashed even harder early game on small maps.  They take too long to get Impact Laser / Rovers.

Even without impact, Believer laser troops can beat enemy synthmetal, and their gun troops can beat enemy unarmored.  They will need to go pod popping and maybe do some trading (or extortion), though.  But they really shouldn't have small maps; everyone is going momentum then.  A better situation for them is the natural one: Medium sized maps where the factions start at random distances from each other.  That way, everyone will tend more toward their natural tendencies, and everyone has a chance.

Quote
The +2 SUP is decent for early large armies but not enough on its own.  At the least I think they should have IMPUNITY on Fundamentalism. 

IMPUNITY on fundamentalism wouldn't really help here, as they're not going to go for fundamentalism until they have the techs needed for a proper attack force.

Quote
Yea I may be underestimating/misunderstanding Morale.  Hitting +3 SE morale seems to be the big breakpoint.

Generally, it's probably +3 for land (which can use monoliths) and +4 for sea/air (which can't).

Quote
I will have to play more games aiming for high morale as I thought troop rank was determined at the time of troop production.  Therefore Power switching isn't that good.

When you switch SE, it does affect the morale of units already created.

Quote
-2 IND is a lot less troops.

16% less, to be exact.   But a 25% boost to effectiveness is quite nice (remember, higher-effectiveness troops can be worth more by more than the effectiveness boost if you use a "rolling repair" strategy), and if you get high enough that your infantry are as mobile as your opponent's speeders, then you have the production advantage.

Quote
Hitting Elite is definately huge against human players.  +1 move is absolutely insane, imo even better than the +% modifiers, as a lot of battles come down to positioning.  You can pretty much deadzone the enemy backwards a tile a turn and there's little they can do to counter it.

What's "deadzoning"?

Quote
Put Gaia to +1 EFFIC for a small nerf.

It's an idea, though I don't really like it that much.  I think the question is more: How can a POLICE penalty be made significant but not crippling?
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 17, 2013, 01:35:24 AM
Yea I've really gotten on a tangent regarding many relatively small balance tweaks.

Generally I think the AI builds well in SMAC.  But I can appreciate trying to fix things like their SE settings, lack of crawlers, not making satellites, etc.  Ideally the AI would play this side well without cheating.

If I had one wish for SMAC/X though it would be for the AI to use military units better and consider 'deadzones'.  It just does not fight well at all.
 
A 'deadzone' is the area in which troops will decisively lose to the opponent's next move. 

Here's a simple example:
I have a 2/1 infantry unit, and so does my enemy.  It's my turn and they are separated by 1 tile.  If I move into the tile, my unit will die.  Therefore the tile between us is a 'deadzone' for me.  I cannot go there.  Conversely, neither can my enemy because I will kill them if they move beside me.  The tile is therefore a 'deadzone' for both of us, meaning a stalemale.  Neither side can advance.

Now let's instead say I have a 2/1 infantry, and my enemy a 2/1 rover.  There is still one tile between us and it's my turn.  My unit is now on a 'deadzone' because if I don't move backwards, my unit will die.  I cannot move forward because I cannot attack his unit.  So naturally I retreat backwards out of my 'deadzone'.  There is no sense in losing my infantry to nothing.  My enemy can now move forward one square (but not two, because that's a 'deadzone' for him).  I lose one square of territory.  And so this repeats until I get more troops.

What does this mean?  Since offense vastly overpowers defense in SMAC, +1 movement generally puts the enemy into a deadzone.  You can generally force the enemy to retreat indefinately.  This is also why airpower is so good.

I'll go more into the implications this has for effective combat tactics for 'invaders' and 'invadees' in my next posts.  Most of the time two sides have equal movement, and this is the more interesting case. 
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 17, 2013, 04:45:14 AM
Yea I've really gotten on a tangent regarding many relatively small balance tweaks.

Small tweaks tend to be better, when they work.

Quote
Here's a simple example:
I have a 2/1 infantry unit, and so does my enemy.  It's my turn and they are separated by 1 tile.  If I move into the tile, my unit will die.  Therefore the tile between us is a 'deadzone' for me.  I cannot go there.  Conversely, neither can my enemy because I will kill them if they move beside me.  The tile is therefore a 'deadzone' for both of us, meaning a stalemale.  Neither side can advance.

Of course, if you have defensive units, you can use them to advance, and push the enemy back that way.  Or if you've got enough of a material advantage, you can just move forward along a whole line, and he won't be able to kill them all.

Quote
What does this mean?  Since offense vastly overpowers defense in SMAC, +1 movement generally puts the enemy into a deadzone.  You can generally force the enemy to retreat indefinately.  This is also why airpower is so good.

Actually, once AAA becomes available, defense is comparable or superior to offense when the attacker is an air unit (though choppers still get overpowered simply because they get so many attacks per turn even at a sizable distance).  ECM accomplishes the same thing vs. rovers, to a lesser extent.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 18, 2013, 06:21:13 PM
Quote
Actually, once AAA becomes available, defense is comparable or superior to offense when the attacker is an air unit (though choppers still get overpowered simply because they get so many attacks per turn even at a sizable distance).  ECM accomplishes the same thing vs. rovers, to a lesser extent.

It seems in my games that there's always a rush to Fusion (10) and Shard (13) while defense is still at 3.  Fusion Power and Satellites are just so good.  AAA seems to be a side tech, as are the mid-game defensive upgrades.  Thus the 'advancing column' never really works that well even when air comes into play.  The issue is that once your defensive garrison dies, the 'attack' troop dies to the next air unit with no losses (as it's usually a Rover with defense of 1).  Of course you can counter that air unit with an anti-air, so there does get to be a battle of attrition.  In the earlier game ECM is only +50%, which in most circumstances is not quite enough to offset the power of attack.  I will agree that once defense catches up to "6", air units can again be somewhat safe in a city with an Aerospace Complex and AAA garrisons.

Back on the topic of Gaia, you could make it so that -1 POLICE (or more) means you can't commit any atrocities, perhaps even with UN Charter repealed.  As well I don't feel punishment spheres should work with -POLICE SE.  -POLICE SE already applies to nerve stapling...I'm not sure why they didn't extend it.  But this would likely be too much to fix/change...I assume anyways.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 18, 2013, 07:48:41 PM
Quote
Actually, once AAA becomes available, defense is comparable or superior to offense when the attacker is an air unit (though choppers still get overpowered simply because they get so many attacks per turn even at a sizable distance).  ECM accomplishes the same thing vs. rovers, to a lesser extent.

It seems in my games that there's always a rush to Fusion (10) and Shard (13) while defense is still at 3.  Fusion Power and Satellites are just so good.

I wonder if allowing 4 defense at Advanced Subatomic Theory would help balance things (together with nerfing satellites and making Doctrine:Air Power require AMA).  Then Silksteel could be 5, Photon could be 6, and Probability Mechanics is a powerful defensive tech anyway due to tachyon fields.

Quote
In the earlier game ECM is only +50%, which in most circumstances is not quite enough to offset the power of attack.

Well, before Missile weapons it gets fairly close (enough that a 2-to-1 numerical advantage should be enough to compensate); it's only after that that it lags behind.

Quote
I will agree that once defense catches up to "6", air units can again be somewhat safe in a city with an Aerospace Complex and AAA garrisons.

Why both?  Even with only one, you're looking at 6X2X1.25=15 against an attack of 10 or 13.

Quote
Back on the topic of Gaia, you could make it so that -1 POLICE (or more) means you can't commit any atrocities, perhaps even with UN Charter repealed.

I think this might be a good idea.

Quote
As well I don't feel punishment spheres should work with -POLICE SE.  -POLICE SE already applies to nerve stapling...I'm not sure why they didn't extend it.  But this would likely be too much to fix/change...I assume anyways.

Nah, it shouldn't be that too hard for the punishment spheres.  For atrocities it'd be harder, but that can always be house-ruled.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 18, 2013, 09:05:26 PM
I think even with those changes you'd still see a rush to Fusion and Shard.  The problem is moreso that they lie along the Fusion reactor line and that the defense techs can be skipped.  A fusion reactored 3 unit is as good as one without and 6 (though, the former may cost a bit more in some cases).  I noted that neither 8 (Chaos) nor 4 (Silksteel) are required for Fusion Power.  Now the intent was probably that you could 'rush' these at a long term cost, like with Gatling Gun (5).  So that might be okay as-is, if satellites didn't pile onto this problem.  A better fix might be to make Organic Superlubricant require Applied Relativity rather than Synthetic Fossil Fuels (a requirement you'll always have by getting Fusion, which leads me to think that was an error on the part of the developers).  I think it would be interesting if the satellites required Chaos and Silksteel instead, but I haven't really analyzed what that would imply for strategy.

[/quote]Why both?  Even with only one, you're looking at 6X2X1.25=15 against an attack of 10 or 13.[/quote]

Because once the AAA unit(s) die, the rest of your needlejets/copters on the tile with 1 defense get routed.   It's even more relevant with copters than needlejets because copters can attack over and over, whereas a needlejet can sometimes be counterhit (upon landing, or in the air.  Though I don't think it's too common to run very many AAA air units, they are generally not useful).  Therefore you have to have enough defense to prevent this from happening.  Running a lot of defense units is not that viable either, because you typically have many cities to defend.  Eventually you'll get surrounded, or your crawlers and formers with 1 defense will be whittled down.

A battle has to be considered in terms of total mineral losses (factoring for mineral production of both sides) and not just who has a higher number.  I have to do a more full analysis of optimal combat still, where not all units have the same cost.  Intuitively I think we all already know copters are too good, whereas infantry and attack ships are generally not so good.



I still feel the whole punishment sphere to avoid p-drones seems like an exploit.  In fact a lot of things that have this kind of an SE impact tend to be game breaking.  -5 POLICE from FM was intended to make it very hard to fight offensively with FM on.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 18, 2013, 10:38:40 PM
I think even with those changes you'd still see a rush to Fusion and Shard.  The problem is moreso that they lie along the Fusion reactor line and that the defense techs can be skipped.

On the other hand, Fusion reactors are fairly high up as compared to Subatomic; furthermore, by the time you get Fusion reactors and are ready to get the fusion laser you've already had the ability to get Subatomic for quite a while.  Basically, even after getting fusion reactors (and other stuff you're likely to get such as synthetic fossil fuels), fusion laser is one more tech and Plasma Shard is three after that; Subatomic is 1 more tech, and silksteel is one after that.

Furthermore, Fusion power requires 18 techs, more if you stop to get weapons better than 5 or armor better than 2.  So yes, your 5/2/*2 unit is going to be stronger than your opponent's 8/4/*1...but that 5/2/*2 requires 18 techs, whereas your opponent could grab 6/4/*1 well before then and attack you (and his tree includes Industrial Automation whereas yours doesn't.)

Quote
I noted that neither 8 (Chaos) nor 4 (Silksteel) are required for Fusion Power.

Neither are 6 (Missile) or 3 (Plasma).

Quote
So that might be okay as-is, if satellites didn't pile onto this problem.

Satellites need nerfing in any case.

Quote
rather than Synthetic Fossil Fuels (a requirement you'll always have by getting Fusion

False; Synthetic Fossil Fuels is not required for Fusion.

Quote
I think it would be interesting if the satellites required Chaos and Silksteel instead, but I haven't really analyzed what that would imply for strategy.

Better idea: Switch the positions of Orbital Defense Pod and Sky Hydroponics Lab.

Quote
It's even more relevant with copters than needlejets because copters can attack over and over

Nerfing copter speed would reduce this substantially.

Quote
Running a lot of defense units is not that viable either, because you typically have many cities to defend.  Eventually you'll get surrounded, or your crawlers and formers with 1 defense will be whittled down.

Ok, that makes sense; while one of the two would be sufficient for a single city, or for a short period of time (or to defend an assault force from aerial defense), a long-term assault along a lengthy border does require both, plus some way to take out air units in the air.)

Quote
A battle has to be considered in terms of total mineral losses (factoring for mineral production of both sides) and not just who has a higher number.  I have to do a more full analysis of optimal combat still, where not all units have the same cost.  Intuitively I think we all already know copters are too good, whereas infantry and attack ships are generally not so good.

I think the answer for copters is to reduce them to 8 movement, rather than 8+2*reactor, and (because a lot of attacks won't help if you're usually going to die on the first one) to make AAA available before Air Power, and strengthen defense in general.

Quote
I still feel the whole punishment sphere to avoid p-drones seems like an exploit.  In fact a lot of things that have this kind of an SE impact tend to be game breaking.  -5 POLICE from FM was intended to make it very hard to fight offensively with FM on.

I'd agree here.
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on April 18, 2013, 11:44:01 PM
The multi-attack capability of copters is just broken; copters should not have multi-attack.  If there is no way to mod copter to not have multi-attack, they should be banned, because they screw up the balance.

Even if you greatly reduce speed, copters are still very powerful in an active defense role.  They are way too powerful on defense even with low speed.

Copters are the reason air power is considered by many to overpowered.  With Needlejets only being able to attack every other turn, it is difficult for Needlejets to demolish the remaining defenders once the AAA defenders are destroyed.  In contrast, often a single copter can wipe out 4+ units in a stack once the AAA defenders are destroyed.

Even if you drop the speed of copters, a single copter can devestate a whole stack of attackers once the AAA defender(s) in the attacking stack have been destroyed.
Title: Re: SE choices for AI - suggestions
Post by: Buster's Uncle on April 18, 2013, 11:45:55 PM
Copters would be almost completely useless if you nerfed the multi-attacks, wouldn't they?
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on April 18, 2013, 11:49:11 PM
Copters would function as an attack/retreat unit, able to make a single attack then retreat to safety in a single turn.  This would still give them double the attack capability of needlejets, and they are much less subject to counterattack than needlejets.

Plus, if you need to travel particularly far, the copter can go twice as far as a needlejet, as long as you are willing to repair damage when you get there.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 19, 2013, 12:02:40 AM
The multi-attack capability of copters is just broken; copters should not have multi-attack.  If there is no way to mod copter to not have multi-attack, they should be banned, because they screw up the balance.

Even if you greatly reduce speed, copters are still very powerful in an active defense role.

Unless you also boost defense enough that AAA units can defend their stack (which is not a whole bunch of spread-out bases and so does not suffer from the need to spread out defenders) at good cost-effectiveness (remember, air units don't do collateral damage, so stacking is a lot safer against them.)

Quote
Even if you drop the speed of copters, a single copter can devestate a whole stack of attackers once the AAA defender(s) in the attacking stack have been destroyed.

True, but if you spend 300 minerals of copters to kill 200 minerals of AAA defenders and 100 minerals of attackers, you haven't really come out ahead, have you?  So that means that, at the very least, there is some amount of boosting to AAA defenders that will solve that problem.  Exactly what that amount is may require testing, though.
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on April 19, 2013, 12:52:48 AM
This is just the point.  Copters should not be used to justify any changes, because if you compensate for overpowered copters, you make Needlejets nearly irrelevant.

Far better to just ban copters.

As for mineral balance, if I am trying to attack, and  bring up a stack of 10 attackers, the defender only needs to get rid of the AAA defender(s) and then a copter or two can wipe out the rest of the stack, a single copter eliminating 4+ attackers, whereas a Needlejet can only take out one attacker every other turn.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 19, 2013, 03:04:10 AM
Hmm seems I was going by an incorrect tech tree chart.  :-[

Though the point remains, offense techs are king and you should always beeline to Shard before Silksteel.

Air Superiority units are generally a sub-optimal purchase for many reasons:
higher cost than non-AS
-50% attack to ground
less movement speed than non-AS
AS takes up a slot (deep radar or clean have to be excluded)
A non-AS unit is on par when hitting an AS unit

Choppers to 8 move would still leave them overpowered.  They're moreso problematic because they also don't crash when out of fuel.  Why does this matter?  You can take a risk and sneak in a small stack to somewhere other than a city.  With infiltrate, you can know exactly where the enemy airforce is.  Since choppers have multiple attacks, it's a worthy gambit to wipe out an enemy's entire air force.  I would say air in general has way too much movement speed.  Needlejets, Copters, and Gravships alike.  Ideally the main benefit of air would not be the crazy movement, but its ability to fly over enemy troops.  Or at least ignore terrain which it already does.  Once air comes into play pretty much the only attacking ground units you want are Scout Rovers, just to capture cities.  This later becomes Drop Rovers.

Granted the AI isn't smart enough to play as a human would/should.  I'll have to do an analysis of AAA counterattack chains.  It gets more interesting, if both sides have the same movement on their air units.  AAA makes a defensive infantry unit cost effective against air even in the open.  But not quite cost effective enough to protect offensive land units under them.  Therefore you still end up making all Needlejets on offense.  Hence they are also overpowered, just less so than Copters.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 19, 2013, 04:24:59 AM
This is just the point.  Copters should not be used to justify any changes, because if you compensate for overpowered copters, you make Needlejets nearly irrelevant.

Not really.  Needlejets can bomb enemy improvements (if copters try to do that, they land, take damage, and can be attacked by non-SAM units.)  Needlejets can attack enemy formers/crawlers more than 4 spaces away from where they're based.

Quote
As for mineral balance, if I am trying to attack, and  bring up a stack of 10 attackers, the defender only needs to get rid of the AAA defender(s) and then a copter or two can wipe out the rest of the stack, a single copter eliminating 4+ attackers, whereas a Needlejet can only take out one attacker every other turn.

Firstly, for a single copter to eliminate 4 attackers would require them to be only 3 spaces away (counting "right next to" as one space away.)  That's not exactly long-distance.  Secondly, with defense boosted, AAA would be quite a hassle to take out with only copters, probably costing you more than you gain (or at least giving a lower percentage bonus than simply waiting for the enemy to approach and them attacking your perimiter-defensed base*).  Thirdly, as the game progresses and more advanced reactors are used, it becomes possible to put some defense on attackers at no extra cost...it's not going to be enough to beat an attack, but it might damage the attacker enough to make 4 attacks in a row unfeasible.

*And if they try to go around the base, you should have enough time to take them out with rovers or infantry, a lot more efficient than copters.


I think you are being misled by your experience of copters being overpowered under the current rules and tech tree.  But that experience need not hold if things are changed.

Though the point remains, offense techs are king and you should always beeline to Shard before Silksteel.

True under the current tech tree; not necessarily under a changed tech tree.

Quote
Air Superiority units are generally a sub-optimal purchase for many reasons:
higher cost than non-AS
-50% attack to ground
less movement speed than non-AS
AS takes up a slot (deep radar or clean have to be excluded)
A non-AS unit is on par when hitting an AS unit

I discussed AAA rather than AS.  AS is useful when you want to attack air units, and that's about it.  (Note, however, that AS can be placed on ground units, in which case it does not affect movement speed or attack strength, and they probably wouldn't use deep radar anyway but it allows them to attack air.)

Quote
Choppers to 8 move would still leave them overpowered.  They're moreso problematic because they also don't crash when out of fuel.  Why does this matter?  You can take a risk and sneak in a small stack to somewhere other than a city.  With infiltrate, you can know exactly where the enemy airforce is.  Since choppers have multiple attacks, it's a worthy gambit to wipe out an enemy's entire air force.

Which circumstance are we talking about here?  If the enemy airforce is in the base, you're attacking the enemy's defenders (with damaged choppers).  If not, then you need AS (so your choppers have even lower speed and even fewer attacks), and you also need some way to lure them out of the base (not so easy.)

Quote
I would say air in general has way too much movement speed.  Needlejets, Copters, and Gravships alike.  Ideally the main benefit of air would not be the crazy movement, but its ability to fly over enemy troops.  Or at least ignore terrain which it already does.

I think that their movement speed does make sense; the problem is that it doesn't come at a larger penalty to combat effectiveness (which would do a lot more to require ground units than weakening air units in a way that leaves them superior to ground units).

Quote
Granted the AI isn't smart enough to play as a human would/should.  I'll have to do an analysis of AAA counterattack chains.  It gets more interesting, if both sides have the same movement on their air units.  AAA makes a defensive infantry unit cost effective against air even in the open.  But not quite cost effective enough to protect offensive land units under them.

AAA alone wouldn't, but a cheaper (read: free on defensive units) AAA plus an overall boost to defense might.
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on April 19, 2013, 12:59:14 PM
There is no justification for Copter multi-attack.  It is the only air (or sea) unit with this capability.  Even higher tech air units do not have this capability.

Anything you do to increase defense against Copters with also affect all other air units (who don't need the fix).

Dropping copter mobility is a start, but it has little effect on defensive roles for Copters.  If I wait until a stack is 1-2 squares away from a city being defended, I can get 4-6 attacks from a single 8 speed chopper and still land safely.  This is a lot different than the attack every other turn from a Needlejet, which is then vulnerable to counterattack.  Choppers are why air power is widely considered overpowered, and they can still be overpowered even with only 8 speed.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 19, 2013, 02:37:16 PM
There is no justification for Copter multi-attack.  It is the only air (or sea) unit with this capability.

And gravships are the only air unit that doesn't need fuel, and infantry are the only units with +25% vs. bases...different chasses have different strengths, and that's perfectly ok, as long as they pay for it somehow.  IMO, the best way to make copters pay for it (as compared to other air units) is with fairly small range.

Quote
Even higher tech air units do not have this capability.

But gravships (the only higher-tech air unit) have other strengths instead.

Quote
Anything you do to increase defense against Copters with also affect all other air units (who don't need the fix).

True to an extent, but copters will suffer a lot more, as (1) boosting defense hurts a unit that attacks several times a turn or is underpowered a lot more than it hurts a unit that only attacks every other turn anyway, and (2) needlejets might have the range to attack undefended targets such as terraforming or formers/crawlers near the border; copters with reduced range are only useful for defensive roles, which is a lot easier to defend against.

Quote
Dropping copter mobility is a start, but it has little effect on defensive roles for Copters.  If I wait until a stack is 1-2 squares away from a city being defended, I can get 4-6 attacks from a single 8 speed chopper and still land safely.

True, but at that point why not just use infantry or rovers to attack even more cost-effectively, and if you don't kill them all the first turn...that's what perimeter defenses/tachyon fields are for.

Quote
This is a lot different than the attack every other turn from a Needlejet, which is then vulnerable to counterattack.  Choppers are why air power is widely considered overpowered, and they can still be overpowered even with only 8 speed.

Only 8 speed would leave them overpowered, but I think that only 8 speed plus a small boost to AAA (making it come before choppers on the tech tree and reducing its cost would probably be enough) plus a boost to defense in general (which would both reduce their attacking effectiveness and thus ability to make multiple attacks safely in some circumstances, as well as reduce the threat posed by allowing offensive units to reach your base and thus make it less important to kill them quickly) would leave them useful but not overpowered.

Another thing you might not have considered: Once fusion reactors come into play, it is often possible to add a small amount of armor to offensive units without increasing the cost.  Now, 2 or 3 defense against 8 attack isn't going to win you the battle, but it will deal a few points of damage in losing, and that places a limit on how often a chopper can attack before it has to retreat or face a substantial chance of being destroyed.

I think the real issue, though, is that you're still thinking in "offense is supreme" mode, and multiattack is indeed irreparably broken under those circumstances.  But getting out of that mode at least somewhat will greatly weaken the effectiveness of multiattack, and thus of copters.  (Even if it's still be necessary to kill enemies in your territory before they can target your formers and crawlers, as long as they can't take bases without an overwhelming advantage, they have no reason not to use defensive units for that purpose...and copters aren't that great against defensive units even without AAA and are horrible against defensive units with AAA).
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 19, 2013, 06:05:46 PM
There are a lot of issues with combat and combat modifiers.  It goes beyond just Copters, though they are the worst.  Almost all unit types are not viable.  In fact there's only 3 unit types that are useful all game.

Rovers (obsolete with Needlejets)
Needlejets (obsolete with Copters)
Copters
Once air comes out you run a few capturing rovers, and probes to smash defenses.

I know each new unit chassis should be important, but each new unit type makes the old completely obsolete.  +Move is an undervalued thing in mineral cost (I think we all agree on that).

Now I'd start with balancing infantry vs infantry battles.  With the +25% to base, I think they intended you to 'break' a base with infantry.  Unfortunately, with infantry having double the attack as defense, this doesn't work.  You can never get infantry next to a base safely.  So even at a basic level it's broken.

I feel that they really got the whole combat system backwards in SMAC.  Armor values should be equal to or greater than weapon values.  And then, bigger modifiers should be on the attack of a unit.  For example, give infantry +100% attack to base.  Rovers, +100% in the open (attacking from a base would not count).  Air units, +100% against fast units and non AAA-ships.  This sort of thing.  This creates more of a 'I make this unit to kill this unit' dynamic.  Likewise, rovers/air should have armor on them for balance purposes.  Though the concept of scrimping on weapon/armor to get mineral cost down was a decent one, they made it much too costly to have both on a unit.  The only unit with even close to sane costs with both is infantry, and its still much too high.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 19, 2013, 06:20:18 PM
Now I'd start with balancing infantry vs infantry battles.

I would agree that that's the place to start, especially as once that's balanced, rover vs. infantry will be a lot easier.

Quote
With the +25% to base, I think they intended you to 'break' a base with infantry.  Unfortunately, with infantry having double the attack as defense, this doesn't work.  You can never get infantry next to a base safely.  So even at a basic level it's broken.

However, I think that can be fixed via tech tree manipulation.  If attack is only 1.5 times the defense until Probability Mechanics, then the attackers inside the base will still have a 1.5X advantage, but defenders in the base (assuming a perimeter defense) will have a 1.33X advantage even against rovers (more with ECM against rovers), so while the base defenders have the advantage (as it should be, to provide a level of stability to the game), there isn't such an advantage to rovers over infantry; certainly not enough to justify a 2X cost difference.

Quote
I feel that they really got the whole combat system backwards in SMAC.  Armor values should be equal to or greater than weapon values.  And then, bigger modifiers should be on the attack of a unit.  For example, give infantry +100% attack to base.  Rovers, +100% in the open (attacking from a base would not count).  Air units, +100% against fast units and non AAA-ships.  This sort of thing.  This creates more of a 'I make this unit to kill this unit' dynamic.

I don't think that's the right dynamic; that's a good dynamic for a war game, not so good for an empire builder.  I think weapon values higher than armor values, with the advantage being potentially negated when using more mobile units and reversed when attacking a base, would provide a gameplay that's focused more on empire-building but provides the stability and need for variety which is lacking in the current game.

Quote
Likewise, rovers/air should have armor on them for balance purposes.

Air shouldn't have armor on them; rovers can have armor on them, just not weapons as well without being very expensive.  Which I think is good, as it encourages variety of unit types...or would, if armor weren't less than weapons by so much.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 19, 2013, 09:06:00 PM
At 3:2 the optimal strategy would still be 100% offensive rovers I think.  Maybe less so if ECM is also powered up to 100%.  One thought might be to give infantry one of ECM or AAA for free, but  unable to put both on the same unit.

Defensive rovers are an interesting unit, and decent to sometimes have 1 of, if only because they can hit mindworms at 100% attack prep.  But they cost too much compared to infantry and generally I find Scout Police Rovers are better sentinels.

Is it more interesting to have 'offense' and 'defense' troops' rather than troops where you max both?  Perhaps.  My thought was that if you make weapon and armor relevant to every unit, that might make the tech choices a little more interesting.  Where armor doesn't help air, I think I'd still tend down the 'weapon' tree even if armor was boosted up.

I do agree with the lack of variety.  It's very dull as it is, you more or less make the same troops every game.  Do you feel the game is too war oriented?  I would say that war is quite destructive in SMAC.  If the AI managed and picked units better, being declared on by more than one AI would usually mean your death. 
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 19, 2013, 10:13:54 PM
At 3:2 the optimal strategy would still be 100% offensive rovers I think.

I don't think so.  At 3:2 in the field (and thus 5:3 to the defender's advantage when attacking a base with non-infantry) if the enemy uses only offensive rovers, then they can't take your bases without expending substantially more than you (and then they don't have defensive units there, so you can take them back fairly easily), so all they can do is try to go after your formers/crawlers/terraforming.  But that'll be difficult, as if they stick around in your territory without taking your bases, you have a natural "safe zone" (your bases) from which to attack them with your own rovers or even infantry (depending on how good your road system is and how well they're armored).

Quote
Defensive rovers are an interesting unit, and decent to sometimes have 1 of, if only because they can hit mindworms at 100% attack prep.  But they cost too much compared to infantry

They do tend to be a bit more expensive (until you get quantum or singularity reactor), but can help keep your force somewhat resistant to counterattacks without sacrificing mobility.

Quote
Is it more interesting to have 'offense' and 'defense' troops' rather than troops where you max both?  Perhaps.  My thought was that if you make weapon and armor relevant to every unit, that might make the tech choices a little more interesting.  Where armor doesn't help air, I think I'd still tend down the 'weapon' tree even if armor was boosted up.

Actually, armor does help air by keeping the enemy from taking your bases and killing your air units.  Not to mention that a lot of armor is on the track to Monopole Magnets, which I'm pretty sure can be quite useful.

Quote
Do you feel the game is too war oriented?

No, but if it were substantially more then it would be.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 20, 2013, 03:23:31 PM
Well you're always going to bring a probe to down the Perimeter Defenses.

I'm not so convinced the early game or the game in generally is actually even that unbalanced at 2:1, as I think more and more on this.  The problem is, few games are played at a high level between humans with a warring mindset.  The AI wars poorly, and humans prefer to build (logically, growth/teching is exponential so you get left in the dust to be too aggressive early). 

I had the thought that a more sensible way to balance might be around chassis cost only, rather than tweaking the movement and rules of everything which is a lot more complex.  That takes a lot more thought in terms of new tactics and their countertactics.

Here are some common early-mid game unit costs (pre-Fusion/Chaos/Silksteel):

Infantry:
1/2-ECM: 20
1/3-ECM: 20
1/3-AAA: 30

4/1: 20
6/1: 20

6/3: 50  (6/3 AAA/ECM is more than 50, not worth it)

Rovers:
2/1: 20
4/1: 30
5/1: 40
6/1: 50

Needlejets:
6/1: 40
6/1-AAA: 50

Copters:
6/1: 40
6/1-AAA: 50

The problem here is that air units get a flat quartering cost reduction.  This makes all 3 air units (8+1)/4 -> 2.25 factor, universally cheaper than rovers (2+1) -> 3 factor, or hovertanks (3+1) -> 4 factor.  Logically air should cost more than a rover or even a hovertank.

What if we change copter's movement cost factor to 30, and Needlejet to 18?  Then we get:

Needlejet
6/1: 80
6/1-AAA: 100

Copter
6/1: 120
6/1: 150

You'd still buy them, just not in mass quantities.  Then you can keep AAA as-is (100% with a cost: 1)
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on April 20, 2013, 08:27:46 PM
I agree, Nexii.  Aircraft are undercosted with respect to Rovers.

I like the fix for Needlejets.  I think Copters should cost double the amount of Needlejets (if they are not banned outright).

But if we change this, is it possible to make AAA no cost on Air units (like deep radar).  I see no reason for intercepters to cost more, since they take a 50% penalty to ground to compensate for the double air attack.

As long as we are mucking with things, I think armor on Hovertanks should be costed the same as Infantry.  As it is, it is prohibitively expensive to put armor and a weapon on a Hovertank.  Which seems absurd, since tanks for known for armor.  Yet, in a real game, nearly all of my Hovertanks are unarmored.

I am OK leaving Rovers alone.  But by the time you have figure out the tech for Hovertanks, surely we have figured out how to make a decently priced mobile armor unit.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 20, 2013, 10:39:17 PM
Agreed with the Air Superiority on air units.  I gave the reasons on why I feel they are overcosted earlier.  The -move of AS is as much a penalty as the -50% to ground, IMO.  Actually they don't get double (in most cases) vs air, though they probably don't need it where non-AS air units will have 1 armor?  Giving the AS attacker double against the defender, might imbalance things though.  Scrambling would always be a good thing for the aggresor, i.e. you'd send an interceptor in at their formers, and hope the defending jets fight you.  Unless of course the defender got the +100% AS on defense too.  But again thats redundant.

http://strategywiki.org/wiki/Sid_Meier's_Alpha_Centauri/Units (http://strategywiki.org/wiki/Sid_Meier's_Alpha_Centauri/Units) - good source for unit cost, it seems close/accurate

EM do you ever make infantry with both attack and defense?  I find that its overcosted but maybe you find it has a niche role?  You can get a x/1 infantry and a 1/y infantry for less cost as an x/y infantry.  Plus the defensive special abilities on the 1/y will be cheaper %-wise.

Changing "S" in the formula is easy, it's just the "Cost" variable in alphax.txt.  Yitzi would have to comment on how much harder it is to edit the hardcoded parts like the 1/4 air cost reduction.

On a side note Gravships also need a raising in their 'move' for cost purposes.  Likely to somewhere between Copters and Needlejets.  I haven't analyzed foils but I feel they're also overcosted, and are also tilted towards 'all attack' or 'all defense' units.

If I put Needlejet to 17 (17+1->18/4 - 4.5 factor), and Copter to 35 (35+1->36/4 = 9 factor), we get:
6/1 Needlejet - 70
6/1 Copter - 140

The thing is though in the later game, you'll see copters and other 'expensive units' not have the same gap in cost vs rovers.  This is because the better reactors cut down costs non-linearly.  I'm not sure if this was intended, and it's something else to consider.  It impacts all units, and although it makes units with both attack and defense more affordable, it also has the effect of making the best chassis cost a lot less.

Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 21, 2013, 03:58:03 AM
Well you're always going to bring a probe to down the Perimeter Defenses.


On the other hand, the other guy will have probe teams to defend against them...and he can put an infantry chassis on his, making them cheaper.

Quote
I'm not so convinced the early game or the game in generally is actually even that unbalanced at 2:1, as I think more and more on this.  The problem is, few games are played at a high level between humans with a warring mindset.  The AI wars poorly, and humans prefer to build (logically, growth/teching is exponential so you get left in the dust to be too aggressive early).


On the other hand, if you take over enemy bases, that's also a form of growth...and if the bases are fairly high population/infrastructure (you'll lose some facilities, but probably not all), it can be a lot cheaper than building your own bases (assuming you have the POLICE rating to keep the captured bases under control).

Quote
I had the thought that a more sensible way to balance might be around chassis cost only, rather than tweaking the movement and rules of everything which is a lot more complex.


Well, rovers are already over twice the cost of infantry, and hovertanks are fairly late, so it really just becomes a question of air units.

Quote
Here are some common early-mid game unit costs (pre-Fusion/Chaos/Silksteel):


Keep in mind that that's with the current tech tree.  If you get defense 4 at the same time as ECM (as I'm proposing) and reduce AAA to ECM's cost (also a good idea IMO), that would affect things.

Quote
Infantry:
1/2-ECM: 20
1/3-ECM: 20
1/3-AAA: 30

4/1: 20
6/1: 20

6/3: 50  (6/3 AAA/ECM is more than 50, not worth it)

Rovers:
2/1: 20
4/1: 30
5/1: 40
6/1: 50

Needlejets:
6/1: 40
6/1-AAA: 50

Copters:
6/1: 40
6/1-AAA: 50

The problem here is that air units get a flat quartering cost reduction.  This makes all 3 air units (8+1)/4 -> 2.25 factor, universally cheaper than rovers (2+1) -> 3 factor, or hovertanks (3+1) -> 4 factor.  Logically air should cost more than a rover or even a hovertank.


You generally won't want AAA on an air unit; air superiority makes sense, but AAA is for defensive units.

More to your point...I think the idea is that the extra movement and slightly lower cost (as compared to rovers) is supposed to be counteracted by not being able to go far from your bases/airbases, so it can't be used to threaten assets away from the border.  Plus, of course, by AAA giving +100% whereas ECM only gives +50%.

A 1/4/1 ECM (and thus 1/4/1 AAA if AAA is changed to cost the same as ECM) costs only 30, less than the 6/1 needlejet that the AAA version would defeat quite easily.  At 1/3/1, it'd cost only 20, half the cost of the needlejet.

Air does substantially better against rovers, but rovers in turn can threaten infantry by outmaneuvering them and not granting as extreme a defensive bonus.  (Air can't outmaneuver infantry in the same way, as they have to stay fairly close to an airbase/base that can't move at all.)

Quote
What if we change copter's movement cost factor to 30, and Needlejet to 18?  Then we get:

Needlejet
6/1: 80
6/1-AAA: 100

Copter
6/1: 120
6/1: 150

You'd still buy them, just not in mass quantities.  Then you can keep AAA as-is (100% with a cost: 1)


Those are really expensive; at that cost, they're not really worth using except under extremely specific circumstances (undefended, perhaps even non-combat, units with no SAM units around.)  On the other hand, Fusion reactors would then throw things off horribly.

But if we change this, is it possible to make AAA no cost on Air units (like deep radar).  I see no reason for intercepters to cost more, since they take a 50% penalty to ground to compensate for the double air attack.


Making SAM have the same cost as deep radar does seem to make sense.  Good idea.

Quote
As long as we are mucking with things, I think armor on Hovertanks should be costed the same as Infantry.  As it is, it is prohibitively expensive to put armor and a weapon on a Hovertank.


Armor alone isn't; it's only armor+weapons that is.

Quote
Which seems absurd, since tanks for known for armor.


And the result is quite expensive; in the real world we do that because there's no such thing as "attacking units only use their weapon values, defending units only use their armor values", but Alpha Centauri does have such a rule.

http://strategywiki.org/wiki/Sid_Meier's_Alpha_Centauri/Units (http://strategywiki.org/wiki/Sid_Meier's_Alpha_Centauri/Units) - good source for unit cost, it seems close/accurate


A full formula can be found here (http://alphacentauri2.info/index.php?topic=3038.new#new) (reply #8).

Quote
Yitzi would have to comment on how much harder it is to edit the hardcoded parts like the 1/4 air cost reduction.


Not too difficult, but they'd have to be optional which would increase the difficulty somewhat.  Doable, but it'll have to wait its turn.

Quote
On a side note Gravships also need a raising in their 'move' for cost purposes.  Likely to somewhere between Copters and Needlejets.  I haven't analyzed foils but I feel they're also overcosted, and are also tilted towards 'all attack' or 'all defense' units.


I think that a decrease to move might be a better choice for gravships more than a cost increase.  That way, a fleet of gravships isn't as fast as needlejets...but it doesn't need bases and so can move toward your territory.

Quote
The thing is though in the later game, you'll see copters and other 'expensive units' not have the same gap in cost vs rovers.  This is because the better reactors cut down costs non-linearly.  I'm not sure if this was intended, and it's something else to consider.  It impacts all units, and although it makes units with both attack and defense more affordable, it also has the effect of making the best chassis cost a lot less.


Yes...that's a large part of why I don't think cost is the right way to handle air.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 21, 2013, 05:51:16 AM
Re: Perim defenses.  Yes it's true that a lot of wars come down to whether you can hold the enemy at a single chokepoint.  Generally if there's only one base in reach, it is quite difficult to break.  As well CDF and CBA make you more or less unbreakable at a single choke city.

Air does need a higher minimum cost and a higher overall cost.  Perhaps not as dramatic increase in cost as I had above.  Especially not as big an increase if AAA is made free (I assume you mean -1 modifier like ECM, in that its free on defense units only).  Having to take ECM and AAA on most of your infantry not made for cities, I think is a good thing. 

There are many points in the game (New reactors, when air first is available, and after Centauri Psi) where 100% air would still counter AAA infantry, even if AAA is made free.  Where air counters rovers, attack infantry, foils, cruisers, and anti-air (on par) also, you'd largely still see 100% air-to-ground copter armies.

For example, Chaos Fusion Chopper is 30 in cost.  That's absurd, it has the same cost as a 1/4/2*AAA-ECM infantry, with free AAA.  Even ignoring standard combat, if you suicide that into 2-3 crawlers or formers it pays off.  You can even make these even into the endgame.  This is why I'm also in favor of letting non-combat units (formers and crawlers specifically) get armor for free.  Granted, they would perhaps need to be changed to -75% combat power to compensate.

The way reactors reduce costs really messes with overall balance, and a single minimum unit cost across all chassis types is a big part of this.  I feel that reactors were plenty strong without also reducing unit costs in the way that they do.  A better formula would have made the later weapons/armor a bit less costly instead.  And then a modifier based on chassis alone rather than the strange move+chassis hybrid system they had.
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on April 21, 2013, 06:26:22 AM
I believe air combat only uses the attack strength of the defender, so armor is of no value on an air unit, except maybe defending against non-air units.

I see no reason that Hovertanks should be penalized so heavily for having both armor and weapons.  I think it should be treated exactly like infantry.

What really needs help is ships.  Armor on transports (and other non-combat ships) should be much less expensive.  AAA should be free on ships, just like air superiority should be free on aircraft, because ships are sitting ducks for aircraft since they do not get a bonus for terrain.  Sea terraformers should also be cheaper.  Sea colony pods are costed OK because they effectively include a free Recycling Center.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 21, 2013, 07:48:34 AM
Re: Perim defenses.  Yes it's true that a lot of wars come down to whether you can hold the enemy at a single chokepoint.  Generally if there's only one base in reach, it is quite difficult to break.  As well CDF and CBA make you more or less unbreakable at a single choke city.

Not quite unbreakable; a sufficient resource advantage will do it.  But yes, a single chokepoint does make it very difficult.

But let's say one player doesn't have the chokepoint.  (Let's use the example of Morganites and University, as they're both fairly neutral in direct combat ability, and say Morgan lacks a chokepoint.)  Now the University uses air units (as that's what we're discussing) to take Morgan Metagenics.  Even if it's not a chokepoint, air units still are fairly weak against AAA, and the base can have an aerospace complex as well to make it even harder, so say that they spend 3 times as much resources to take the base as the Morganites lose in defenders.  Now, the University can still do that; air units are fast, so they can concentrate 5 or 6 times as much force on one base as the Morgans have to defend it, but they do lose quite a bit more.

Now the Morganites can try to take the base back...they, however, are trying to take a base that's in their territory, connected to their road/maglev system, so they have the freedom to use infantry.  Infantry still is at a disadvantage when attacking a base with perimeter defense, but not as much, so say the Morganites lose 1.5 times as much as the University does.  Still, at the end of the day, they've come out ahead unless the University defends the base with 4 times as much as the Morganites did...possibly doable (though difficult) with one captured base, unfeasible with a significant amount.

Quote
Air does need a higher minimum cost and a higher overall cost.

Minimum cost is difficult to do.  But why is an increase in cost any better than a decrease in effectiveness?  Either one increases the number of minerals' worth of air units you need to attack a target...

Quote
I assume you mean -1 modifier like ECM, in that its free on defense units only

Indeed.

Quote
Having to take ECM and AAA on most of your infantry not made for cities, I think is a good thing.

And it doesn't hurt for infantry either.

Though only the defensive units would get it.

Quote
There are many points in the game (New reactors, when air first is available, and after Centauri Psi) where 100% air would still counter AAA infantry, even if AAA is made free.  Where air counters rovers, attack infantry, foils, cruisers, and anti-air (on par) also, you'd largely still see 100% air-to-ground copter armies.

Let's take those one at a time:
1. New reactors: Fusion is the big one here (quantum is only a 33% increase to effectiveness, and I think it may be advisable to double the benefits of AAA and ECM with Probability Mechanics anyway to offset the fact that attack later in the game is twice the corresponding defense).  Now, it's certainly a substantial advantage, but it applies to everything; your rovers infantry also get an advantage with it.  So the real question is whether fusion reactors themselves are too overpowered.

I think not, simply because of where they fall in the techtree; consider some of the other things that you might get instead (particularly if playing by game-normal rules, i.e. blind research, which limits beelining):
-Monopole Magnets.  Arguably the best mobility tech in the game.  And you know what they say about the importance of logistics...
-Orbital spaceflight.  While I would move the best satellite in the game to substantially later, replacing it with a weaker satellite (say, Orbital defense pod) would still give the advantage of seeing the whole map...and of course missiles are useful too.
-Retroviral Engineering.  +50% to minerals is quite useful, especially in wartime.
-MMI.  Speaking of copters...

2. When air first is available: This is why it's not enough just to make AAA cheaper; it also has to come earlier in the tech tree than air units do.  I think making Advanced Military Algorithms a prerequisite for Doctrine:Air Power (replacing Doctrine:Flexibility) is a good way to do that.

3. After Centauri Psi: No, no, and absolutely not.  Psi units ignore base defenses, but they do not ignore AAA, and they only get a 3:2 attack ratio anyway; worse yet, those infantry can be made AAA/trance, and they'll absolutely demolish psi air attackers.

Quote
For example, Chaos Fusion Chopper is 30 in cost.  That's absurd, it has the same cost as a 1/4/2*AAA-ECM infantry, with free AAA.

Yes, but it's a lot higher on the tech tree.  The Chaos copter requires 30 techs, whereas the 1/4/1*2 infantry (assuming we go with my idea of moving defense-4 down to advanced subatomic theory) needs only 20.  A better analogy would be missile copters vs. 1/4/1/*2, or chaos copters vs. 1/5/1*2 silksteel defenders.

Quote
Even ignoring standard combat, if you suicide that into 2-3 crawlers or formers it pays off.

Yes, copters are good for that...but in that case, wouldn't it be easier to make 4/1/8*1 copters for only 20 minerals?  The fact is, keeping your formers and crawlers that close to your enemy's bases without defending them is a fairly stupid idea.

Quote
This is why I'm also in favor of letting non-combat units (formers and crawlers specifically) get armor for free.  Granted, they would perhaps need to be changed to -75% combat power to compensate.

And then how would you make "armored formers" for use in battle zones?  (Note that if they have armor they don't even take the -50% combat power.)  Better to have normal crawlers/formers be vulnerable (and then whoever owns them just needs to keep them out of the line of fire), and then if you want to pay the extra cost to armor them you can.

Quote
The way reactors reduce costs really messes with overall balance, and a single minimum unit cost across all chassis types is a big part of this.

So...perhaps increase the minimum cost, but apply the "infantry is half cost" after that?

I believe air combat only uses the attack strength of the defender, so armor is of no value on an air unit, except maybe defending against non-air units.

I think it does get more complicated than that.

Quote
I see no reason that Hovertanks should be penalized so heavily for having both armor and weapons.  I think it should be treated exactly like infantry.

So only rovers would get +2 rows for having both armor and weapons?

Quote
What really needs help is ships.  Armor on transports (and other non-combat ships) should be much less expensive.

Let's see how much it costs for armored transports...
Anything below defense 5 costs 1 row more than unarmored transports, or the minimum of 30 with fusion reactors.
Defense 6 or 8 (or psi for that matter) costs 2 rows more than unarmored; with fusion it's 30 for foils and 40 for cruisers.
Defense 10 or 12 with a fusion reactor is 40 for both foils and cruisers
So simply removing the "+1 row for both weapon and armor cost above 1" rule for noncombat units would probably be the best way; it'd have to be made optional, but shouldn't be too hard.

Quote
AAA should be free on ships, just like air superiority should be free on aircraft, because ships are sitting ducks for aircraft since they do not get a bonus for terrain.

How about defensive ships (which are actually fairly cheap; a 1/12 cruiser is only 40 with a fission plant; even infantry cost 60) get AAA for free, whereas offensive ships (which can't use it as much anyway) don't, and might even have it cost more?  (Transports would be "defensive" at armor 5 or more.)

Quote
Sea terraformers should also be cheaper.

Agreed, but in exchange the thermocline transducer should be delayed on the tech tree; I think moving the transducer to Planetary Economics and the subsea trunkline to Advanced Ecological Engineering is a good way to do it.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 21, 2013, 10:53:57 AM
Centauri Psi is game breaking because it gives you Dissociative Wave, which lets you ignore ECM/AAA/Trance.  Gas Psi Choppers crush everything, and +Planet/Dream Twist only add onto this.  There isn't a counter, cheap normal units don't really work (due to min unit costing).  But if they do that then you just mix in some copters with conventional weapons. 


For examples, it's highly dependent on tech-level (especially reactors) as to how much it costs to break a base with air (or ground for that matter).  What I would prefer to see is a more consistent balance throughout the game. 

Quote
So...perhaps increase the minimum cost, but apply the "infantry is half cost" after that?

Maybe.  Halving infantry is a lot like doubling everything else (which impacts air).  I think you'd see mostly infantry and air in this case, and not much rovers/hovertanks.  It would really depend what else was tweaked.

What I don't like about the current costing model is that often an increase in weapon is actually a decrease in effective attack/defense per mineral spent.  Sometimes due to rounding, but more often to the wonky way in which reactors affect costs.  The efficiency of weapons/armor should go up on their own accord, and not with reactors (which already double/triple/quadruple both attack and defense).  I'll see what I can come up with.  I don't feel that units like Impact Copters should be relevant once you've well surpassed that tech.  Also they can travel 2 turns out, and hit those former/crawlers from 16 tiles away even if they are hard capped to 8 movement per turn.  That's usually pretty far into your territory.
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on April 21, 2013, 12:23:57 PM
Yes, the idea is for only Rovers to be penalized for having offense & defense; that tech improvements get rid of this limitation for Hovertanks.

I like the idea of defensive ships being able to add AAA for free, as long as it includes armored non-combat vessels (like transports).

Maybe we can add that offensive ships can add air superiority for free if they choose.

My multiplayer games never get to Centauri Psi, so I don't have much experience with it.  But if gas affects combat, one simple fix is to for gas to have no effect on Psi combat.  Then it is just straight morale vs morale; secret projects can still help, but most people don't prioritize secret projects that affect psi combat.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 21, 2013, 04:03:46 PM
Centauri Psi is game breaking because it gives you Dissociative Wave, which lets you ignore ECM/AAA/Trance.  Gas Psi Choppers crush everything, and +Planet/Dream Twist only add onto this.  There isn't a counter, cheap normal units don't really work (due to min unit costing).  But if they do that then you just mix in some copters with conventional weapons. 

Dissociative wave is fairly broken, I'll grant that (though they're useless against Trance, since they're not allowed on psi units.)  It does need to be either increased in cost or banned for air units; I'd favor the latter, on the basis that the equipment is too bulky for aircraft.
Gas psi choppers are also an issue; perhaps ban gas on psi units?  (That said, psi choppers with the Dream Twister will still be very strong against non-trance units...but all projects are strong in their niche.)

Quote
For examples, it's highly dependent on tech-level (especially reactors) as to how much it costs to break a base with air (or ground for that matter).  What I would prefer to see is a more consistent balance throughout the game. 

Yes, perhaps reactor cost effects need to be reworked.  I'll start a thread in Modding for ideas.

Quote
What I don't like about the current costing model is that often an increase in weapon is actually a decrease in effective attack/defense per mineral spent.  Sometimes due to rounding, but more often to the wonky way in which reactors affect costs.

The rounding will probably be compensated for by the fact that you have a better chance of surviving to get healed; the reactor effects are more of a concern.  Perhaps reactor effects need to be decreased, and the reactor-based minimum also decreased.

Quote
Also they can travel 2 turns out, and hit those former/crawlers from 16 tiles away even if they are hard capped to 8 movement per turn.  That's usually pretty far into your territory.

Ah, I didn't count on "suicide" meaning "take damage the first turn".  Of course, that leaves them vulnerable to counterattack (or simply pulling formers and crawlers into the base/another area for a turn or two) after being seen by sensors.  It also means that they'll die easily to scrambled interceptors.

Yes, the idea is for only Rovers to be penalized for having offense & defense; that tech improvements get rid of this limitation for Hovertanks.

Whereas I see it as that the tech is all going into making hovertanks possible at all.

Quote
I like the idea of defensive ships being able to add AAA for free, as long as it includes armored non-combat vessels (like transports).

Depends how armored and what type of combat vessels.  How about:
Transports and probes can get AAA for free with silksteel armor.  Formers need neutronium armor (Matter Compression) for free AAA, crawlers need antimatter (Matter Editation), and colony pods need stasis (Temporal Mechanics).  That should be fairly easy to do.

Quote
Maybe we can add that offensive ships can add air superiority for free if they choose.

Now that I think of it, it'd be way too difficult to make air superiority free for interceptors (which I think we agreed is a good idea) but not for ships.  I think let's just make it cost the same as Deep Radar: Free for air and sea, not free for land.  So any ships would be able to add free Air Superiority (though purely defensive ships would not be so certain to win).
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 21, 2013, 05:05:56 PM
What do you mean never gets to Centauri Psi?  Aren't Hovertanks on the same level of tech as Psi (roughly)?  I'm curious as I haven't actually played a lot of MP for typical tech beelines late.  I'm just thinking more in terms of a war than build mindset.  Maybe I'm in the minority that likes the war side of the game as much as the builder side.  It's just that there's a lot less strategy in it, which is a shame considering the brilliant concept of the unit design shop.

It's Wave that is game-breaking more than Gas.  Gas is -25% enemy morale for +25% cost.  Against higher morale defenses taking Gas actually downs the efficiency of Wave Choppers a bit.  But it's not a huge effect and generally worth taking so you can more reliably preserve choppers.  Since their efficiency ratio is so high, you don't need to put a lot into minerals to have a stronger army.

Psi combat (and native life combat) is a much harder thing to balance.  Generally native life is not that good because you can't put modifiers on it.   


A much simpler and intuitive unit cost formula would be:

M = (W + A + C) * 10

M: the cost of the unit in minerals
W: weapon factor.  would scale from 1 for early game weapons to 6 for the late-game weapons.
A: armor factor.  would scale from 1 for early game armors to 4 for the late-game armors.
C: chassis factor.  0 for infantry, 1 for rover, 2 for hovertank, 3 for needlejets and gravships, 4 for copters. 1 for foil and 2 for cruiser. 

Since weapons have double the values of armor, and a 3:2 cost ratio, this would give a 3:2 effectiveness ratio of attack to defense (the same as PSI).

For rovers, foils, cruisers, needlejets, copters, and gravships, add (W-1)*(A-1) to the above cost
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 21, 2013, 05:26:54 PM
It's Wave that is game-breaking more than Gas.

Indeed, which is why it needs to either have its cost doubled (at a minimum) or be banned on air units.

Quote
Gas is -25% enemy morale for +25% cost.  Against higher morale defenses taking Gas actually downs the efficiency of Wave Choppers a bit.  But it's not a huge effect and generally worth taking so you can more reliably preserve choppers.  Since their efficiency ratio is so high, you don't need to put a lot into minerals to have a stronger army.

Yes; I'd say let air units have both types of gas (nerve and soporific) but not Wave; psi units should be unable to get any of the three (nerve gas is the only one they can get now IIRC).

Quote
A much simpler and intuitive unit cost formula would be:

M = (W + A + C) * 10

M: the cost of the unit in minerals
W: weapon factor.  would scale from 1 for early game weapons to 6 for the late-game weapons.
A: armor factor.  would scale from 1 for early game armors to 4 for the late-game armors.
C: chassis factor.  0 for infantry, 1 for rover, 2 for hovertank, 3 for needlejets and gravships, 4 for copters. 1 for foil and 2 for cruiser. 

Since weapons have double the values of armor, and a 3:2 cost ratio, this would give a 3:2 effectiveness ratio of attack to defense (the same as PSI).

For rovers, foils, cruisers, needlejets, copters, and gravships, add (W-1)*(A-1) to the above cost

It'd be simpler, but would encourage combining the best armor with the best weapon on infantry and tanks (as putting high armor on a high-weapon unit is a lower percentage increase to cost than putting it on a 1-attack unit), whereas discouraging it makes for more interesting unit choices.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 21, 2013, 05:39:46 PM
Yea well the intent was that infantry and tanks with both weapons and armor would be viable.  Right now they're not.  Keep in mind, a lot of these free abilities like AAA/ECM would only apply to defensive infantry and tanks, so they wouldn't have that benefit.  You could also apply the speed increase cost on tanks, if they're too good (I'd argue EM's suggestion would mostly make rovers obsolete, whether this is good or not is debatable).

It should be that either weapons and armor start at 0 and go to 6 (with a factor of W*A), or chassis costs are all 1 less than I listed.   The former probably makes more sense.

An infantry with the best weapon and armor would cost 110.  Alternatively, you can buy a defensive infantry for 50 (with free modifiers) and an attack infantry to go with it for 70.  I think that's a pretty good balance.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 21, 2013, 06:15:36 PM
Yea well the intent was that infantry and tanks with both weapons and armor would be viable.  Right now they're not.

Well, you didn't really think something called "shock troops" or "behemoth" would be cheap, did you?  (Although shock troops are still substantially cheaper than facilities of the same tech level).  I think having to choose a weakness for each unit makes for a more interesting game.

Quote
An infantry with the best weapon and armor would cost 110.  Alternatively, you can buy a defensive infantry for 50 (with free modifiers) and an attack infantry to go with it for 70.  I think that's a pretty good balance.

If you want "best of everything" units to be fairly common.  I think that the game works better if they aren't.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 21, 2013, 06:35:59 PM
A few things to consider:
If you're attacking and you lose your offense infantry, you're only down 70 minerals instead of 110. 
If you lose your defense garrison, the enemy has to commit an additional unit to kill off your offense unit.  This often sets up a counterattack.
It's much more vulnerable to rovers/air, unless you put ECM/AAA on it.  And in this case the 'good at everything' infantry would cost a lot more than 110.  If AAA has -1 for a modifier and ECM, each of these has a cost of 2 when weapon is double armor.  Not very viable to pay 220 for an all around infantry.

I do see some upsides to the all around infantry:
- 10 less mineral cost
- Less transport space
- Less support required
- More suited to Drop and other Cost: 1 modifiers not skewed to defense.  Where the defense unit would have to give up one of ECM/Trance/AAA to follow along.

Although if you really hated all around infantry it could also have the option of (W*A) applied.
Title: Re: SE choices for AI - suggestions
Post by: Earthmichael on April 21, 2013, 07:08:27 PM
Just as Cruisers usually make Foil obsolete, I think Hovertalk should do this to Rovers.

I think the roles could be better defined with the special abilities, not the weapon and armor.  For example, the default unit would be your best reactor, best weapon, and best armor.  But for certain special roles, such as air defense, it might be costed so that it is cheaper (or even free) to add air defense to a defensive unit.  So there would still be a reason to create specialized units.

Most of my games do not get to Hovertanks, either.  For a typical game on the vets map, the game is usually over by turn 150 or less.  On a really huge map, I could see getting to higher techs before things get decisive.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 21, 2013, 07:25:14 PM
If you wanted to actively encourage max weapon/armor on all units, you could instead make the formula:

M = (max (W, 1.5*A) + C) * 10

With no (W*A) modifier on faster units.  You would also need to increase W/A cost amounts slightly.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 21, 2013, 08:12:59 PM
A few things to consider:
If you're attacking and you lose your offense infantry, you're only down 70 minerals instead of 110. 
If you lose your defense garrison, the enemy has to commit an additional unit to kill off your offense unit.  This often sets up a counterattack.
It's much more vulnerable to rovers/air, unless you put ECM/AAA on it.  And in this case the 'good at everything' infantry would cost a lot more than 110.  If AAA has -1 for a modifier and ECM, each of these has a cost of 2 when weapon is double armor.  Not very viable to pay 220 for an all around infantry.

I do see some upsides to the all around infantry:
- 10 less mineral cost
- Less transport space
- Less support required
- More suited to Drop and other Cost: 1 modifiers not skewed to defense.  Where the defense unit would have to give up one of ECM/Trance/AAA to follow along.

Quote
Although if you really hated all around infantry it could also have the option of (W*A) applied.

It's not a question of "hate all around infantry", it's that discouraging all-around infantry is neither broken nor unbalancing, and so I don't think it's worth changing it from the original rules.

Just as Cruisers usually make Foil obsolete, I think Hovertalk should do this to Rovers.

They already do, for the same reason.  But at least foil is cheaper (and so is often better for formers, crawlers, and maybe colony pods); rovers should at least have that advantage.

Quote
I think the roles could be better defined with the special abilities, not the weapon and armor.  For example, the default unit would be your best reactor, best weapon, and best armor.  But for certain special roles, such as air defense, it might be costed so that it is cheaper (or even free) to add air defense to a defensive unit.  So there would still be a reason to create specialized units.

Interesting idea, but it'd just be too much unnecessary change for me.  But feel free to add your in the thread in Modding, and when I get to adding varying formulae I'll probably add whatever's there.
Title: Re: SE choices for AI - suggestions
Post by: Kilkakon on April 28, 2013, 07:23:04 AM
What do people think about Gravships? Are they like ever used?
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on April 28, 2013, 03:38:11 PM
What do people think about Gravships? Are they like ever used?

I don't see why they wouldn't be, once the necessary tech becomes available.  They don't have multiattack like choppers, but still have good speed without needing a base or airbase nearby.
Title: Re: SE choices for AI - suggestions
Post by: Nexii on April 28, 2013, 03:47:09 PM
Gravships are also cheaper than Drop troops and defense foils for capturing bases.  And they're much faster than ordinary Rovers and Hovertanks.  So while Copters will still make up most of your force, Gravships have their place.

Also Gravship formers and colony pods are nice to have, once you can spare the extra cost.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on May 29, 2013, 11:42:09 PM
As I understand it, superdrones are only harder to suppress via psych, facilities turn them into citizens as easily as normal drones.

In the process of studying the drone rules, I found that this is not in fact true.  Facilities also take two units of drone suppression to suppress a superdrone, though they seem to work on them last.
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on May 30, 2013, 10:07:42 PM
As I understand it, superdrones are only harder to suppress via psych, facilities turn them into citizens as easily as normal drones.

In the process of studying the drone rules, I found that this is not in fact true.  Facilities also take two units of drone suppression to suppress a superdrone, though they seem to work on them last.

Well, I've investigated some more, and it is true, but doesn't show up in the final base display.  (It does show up in the psych breakdown, though; furthermore, it does not lead to actual drone riots.)  Thus, this raises the question: Is superdrones not being harder to suppress via psych a bug, or is the bug just that the drones in the main base screen display wrong?
Title: Re: SE choices for AI - suggestions
Post by: Buster's Uncle on May 30, 2013, 10:29:33 PM
...It's a kind of bug in general that drone stuff is so impossible to anticipate and understand, and this particular thing is definitely part of that...
Title: Re: SE choices for AI - suggestions
Post by: Yitzi on May 30, 2013, 10:33:00 PM
Once I finish looking over the code, I plan to post a full description of the rules, including options I plan to include in version 2.1.
Templates: 1: Printpage (default).
Sub templates: 4: init, print_above, main, print_below.
Language files: 4: index+Modifications.english (default), TopicRating/.english (default), PortaMx/PortaMx.english (default), OharaYTEmbed.english (default).
Style sheets: 0: .
Files included: 33 - 892KB. (show)
Queries used: 20.

[Show Queries]