Turning SMAX back into strategy game

[bvanevery]

Modding the reactor sizes doesn't seem to do anything.  I tried decimal numbers, no change.  I tried the 3, 4, 5, 6 sequence, no change.  Reactor sizes 1, 2, 3, 4 seem to be used regardless.

[Alpha Centauri Bear]

bvanevery,
By the way what does this mean in your alphax.txt?

30,  ; 12 Continent base   (Base size of a land mass seed)
60,  ; 24 Continent modif. (Increased size from LAND selection: x0, x1, x2)
0 ; 36 Islands          (Higher # increases island count)

[bvanevery]

It means I changed those settings from the original.  I suppose I could either add the notation "original" to make it more clear, or just delete the commented value and let people figure it out with a diff tool.  Not like I've got a running commentary about original values for anything else.

[Vidsek]

       Hello Tnevolin, I'm the infamous stealthy Vidsek

  As bvanevery mentioned, I have been working on merging his AI Growth mod with Yitzi's latest.   I had it functioning smoothly (as in no crashes or obvious bugs) as of AIG v. 119.   I haven't had the time to update it to AIG v. 1.22 as yet, but very much want to and will.

  The process is fairly simple, if somewhat time consuming in places. 
  First: Yitzi made no changes to the Faction Files, so I simply use bvanevery's.  I did make some changes to my personal custom faction's bonuses to keep it in line with the AIG philosophy and not be overly OP.

  Second: for most of the sections in alphax.exe there is either no change or you can simply replace Yitzi's with the AIG one, making sure to preserve line spacing and so on where it is important.
The one change I did make on my own was to line up everything in neat columns for ease of reading and comparing entries.  Since this only required adding or deleting blank spaces, I assumed it would not be a perversion of either Yitzi's or bvanevery's work.
The RULES section takes a bit more work, as Yitzi both added new content and rearranged the order of some of the existing lines.  I just kept a checklist as I went through Yitzi's alphax and made the changes.  And yes, it was easier for me to use Yitzi's as the base and modify it than the other way around.

  Third: Yitzi added the ability to adjust drone rules, unit costs, and the cost of switching between different types of production (plus a couple other minor things).
To keep the merged mod functioning closely to what bvanevery intended, I studied the options for each of those and tried to choose the one that most closely matched what happens in un-modded SMACX.

  Naturally, at several stages in this process, I fired up a game to check for crashes and malfunctions.  The only place I found a problem that wasn't due to a typo, was (as best as I can tell) with the "facility" Stockpile Energy.  When it was changed from Yitzi's cost of 2 to cost of 0 it caused a crash, so I left the cost at 2 (not a notable game-changer).  I think in more recent work, it mysteriously stopped doing that.  I will check it out further when I do the AIG v. 1.22/Yitzi merge.

   Now, in the (admittedly somewhat limited) game testing I did with the merged mods, I found the AI and game in general to function very similarly to what bvanevery described from his testing, and showed in his AARs.  It will definitely need much more testing, especially in the late-middle and late game stages I haven't yet had time to play and study.

   You are welcome to work up your own merge if you become impatient with my pace of getting this finished (just turned 65, so I've been mired in getting paperwork done for medical and retirement type stuff recently).
Hopefully what I've outlined above will be a help, if anyone needs more explanation or clarification, I'll be happy to give it.

[Alpha Centauri Bear]

Vidsek, that's great!
I was about to do the same just now but realized it is not just merging two files together. So I'm glad that you have already done all the work.


I definitely can upgrade to 1.22 if you share whatever you ended up with. I guess incremental change will be much easier.

One thing that struck me is the additional variables Yutzi added for faction bonuses. He rearranged their keywords too. So I was thinking that'll require checking if AIG faction files use same exactly keys as in alphax.txt file. That was my only concern.

[bvanevery]

One thing that struck me is the additional variables Yutzi added for faction bonuses. He rearranged their keywords too. So I was thinking that'll require checking if AIG faction files use same exactly keys as in alphax.txt file.

I did not change any keys.  I seriously doubt Yitzi would have changed existing keys, that would be an extremely daft thing to do.  He probably just added keys.

[Alpha Centauri Bear]

That what I thought. He added more and then rearranged them. However, if you were able to play with unchanged AIG faction files. Then probably appended them keeping backward compatibility.
Whatever is it, can you share you latest merge result?

[bvanevery]

Whatever is it, can you share you latest merge result?

I guess you're addressing Vidsek?  I've never merged Yitzi's alphax.txt with my own.  I've merely stared at differences using the WinMerge tool.

[Alpha Centauri Bear]

Whatever is it, can you share you latest merge result?

I guess you're addressing Vidsek?  I've never merged Yitzi's alphax.txt with my own.  I've merely stared at differences using the WinMerge tool.

I've done it already. Thank you.

[Alpha Centauri Bear]

Continuing discussion with PvtHudson started in http://alphacentauri2.info/index.php?topic=21013.msg116424#msg116424

You can build like 10 needlejects capable of knocking out enemy base defenders and just bomb-heal-repeat conquer them without casualties.

So can them. AIs even before ThinkerMod were capable of amassing needlejets and destroying every former, colony pod and crawler in range, then switching to base defenders.

For AI it is even worse as it doesn't know about indestructible army tactics and doesn't use it. Humans do and that's why they beat it every time.

Sorry, I don't catch. AI builds units with highest available attack, and it uses healing, doesn't it? But probably we should talk about this in your thread, not here.

Sorry about wording. It is quite difficult to explain these terms in short paragraph.

I am talking about two things those are different but can be used together.

1) Indestructible attacker. This happens when a faction possesses a unit with attack strength enough to knock out a defender with near 100% probability. If this unit can also escape retaliation then it returns to safety, heals and repeats. This usually applicable to needlejets and ships as there are not much units capable of attacking them from base. However land units can be that too as AI often does not counterattack weakly defended attacker near base for some reason. Bot AI and human can get such indestructible units with technology advantage.

2) Concentrated fire tactics. You gather attackers near the base but do not attack until you are sure you can clear it with one turn. Then you go. Then you can take it and avoid retaliation at least from this base. This tactics is used mostly by human and not AI. As AI usually does not "wait" if they have an opportunity to use the unit. Each AI unit acts more or less alone by its own program.

Summarizing above, I am saying that AI naturally can get indestructible units when they have weapon/armor advantage. However, it doesn't use concentrated fire tactics on purpose. To get to the point when it incurs no casualties it has to have both weapon/armor advantage and sheer number of units. So that even though each unit acts independently there are many units acting at the same turn. The only difference between AI and human is that human can achieve this level with lesser number of units. Where AI needs 20 needlejets and 10 ground units human can be effective with 5 needlejets, 2 rovers, and 1 probe. As a result of being more effective human usually wins overall.

Once again. I understand this is kind of a gray area. As with any exploits both healing and skewed combat odds benefit human more just because they consciously exploit them while AI does this unconsciously and, therefore, benefit less.

Usual disclaimer. I do not state this reasoning is 100% accurate. This is merely an opinion to support the discussion.

[bvanevery]

You gather attackers near the base but do not attack until you are sure you can clear it with one turn. Then you go. Then you can take it and avoid retaliation at least from this base. This tactics is used mostly by human and not AI.

This phenomenon occurs in every computer game ever made that doesn't have much in the way of AI brains, but instead relies on large resource advantages and the spawning of units to provide a challenge to the player.  Formally I call it "the spawning problem".  It is no different than a game of Gauntlet, with all the units coming out of some designated spawning box.  It just occurs in turn-based slow motion.  I first fully realized how this works when playing Panzer General II.  If you knock out the city, you make progress.  If you don't, it tediously pulls new units out of its ass.

Others have called this a "thresholding" problem.  The AI has a threshold of what it will or won't detect as a threat.  Humans are good at noticing these patterns and find ways to hang around just outside the range of perceived threat.  This of course works in real life as well, when hunting game, or ambushing an enemy.

[PvtHudson]

Continuing discussion with PvtHudson started in http://alphacentauri2.info/index.php?topic=21013.msg116424#msg116424

You can build like 10 needlejects capable of knocking out enemy base defenders and just bomb-heal-repeat conquer them without casualties.

So can them. AIs even before ThinkerMod were capable of amassing needlejets and destroying every former, colony pod and crawler in range, then switching to base defenders.

For AI it is even worse as it doesn't know about indestructible army tactics and doesn't use it. Humans do and that's why they beat it every time.

Sorry, I don't catch. AI builds units with highest available attack, and it uses healing, doesn't it? But probably we should talk about this in your thread, not here.

Sorry about wording. It is quite difficult to explain these terms in short paragraph.

I am talking about two things those are different but can be used together.

1) Indestructible attacker. This happens when a faction possesses a unit with attack strength enough to knock out a defender with near 100% probability. If this unit can also escape retaliation then it returns to safety, heals and repeats. This usually applicable to needlejets and ships as there are not much units capable of attacking them from base. However land units can be that too as AI often does not counterattack weakly defended attacker near base for some reason. Bot AI and human can get such indestructible units with technology advantage.

2) Concentrated fire tactics. You gather attackers near the base but do not attack until you are sure you can clear it with one turn. Then you go. Then you can take it and avoid retaliation at least from this base. This tactics is used mostly by human and not AI. As AI usually does not "wait" if they have an opportunity to use the unit. Each AI unit acts more or less alone by its own program.

Summarizing above, I am saying that AI naturally can get indestructible units when they have weapon/armor advantage. However, it doesn't use concentrated fire tactics on purpose. To get to the point when it incurs no casualties it has to have both weapon/armor advantage and sheer number of units. So that even though each unit acts independently there are many units acting at the same turn. The only difference between AI and human is that human can achieve this level with lesser number of units. Where AI needs 20 needlejets and 10 ground units human can be effective with 5 needlejets, 2 rovers, and 1 probe. As a result of being more effective human usually wins overall.

Once again. I understand this is kind of a gray area. As with any exploits both healing and skewed combat odds benefit human more just because they consciously exploit them while AI does this unconsciously and, therefore, benefit less.

Usual disclaimer. I do not state this reasoning is 100% accurate. This is merely an opinion to support the discussion.

1) I agree there is some problem, although, IMHO, you overstate the scale of it. In theory, in the SMAC ruleset attackers aren't 100% indestructible. Especially needlejets and gravships, that after bombing run are sitting ducks waiting for enemy interceptors and SAM-capable ships/land units. If you provide fighter escort, it's just 1:1 slugfest at best in the enemy turn, and you need to produce and support bombers AND escorts. Similar counter-attackers exist for other types of attackers. Worst offenders here are attack choppers, that can deal massive damage and withdraw to relative safety in the same turn. OP choppers are near consensus on the forums, and many mods reduce their range. In general, I believe the issue is not in unit specs per se, but in AI inability to counter-attack properly, and hence the solution should be AI improvement and not rule change. Disabling healing and, especially, buffing defenders will promote bloody grinding and make the game less enjoyable. Not mentioning balance bias to high-production players. IMHO, of course.
2) This is obvious AI problem. I think, partly it can be explained by AI's desire to reduce splash damage to its attacking troops. Indeed, splash damage from defending artillery and losses in attackers' ranks is the best counter to concentration tactic. Once again, it's AI's fault, not ruleset's. However, even with current AI there are glimpses of hope - its artillery will start duels with yours if in range, its air will attack your exposed reinforcements nearing assembly point, and amassing of forces in one nearby base is a prompt for AI's nuke.

You should play the game Pandora: First Contact. Its tactical AI can teach SMAC's one a few lessons. Just use expansion Eclipse of Nashira, not the base game.

[bvanevery]

OP choppers are near consensus on the forums, and many mods reduce their range.

Including mine.  Reduced to 6 movement.

Disabling healing

That's not .txt moddable so I don't know the actual testable consequences of this.

and, especially, buffing defenders will promote bloody grinding and make the game less enjoyable.

Armor is slightly buffed in my mod.  Armor in the original game is widely regarded to be pathetic to the point of uselessness.  My solution is armor and weapons strengths are matched at every Tier of tech.  So for instance, Silksteel Alloys and Nonlinear Mathematics are both Conquer 4 techs in my mod.  The end result is a lot of the game gets played with low caliber weapons and armor, which is fine by me.  There are other ways than raw weapon size to gain an advantage in battle, and my mod challenges the players to think somewhat about what those might be.

So, the consequences of "better armor" are fully testable right now, by playing my mod.  I doubt that's going to relieve all AI concerns raised in this discussion, but it does improve things IMO.  I also wouldn't make armor any stronger than I have.  I think it's important to move discussion about armor out of the realm of conjecture and ground it in fact, i.e. play my mod before having more debate about armor.

Or hey really, about any of the combat stuff.  Because when you change the fundamental pacing of the technology landscape, you don't have the same problems anymore.  I'm sure there are still problems, as I'm only .txt modding, but I'd like to think I've gotten as far as one can without .exe patching.  That's an important baseline IMO for "what should be done next".  I don't see the point in discussions without such baselines.  People have tried to fix a number of things over the years in various ways.  Don't debate plain SMAC, debate from the delta of some mod.

Not mentioning balance bias to high-production players.

What's a "high production player" ?  FWIW, only Domai gets an INDUSTRY bonus in my mod, and it's only +1.  INDUSTRY is overpowered, and already assigned egregious bonuses by Transcend difficulty anyways.  So I nerfed it a little. 

[Alpha Centauri Bear]

1) I agree there is some problem, although, IMHO, you overstate the scale of it. In theory, in the SMAC ruleset attackers aren't 100% indestructible. Especially needlejets and gravships, that after bombing run are sitting ducks waiting for enemy interceptors and SAM-capable ships/land units. If you provide fighter escort, it's just 1:1 slugfest at best in the enemy turn, and you need to produce and support bombers AND escorts. Similar counter-attackers exist for other types of attackers. Worst offenders here are attack choppers, that can deal massive damage and withdraw to relative safety in the same turn. OP choppers are near consensus on the forums, and many mods reduce their range. In general, I believe the issue is not in unit specs per se, but in AI inability to counter-attack properly, and hence the solution should be AI improvement and not rule change. Disabling healing and, especially, buffing defenders will promote bloody grinding and make the game less enjoyable. Not mentioning balance bias to high-production players. IMHO, of course.
2) This is obvious AI problem. I think, partly it can be explained by AI's desire to reduce splash damage to its attacking troops. Indeed, splash damage from defending artillery and losses in attackers' ranks is the best counter to concentration tactic. Once again, it's AI's fault, not ruleset's. However, even with current AI there are glimpses of hope - its artillery will start duels with yours if in range, its air will attack your exposed reinforcements nearing assembly point, and amassing of forces in one nearby base is a prompt for AI's nuke.

You should play the game Pandora: First Contact. Its tactical AI can teach SMAC's one a few lessons. Just use expansion Eclipse of Nashira, not the base game.

Thanks for feedback, PvtHudson. If you think you are arguing with me, you are not. In fact you are repeating all the same arguments just with other wording. Why we always discussing the wording??? Let's look at the big picture.
 

I repeat myself: this is a gray area. The indestructible attackers is not an exploit that is there for all to take. Some number of conditions need to align so it starts to manifest itself. Yes, it is all conditional and "indestructible attacker" is not 100% indestructible. If you remember, I said near 100% winning chance, etc. There are a lot of ifs. Obviously, these conditions also include the defender inability to organize proper and active defense to equalize chances, etc. What I am saying is that as faction technology levels and economical powers diverge there will be point when attacking side suddenly stop incurring casualties! This happens sooner or later. I observed it in many games and many other people did too. It's a nature phenomenon which you cannot close you eyes on. You see two AI faction fighting for long long time. And for some (could be pretty long) time neither side cannot advance significantly. Then, suddenly, one side starts advancing their borders at a highest possible speed. Moving units to next enemy base takes longer than actually capturing it. Yea, destroying whole faction with large territory may actually take some 20-50 turns just to move your military units across it. So some people may get an illusion that this happens not instantly but it doesn't really matter - the fate of loser is already sealed. Even temporary truce is just a delay. What bothers me is that losing faction spent some time to populate territory, grow up bases, build facilities, terraform, and protect this all. Now other faction gets it for free. The explosive conquest territory advancement is even faster than populating unclaimed territory with own colonies. That leaves you with only viable strategy and this is what all this post is against.

I do not absolutely mind anyone taking enemy territory if they so decided. That is a strategy game after all. I argue that benefit/cost of this strategy should be comparable or less than the same of other viable strategies (expand, grow, build, trade, ...). Combined power of others should be able to stop faction growing large by eating others. Currently as you reach the point where you just cakewalk through enemy territory public opinion doesn't matter at all. Seriously, repeating same strategy every time is ridiculously boring. Almost every time when I grow big everybody starts hating me and declaring war. This doesn't scare me a bit. I gather my forces and go kill one faction completely leaving my other borders unprotected. Then I turn to next faction and so on. By the time I dealt with first faction other may capture few of my relatively unprotected bases but this is like a drop in a bucket. If they cannot take my territory fast enough I just turn to next one and they all are done.

Now without indestructible army I wont be able to walk through enemy territory with the same army at a unit movement speed. I keep losing units and even with my strong economy I need to build them again and replenish at a front line. It gives other enough time to take my unprotected territory. So I need to keep some active defenders-attackers around my border which thinners my invading army and lowers my attacking potential even more. Now I think thrice whether I want a war against everybody. Politics starts to matter.

In other words, at some point in vanilla strong faction conquers heavily protected enemy territory faster than other faction(s) conquer its own lightly protected territory. It tumbles strategical game into race game and I don't like it. Again. I am not arguing that it shouldn't be like that. Somebody may like it. I am just building a mod that I would like more and share it with others interested in same.

[Alpha Centauri Bear]

That's not .txt moddable so I don't know the actual testable consequences of this.

No, unfortunately. We discussed this many times already. I mention this as a better but hardly reachable option. Without it I'll try to achieve same goal with other txt moddable means.

and, especially, buffing defenders will promote bloody grinding and make the game less enjoyable.

Armor is slightly buffed in my mod.

Don't be modest. It is twice as strong and comparable to contemporary weapon strength.
 
I actually think it's right and maybe it should be even slightly (20-30%) stronger that weapon to account for highly technologically advanced factions who will get better weapon anyway.