SMAX - The Will to Power - mod

[Alpha Centauri Bear]

Hmm. I see a lot of controversial responses on recent AI changes. Let's review them.

Thinker main goal was to improve AI and WTP continues on this path as well. I have heard a lot of complaints about some OP exploits humans can do but AI cannot. Teaching them to AI is usually the best way to resolve that problem. However, this is also the most difficult option. In absence of which, we usually resort to limiting such exploits to equalize human vs. computer capabilities. ICS was one of the most notorious one everybody tried hard to limit. Yet now when I was able to teach computer do the same, making it no more exclusive human exploit, it doesn't seem to be unconditionally liked. What is going on, guys? Were you hypocritical when you complained about it and demanded this problem to be solved for good? I have spent quite substantial amount of time and mental efforts on it. Was it for nothing? 😄🤔😢
Let me know if this is the right direction and/or something need to be done with spamming as well, regardless of AI success.

Keep in mind that neither Thinker nor WTP planned to put any limit on AI improvement posing themselves as challenging mod. So the argument that game becomes unbeatable on toughest level, while pretty valid for game attractiveness in general, is not applicable to this mod.

About AI disregarding defense while expanding without limits. That is true. Building more colonies obviously impacts combat units. This allowed me to amass an army and easily cut into neighbor territory. I plan to work on defense more. However, even in current state I can eat some number of surrounding bases and I may even match other factions in number of bases but it takes long and I had to put everything into building combat units so this seriously delays my economical development. By the time I an about to eat next faction they become very strong and easily block my advancement. Tayta reported that she can continuously beat them all conquering whole planet, though. I would love to see her SP game progress just out of curiosity.
Nevertheless, it seems that AI grow stronger comparing to previous settings, overall.

[Alpha Centauri Bear]

My Sensors are 50% and that was originally your idea, a long time ago.  I like them that way just fine.  It's a real drill to remove them from enemy installations.  It's a real choice to decide where to place them for your own benefit, and to remember to do it.  In fact, I may have achieved what you were trying to achieve with your territory bonus.  The difference is my method is not magic, you can get rid of the Sensors.  In fact the stock binary loves to shell Sensors, which is why you have to be careful about where you place them.  Front line Sensors, you're gonna get a warning that someone's shelling you, that's about it.  You have to put them behind your cities, for them to be long term effective.

I see no reason for you to be fretting about 25%.  You should at least try 15%, equivalent to a PLANET bonus.

The exploit about settling on top of a Sensor, well if the land inside a city got shelled, that would solve that.  Or erase all terrain improvements when the city is settled.  I don't know the doability.

Yes. Before I introduced territory bonus. I wanted combined protection from them to be about 50%. Which is what you have achieved with your configuration anyway.

I understand there is a drill. I cannot say if I like it or not. It may be interesting but I don't want any drill to become too convoluted to shade the main game course. For that same reason I oppose complex hurry formula to exclude this unnecessary mini game from the scope and many other things of that sort. Beside, the strongest argument against it is that AI does not build them effectively now. It can be taught but it is another hard task comparing to ease of upping territory bonus.

There is nothing magical about them as about any other bonus in the game. Base bonus was increased from 25% to 50% and everybody seems to agree with it. A combined "territory" bonus is stable stays at 50% in all modded variations too, which speaks for its necessity. The only variations are between splitting it between inherent (territory) and buildable/destroyable one (sensor). It's either that or armor should be stronger than weapon instead which nobody seems to like. I am all for sensor bonus as long as we can come up with its sensible lifecycle and rules - not to use it under base, work out the destruction rules, etc. It may be set indestructible or something like that. Anything would do as long as it doesn't seem like an exploit or laughingstock either way.

[Alpha Centauri Bear]

And way too much emphasis on useless water, but I don't see a parameter for that.  In fact the right exploit for the current nonsense, is probably sea conquest.  Ships can get out there and do the damage a lot faster than on land.  Watching the Morganite Cultist endgame was pathetic, they just traded empty cities.

I would obviously appreciate anyone finding current AI weaknesses. So far everyone says it lacks defense but this is not yet a definitive one until people can beat it with same strategy on a regular basis.

What if someone captured AI base, will it go out and found new one as now it is one below the cap? Sounds kind of stupid. Need some more economical limitation.

I don't know how the code works.  From its description I guessed that it's talking about making new bases, not about whether it's willing to capture bases.  This is how I play as a human anyways.  I don't spam forever, there's no point.  You're going to get a zillion bases in conquest, that you're going to have to put Punishment Spheres on at a minimum.

I mean what if AI stopped to build bases because we set a hard limit on it. Now someone, doesn't matter who - a human or another faction, capture one of AI's bases. Then they can found another one under limit. Allowing them to build another base just because they lost one is stupid rule in my mind.

[bvanevery]

ICS was one of the most notorious one everybody tried hard to limit. Yet now when I was able to teach computer do the same, making it no more exclusive human exploit, it doesn't seem to be unconditionally liked. What is going on, guys? Were you hypocritical when you complained about it and demanded this problem to be solved for good?

I think you're making the mistake of thinking that all player demographics and game design goals are created equal.  The executive summary is, there's a species of player whose only goal in life is to abuse the Meta.

I do not care what the MP crowd wants.  I don't even like some of the people that hang out in the last bastion MP community for SMAC.  Our animosities are discoverable in the archives, and have to do with things like ultimate minimaxing and flogging people that this is some kind of moral imperative to play the game this way.  I think ICS is completely disgusting, a logical abuse of the game's system.  Ultimate minimaxers came up with it and it's gross.

The phenom is not limited to SMAC.  The Freeciv crowd came up with minimax dances, Celebration booms, early use of Diplomats as a poor man's Caravan until those were researched.  Then Caravan abuse of trade routes to make stupendous profits.  So, Miniaturization by 0 A.D, then game over.

Been there done all that, to understand the mentality, and even played some competition games.  That's really the only 4X MP I've done.  It was ok, it was chummy, the main reason being that the people I played against were not that good.  So even though we knew all those things in theory, people weren't able to put them into hardcore practice.  I could do things more like my usual play aesthetics and not worry about somebody pulling a mondo minimax trick out of their ass.

MP animosity isn't limited to SMAC.  In the Battle for Wesnoth community, the design tensions between single player campaigns and multiplayer arenas were palpable.  The latter would always insist on unit ability level-ups being shorter and dumber, because a MP match wouldn't go on long enough for all the leveling up.  All they care about is what's "fair" between human opponents, who are going to destroy everything straightaway.  Whereas, when you're fighting 25 scenarios in a row in a campaign, and giving the whole thing a RPG-style narrative overlay, you want more scope for what can happen over that time.  You don't care what's fair to the computer, that's just your source of resistance for the single player's puzzle solving and narrative benefit.

The endgame is I got run out of that community over issues of showcasing 3rd party campaigns, as the MP crowd had far more power over core development.  Career-wise it was not worth it to fork a pile of GPLed code, so that was that.  Learned some things about shared authorship and moved on.

I will always design for SP first and there's no contradiction in that for me personally. I would only ever make accommodation to someone's MP sensibilities if they playtested my SP work and identified a cogent problem and it was readily solveable for MP without compromising SP objectives.  They get the back seat, that's my party line.  SP more important than MP, and I will never budge on that.

Even when I go to make a new game for commercial release, I will never budge on that.  The game industry has far too many jerk productions catering to the almighty social media and MP flagellations for in-game skin purchases "to look cool" with your peers, rampant toxicity, racism, and sexism in the player bases, etc.  Proper AI has gotten the short end of the production stick for decades.  Ideologically AI is what I want, not the MP "oh we'll just fight humans" crap.  There's a business case to be made for trying to compete in the marketplace on that basis, because it bypasses toxicity and "gaming as a service" business problems.

The downside / risk being, that monetizing AI R&D is a hard problem.  A lot of players seem to think they're "owed" a stellar AI for free, as part of the game when it ships.  As you and I both know, this has nothing to do with the production reality of making it possible.  Even the lowest hanging fruit easily consumes 2.5 calendar years of development, nevermind any new fancy algorithmic work.

So the big experiment is, are there enough people in the niche of 4X, who would actually pay for a strong AI, that open their wallets where their mouths are?  Assuming the AI also respects game design concerns, and doesn't just trounce you with goofy resources, or endless tedium ala sea base spamming.  AIs don't get bored pushing stuff around the map, humans do.

I think within "boutique" 4X, it may be possible to get the customers.  At the scale of Civ 4X, I don't think it's possible though.  I think Civ is popular because by all 3rd party reports I've received, Civ 6 is an easy dumb game.  Brains went out the window somewhere along the way.  Instead they endeavored to be popular, which probably means "a history toy".  This has a lot to do with why the attempted narrative formula of SMAC has never been revisited either.  The unwashed masses don't get philosophy and ideology and science fiction and stuff.  They get Stone Age to Space Age / Nukes.  Everyone's trying to reenact WW II, relive Hiroshima, and possibly start WW III.

[bvanevery]

I cannot say if I like it or not. It may be interesting but I don't want any drill to become too convoluted to shade the main game course.

The stock AI already knows the drill of shelling the crap out of Sensor Arrays.  That's a strong argument in favor of retaining the play mechanic.

Beside, the strongest argument against it is that AI does not build them effectively now.

You think the stock AI doesn't build enough Sensor Arrays?  I think it builds a sufficient number of them to be annoying, and to have to work for my victories.  I don't know if it's easy to convince the AI to build even more of them, but I'd expect it to be the mere increase of a terraforming weight somewhere in the code.

There is nothing magical about them as about any other bonus in the game.

An unearned bonus is magical.  Hitler's home territory is magically dangerous to attack?  FFS why?  It was actually dangerous because they actually put lots of anti-tank guns with a defense in depth towards the incoming Western Front invaders.  Nothing magical happened on the Eastern Front either, the Soviets simply did the vast majority of the dying in WW II.

A Sensor Array is easily destroyed.  That makes it non-magical.  It's this game's version of a Fortification.

Base bonus was increased from 25% to 50% and everybody seems to agree with it.

"Everybody" is too strong a word.  You didn't ask me, for instance.  The magic home territory bonus that nobody can do anything about, it's the most obvious thing to object to.  Combat is clearly dialed up way too high for arbitrary defense, when a Scout is defending against a Chaos Gun.  It just says that making bases is the only important way to advance a battle line.  If I trim the fat from the territory bonus, and find that Scouts are still kicking ass they shouldn't, then I'll look at what other defensive bonuses you've given things.

Code: [Select]

It's either that or armor should be stronger than weapon instead which nobody seems to like.
Again, "nobody" is too strong a word and you didn't ask me.  It's not a secret that my mod favors early game defense.  The combo of Huge maps extending early conquest distance, Clean Reactors available immediately, Synthmetal Garrisons being Tier 1 and not requiring a prototype, Lasers being Tier 2 and requiring a prototype, and prototypes being expensive, makes a Recon Rover rush pretty much impossible.  You could probably go pillage but you're not going to be taking enemy AI bases right away.  They don't garrison with 1 Scout, they garrison with piles of real armor.

I don't care what the MP crowd wants.  MP players probably don't want to play on Huge maps at all.  I know what's necessary to make a challenging experience against the stock AI.  It needs time to "sink down roots".  And I don't mind having some time to myself to twiddle out my roads, railroads, and forests either, because it's not just a MP deathmatch to me.  It's a Builder game.  I want to make an empire that looks cool, that isn't some ICS eyesore.  I want my cities where they're supposed to be, the way that utilizes and covers the land.  And I don't want to be kneecapped because the AI got a ridiculous resource boost to spam bases, at a cost I cannot possibly match.  Nor would I want to be bored pushing that many bases anyways.

[Nevill]

Thinker main goal was to improve AI and WTP continues on this path as well. I have heard a lot of complaints about some OP exploits humans can do but AI cannot. Teaching them to AI is usually the best way to resolve that problem. However, this is also the most difficult option. In absence of which, we usually resort to limiting such exploits to equalize human vs. computer capabilities. ICS was one of the most notorious one everybody tried hard to limit. Yet now when I was able to teach computer do the same, making it no more exclusive human exploit, it doesn't seem to be unconditionally liked. What is going on, guys? Were you hypocritical when you complained about it and demanded this problem to be solved for good? I have spent quite substantial amount of time and mental efforts on it. Was it for nothing?

ICS is a dividing notion. It's a strategy that is widely disliked, yet everyone has to do to stay competitive. Obviously staying competitive is much more of a concern in an MP game.

I am talking from v160 experience where the extra aggressive expansion wasn't toned down, and I had complained it was a bit too much. I would like to amend the statement. Something seems to be wrong with Yang, specifically. All other factions are doing exceptionally well. Research stagnation is a hazard for most AIs, but so far they show no signs of it. And there is no other remedy to it but to work more energy tiles; all the multiplying facilities in the world won't help you if there is nothing to multiply.

Yang, well. He is not doing so hot. He is building scouts and Colony Pods, Colony Pods and scouts. At least he built a single trance plasma garrison in each of his bases so far.

Here, let me show you.

I like what the builder factions (Domai & Miriam) are doing with their empires. They are actually ahead of us humans by quite a margin.

From a pure efficiency standpoint, I like the aspect where AI makes better decisions within the confines of the game mechanics. But I also get how it interferes with the faction personalities and lack of the unique feel if they are taught the most efficient way to play.

In which way? By number of bases AI can produce total? That sounds too crude. What if someone captured AI base, will it go out and found new one as now it is one below the cap? Sounds kind of stupid. Need some more economical limitation.

The drive to expand could get lower with more bases, and higher with other factors (competitive strength, stagnation, number of game turns).

[bvanevery]

Allowing them to build another base just because they lost one is stupid rule in my mind.

No it isn't.  It's a regulation valve.  "When this threshold is reached, stop doing this garbage and start doing something that's actually productive and beneficial."  In the real world of engineering stuff, you put regulation valves on things.  Doesn't matter if you'd rather have a human in charge, making perfect sapient big brain decisions about everything.  An AI is a machine.  Sometimes the coffee drip maker is set to an aperture hole of a certain size, because that's the drip that will make coffee.  The carburetor aperture is a certain size because that's the fluid flow rate that will spin the engine.

Game design, is designing and testing such regulation valves, to see if they work in the real world.  That's how you burn up 2.5 years on the iterative design of a SMAC mod, doing nothing but low hanging fruit.  It doesn't matter that this is how values are given, what matters is if it's the right value, when hand tuned.

Do you know how much better my gas mileage got when I hand tuned my carburetor with a $20 stupid tool I had to buy?  No I didn't like having to buy that stupid tool.  But damn if it worked.

And if I paint a painting, I have to decide whether this field of color over here, is going to be a little bit more green or a little bit more blue.  There is no basic difference in process.  Only in the number of steps I take to achieve a result.  It's my judgment which determines whether I've arrived at a result.

[Alpha Centauri Bear]

The drive to expand could get lower with more bases, and higher with other factors (competitive strength, stagnation, number of game turns).

It could be done easy multiple ways. I didn't ask how to limit it. I asked for some economical reason not to do it. Otherwise, we'll just force AI to give up on beneficial expansion while human continues to do it - a setback to vanilla state.

[bvanevery]

Otherwise, we'll just force AI to give up on beneficial expansion while human continues to do it - a setback to vanilla state.

What part of current expansion not being beneficial, did you fail to understand?  It's obvious that the AI is incompetent at defending sea bases.  Only your Territory bonus is holding it up.  The Pirates had No Minerals.  The Morganites and Cultists were completely incapable of garrisoning anything, they just had empty sea base rockfight exchanges.  I didn't invade by sea because sea bases are worthless.  Also they are highly vulnerable to mind control, the AI is quite fixated on doing that to sea bases.

I also don't believe any AI empire needs to be larger than 40 bases on a Huge map, to be a threat.  That's based on guesswork from previous Thinker games a long time ago.  At some point the AI needs to put its productivity into units that can harm other factions, and not new bases.  This is why I'm currently testing the new parameter, and not speculating.  The circumstantial evidence is overwhelming that the current regime is grossly incompetent, and boring for a human being to mop up.

[bvanevery]

New test game with version 182.  Huge map, 30%..50% land mass, Average settings.  Random opponents.  I drew the Spartans.
expansion_factor=50
faction_placement=0
alternative_combat_mechanics_loss_divider=1.0
combat_bonus_territory=15

Details isolated land

isolated land

Sparta Command was settled first, then Ironholm on top of Mt. Planet.  The fungal tower on Mt. Planet was there to begin with, but I settled there anyways.  I did not end up getting much victimized by mindworms, possibly because my land is isolated from other factions and nobody's been stirring the pot.  I only just met Cha Dawn at sea this turn, MY 2165.  I asked for 3 other commlinks and so far he hasn't had any.

I scouted and expanded until I learned Recreation Commons.  Then I started on those, hoping to gain Completions while doing further scouting.  Shortly afterwards I learned Recycling Tanks.  I've Completed one of those so far.

I was unable to explore the land south of Sparta Command for quite some time.  West and East were the most obviously visible pods and decent looking land, so I went for those first.  The middle fungal tower that's southwest of Sparta Command, it came from a popped pod and messed up my connectivity to Mt. Planet.  I used to have a road through there.  My empire has been expanding in 2 distinct halves, with Assassin's Redoubt the most recently settled city, curving up from underneath the fungal wall.

I'd say this proves that the stock faction placement can stick you in a bit of a fungal maze.  However, it also gave me pretty lush land at the same time.  I do wonder if the supply pod pops somehow measure the quality of the land, or if it's random chance whether you get dealt an adversarial fungal wall.  I got all kinds of great forest popping to the far east, which as you can see, has spread rapidly.  Then I got that fungal tower which cut my empire in half. 

The smaller patches of trees inside my empire, are ones I planted.  I'm trying to undermine the fungus.  I have not had time to make Scouts to destroy the fungal towers.  Popping supply pods is always a higher priority, as is setting up Completions.  I managed to complete a Command Center at Sparta Command on like turn 3.  So when it gets done with Recycling Tanks, I suspect its job will be to put out suicidal Scouts that can get rid of those towers.

The open question for faction_placement=0, is whether you can get stuck with a fungal maze and dry scratchy land.  I haven't seen that happen yet, and my count is 2 games for 2.  I figure my aborted Map of Planet game doesn't count.

The Spartans aren't as affected by mindworms as other factions would be.  Mainly I've lost Unity Rovers that went far afield and popped a mindworm from a pod.  I haven't lost anyone in my home territory.  That's because it was easy to Complete a Command Center, since the Spartans start the game with that capability.  I did the same thing with the Data Angels in a previous game.  So all of my Scouts are trained + they have a MORALE bonus.  Whereas, the Unity Rovers are not trained.

[bvanevery]

MY 2201.  Cha Dawn got uppity and went to war with me.  I'm fighting with Gun Foils because I'm too parsimonious to waste time on a prototype.  Hmm, just realized that the Spartan ability to get prototypes for free, isn't happening in this mod.  And that means this line of the description is wrong:

Code: [Select]

^Prototype units do not cost extra minerals.
But this is weird, I also see this entry in spartans.txt:

Code: [Select]

FREEPROTO, 0,
Has some alteration been made that's bugged the Spartan's ability?  There are a few thinker.ini variables affecting prototypes, but they don't obviously govern this.

Details Morganic spread

Morganic spread

I looked around for an ally and came up with Morgan.  From him I got a couple of techs, but not the Skunkworks I need.  Here you can clearly see he's got a huge pile of bases, and doesn't need to be spamming any more to make a tough empire.  I checked his city count using F4 and currently he's got 34 cities.  Checking his unit production, he's got 2 Colony Pods and 1 Sea Colony Pod in progress.  1 active Sea Colony Pod trying to settle somewhere.

Don't know how it'll go with the Free Drones next door.  Currently they have a Treaty.

Morgan is pursing Democratic Free Market Wealth, so at least his SE choices are in character.  Hope it helps him.

Cult of Planet- Police State Simple Survival 
Domai, Zhakarov - Democratic Planned Survival
Svensgaard - Fundamentalist Free Market Survival
Yang - Police State Simple Survival

I'm nothing.  I only have Police State, Democratic, and Planned as choices, and those don't yield good combos.

I have cleared most of the fungal towers from my empire and am approaching the last one.  I have not popped any more pods lately, as my ECM Scouts have been busy with the towers.  I only have the meager number of cities as shown in my previous screenshot.  I managed to get the Human Genome Project done, so I'm really not under any obligation to spread more.  I used some Artifacts for that and spent a pretty good wad of cash I'd been saving.  I'm now poor and have 3 copies of the Command Nexus underway.

[bvanevery]

Details Pirate empire

Pirate empire

MY 2205.  Yang gets uppity with me, and he was at war with Svensgaard, so I have another ally now.  I'm so po that he catches me up on some techs and gives me a free one as well.  I still can't build a Skunkworks though.  I just realized I actually have the tech for it, I'm just not allowed to build a Skunkworks.  I'm not supposed to need it, so previous observations about prototype cost, clearly indicate a bug somewhere.

Checking F4, the Pirates have 28 bases.  Only 1 is showing No Minerals.  So far so good.  Checking units, they have 1 active Colony Pod and 1 in production.  5 Sea Colony Pods and 2 Fusion Sea Colony Pods in production.  They're not done settling.

[bvanevery]

Wow, flat_hurry_cost_multiplier_project=6 ?  That's harsh.  I lost the Command Nexus because I didn't understand the difference between paying cash for minerals and cashing an Artifact.  I had the Artifacts to prevail.  Only just found that out now when attempting my followup project, the Virtual World, and massively overshooting it.  Playing this turn over again, MY 2217.

With only a +15% territorial defense bonus, one of the Cultist cities defended by mere Scouts, is getting properly shelled by mere Gun Foils.  If I get any weapon prototype done at all, that sea base is gonna die.  I'm finally bulldozing through the buggy expense of the prototypes that should be free for me, or documented as not being free, case may be.  I thought about starting over, and not playing the Spartans, but I'll see how this game goes for now.

[bvanevery]

What's up with these ability costs?  There's no predictable consistency between abilities that cost 16 and abilities that cost 1.

Details fusion foil no ability

fusion foil no ability

Details ff empath cost 16

ff empath cost 16

Details ff high morale cost 1

ff high morale cost 1

Details ff hypnotic trance cost 16

ff hypnotic trance cost 16

[bvanevery]

Details Morganic battle

Morganic battle

MY 2243.  My allies the Morganites lose their 1st city to my enemies the Free Drones.  Has Morgan produced enough units to actually fight or defend?   Nope.  Almost his entire empire is Scouts.  He's displaying gross incompetence.  Using F4 his city count is now 54.  3 active Colony Pods, 2 in production.  1 Rover Colony Pod in production.  5 Sea Colony Pods in production.  1 active Quantum Sea Colony Pod.  1 Fusion Sea Colony Pod in production.  This says to me that expansion_factor is still dialed up too high, or that it doesn't work.  I will quit this game and dial it down to 25, to see if that helps.

Details my Spartan extent

my Spartan extent

My empire is fine.  I'm not in direct contact with anyone.  I've got a huge land buffer like last game.  I do have a Fusion Foil which is finally killing Scouts in that 1 Cultist sea base.  They never made any response, other than to try to regarrison their base with more Scouts, or the occasional artillery piece.

At this distance it is not profitable to make war, so I've been concentrating on building my infrastructure.  I went Fundamentalist only as I didn't think other choices were all that good.  Democratic might be useful for GROWTH now, but previously I didn't want to lose SUPPORT or have to deal with a PLANET penalty.  I made 2 feeble mindworms which are sweeping out some supply pods.

I don't want to play the Spartans again until they can make use of a Skunkworks, or don't need to pay extra for prototypes.  As is, they're at a disadvantage for getting prototypes done.

Details my builderesqueness

my builderesqueness

I did manage to get some SPs done, despite the exorbitant cost of some of them.  The Planetary Datalinks was surprisingly pricey, but I had cities stuck in a "SP climb" for a long time, continuing to produce a defunct project.  Now I know how SP cash cost works, and how valuable Artifacts are for the job.  I typically use them to hurry SPs along in my own mod, but in this one, it would cost you 300 cash to get the same benefit.

There doesn't seem to be any point at being good at research, as long as you've got an ally.  Morgan supplied me with most of my tech, and Svensgaard some.  It takes a long time just to build basic things.  I was working on Tree Farms when I quit.  I eschewed Genejack Factories, because I wanted growth and not unhappiness.  I suppose I could have adjusted my budget to take advantage of the Virtual World network nodes that so many of my cities had.  I think I was waiting for a bit more empire spread before bothering.  I did not do Hab Complexes.