SMACX AI Growth mod

[Vidsek]

   Bvanevery: It just occured to me, with your recent changes to terrain enhancement costs, are you planning to keep some or all of these in your mod in general, or are they special cases only to deal with Induktio's mod/patch?

  I'm not planning to try the Thinker thing anytime soon, but am still working with my AIG/Yitzi merge where I want to be using your pure mod in the form you wish to present it to the game world in general.

  I'll start my next trial (tonight) without those cost changes, as I can always add them in (shouldn't have much affect in the few turns I'll likely complete).  And I can always just start over, anyway.

[bvanevery]

I blame my (illogical) attachment to my hard-won terraforming making me subconsciously resist considering fungus appropriately, even when it is an equal or better option.

It's not just a previous investment fallacy.  Forests are also more aesthetically pleasing than fungus.  My standing policy is to retain all previously completed forests, even when fungus becomes a better deal.  Hey it's partly a builder game, like SimCity.

The one I really have a lot of trouble with, is farms with solar collectors on them, vs. Hybrid Forest forests.  There's a point at which you should just throw away those old farms and go all forest.  Also the conundrum of Rolling Moist land that I've saved "in case I want to farm it".  Lately once I get Tree Farms, I just accept that it's not actually going to be useful to farm, so I forest it.  I don't like covering up a mineral but that's the reality of forests being better by then.

In the real world, turning alien biomass into human food would almost certainly be difficult and not quickly mastered, while extracting minerals and energy from it would be much less challenging.  Applying this reasoning to the game (early minerals and energy, rather late food from fungus) would definitely change how fungus was viewed by both the player and AI.  Whether or not it would be a viable situation from a gameplay perspective is another matter.

Why would you get enough minerals or energy to be worth it though?  From a realism standpoint.

In my experience, the AI prioritizes food over either minerals or energy when choosing squares to work and rarely if ever picks a square with less than 2 food, even if the minerals and/or energy from it would actually be of greater value to that base (eg. when population is currently capped).  Players can be more flexible.

Mostly true.  You have to have a substantial surplus of food.  The AI is more interested if the mineral has to be very valuable i.e. a Mine on a minerals special.  Sometimes though I've had to manually adjust the workers to have Stagnant food, so that I can work my big Mine.

[vonbach]

Maybe its just me but I found attacking cities to be very hard in this mod. You nearly always ended up on parity with attack and defense plus sensors and perimeter defense. It got old really fast. I eventually went back to my base game. Its a shame really.

[bvanevery]

Making cities into defensive strongpoints is definitely a change.  My intent is for people to find other ways around problems.  Like, Sensor Arrays can be destroyed.  Perimeter Defenses can be blown up with probe teams, although I'm finding in practice, you have to trash everything in the city to do it.  Cities can be starved instead of conquered, but admittedly, that takes time.  Cities can be bypassed or lands pillaged with field troops, instead of conquering.  Cities can be softened up with artillery.

But I do hear your vote that "this is a drag".  I haven't really heard enough votes one way or the other, and I'm not settled on whether I've overdone it or have done the right thing.

One thing I like about the more defensive cities, is early enemy Scouts can't just come and summarily kill me in the crib.  Nobody's tough enough at the beginning to do that.

[bvanevery]

Ok, I've had a little time to reflect on various test games I've played, and what I do and don't like about the current defensive regime.

I like that Infantry doesn't get any kind of bonus for attacking cities anymore.  As fun as it might be to do that to someone's city while on offense, it's really irritating when on defense.  Typically it's some beginning of the game Scout vs. Scout battle.  Scouts are on an Infantry chassis so boom, suddenly they're some kind of urban siege engineers.  I really don't like that.  I'm glad I did away with that.  I don't feel like futzing with whether infantry, speeders, or hovertanks are best for attacking a city.  There's still a reason to use infantry: they're cheaper.  Especially when combined with rails, they can be very cost effective for blowing someone away.

I like that Sensor Arrays are worth more and take some thinking about wiping them out before making an assault.  I've come up with various tactics for sneaking around to get the job done.  Only just now I've realized, this would be a good use of the Cloaking Device, which I make available early in the game.  Not so much for being invisible, because the AI cheats on that anyways, but because it allows you to ignore zones of control.  Usually I've been either escorting with probe teams or with Gun Jets.  I should try making other Cloaked units.

I do not like that I'm using piles of probe teams to sabotage everything in a base over and over again, just to get at the Perimeter Defense.  I've been doing it because I felt like maybe I was morally obligated to do it or something.  I do think it is a drag.  Perimeter Defenses make a huge difference in a base's defensability, +100%.  My mod punishes probe teams pretty hard, you can't just get PROBE bonuses unless you are the Data Angels.  So it's taking piles of probe teams to bring down those walls.  I find that targeted attacks just have too low a success rate, you're better off wrecking everything until you finally get to the Perimeter Defense.

So a logical solution would be to cut the Perimeter Defense in half, to +50%, right?  And make it cost half as much.  Problem is, there's nowhere in alphax.txt to do that.

So I'm back to the drawing board.  I guess I have to come up with either +25% or +50% to cut off the defense buffs.  I have to assume the Perimeter Defense is going to stay put, if I don't want to have to mess with Probe Teams all the time.  That means I can't take off Sensor Array buffs, because I already destroy Sensor Arrays.  Base defense is still stiff.  I've already eliminated the infantry bonus.  I don't want to retool the weapons and armor ratios and progressions in the tech tree.  It takes a long time to figure that stuff out.

So really this only leaves the intrinsic base defense as something to waffle upon, prevaricate, and go back to how things were before.  I currently have ibd = +50%.  In the stock game it's +25%.  Going back to +25% would be easy enough.  Is it enough?  I don't know.  I suppose I can change it, keep playing my test game, and see if it helps.

If it isn't enough, I could do ibd = +0%.  That's helpful if a base has a Perimeter Defense, as it makes the combo 25% weaker than the stock game.  But it also means a newly established base is really vulnerable.  The old Scout vs. Scout problem at the beginning of the game.

I will try ibd = +25% and see how it goes.

Actually upon further consideration, since something clearly must be done, I will just do this for version 1.29.  I'm not ready to leave bases inherently defenseless.  I need to hear how people react to the change, before dialing it back even more.  I don't want to get rid of the stronger Sensor Arrays.  I already destroy Sensor Arrays when on offense.  It's a bit of a drill but it makes me think tactically about offense, how to get behind a base to eliminate difficult Sensor Arrays.  I think that's a good thing.

[Vidsek]

     When I first started playing SMAC(X) several years ago (and discovered this site), I remember reading several strategy guides on various sites.  Most of them extolled the uberness of the Early Rush tactic as the best way to win the game.  I thought about how boring a strategy game with no choice of strategy would be and promptly raised the inherent base defense and sensor defense bonus in my rudimentary personal mod. I think I used IBD = 50 and Sensor = 40.
   I also thought the infantry base attack bonus was nonsensical (what?  Cities require room by room clearing maybe?  That would really mean that vehicular units should have an attack penalty).
 I removed that bonus from my games.

   The AI does use sensors fairly well, but I try to one-up them by not only spamming them around and within my empire (partly for early warning and to dispel the fog of war from unsettled interior areas) but also by doing a little trick I read about in some Civ forum or other: build a sensor where you want a base and then pop the colony pod right on top of it.  The sensor is not destroyed and becomes significantly harder to get at and destroy.  The delay in starting the base is a downside, but with some forethought (such as building a lot of rover formers as soon as possible) you can toss the sensors in good squares before the colony pods are even finished constructing.
 This obviously works best if you have some early unmolested build time from a larger map, and now, the AI Growth Mod.

   If early game Perimeter enhanced bases are making the game unpleasant to play, downgrading the IBD is certainly worth looking into, and easy to do.
 Another possibility might be to make Perimeter Defense come later (maybe early mid-game?).  I'm sure you have a much better sense than I as to whether that would cause more problems than it solves.  And I apologize for even suggesting something that requires messing with your carefully crafted tech progression  .

[Vidsek]

   I've been following your game testing of Induktio's Thinker X AI Growth with interest, partly just for the entertainment, and partly to glean the improvements to your AIG mod which are not specific to dealing with Thinker's substrate.
  So I was wondering if you could spare the time to point out the ones you think should go in your version of AIG?

  The changes to the Peacekeepers, Pirates, and Believers seem to fall into this bin, and I can see their value outside of Thinker from my earlier non-Thinker game tests.
  Same goes for the changes to Defense techs and facility costs, the cost changes to Terraforming and moving Condensers to EcoEng2.

  It looks like much of your most recent (Attempt #10) changes to the Build techs and facilities will also have general utility.

  So I guess the real question is if there are any that are specific to the Thinker testing and won't make it to the next AIG version.

  To get my testing of AIG on a Yitzi substrate closer to your current work-in-progress version of AIG, it would be useful to know a few additional things:

  Perimeter Defense -  *which* tier 3 Tech is it in now?  (I really like this change since I too have seen the AI obsess over it to the detriment of things more valuable to them in the earliest stages of the game)
  Supply Crawler cost - crawler chassis costs 10, so will the Supply unit now cost 70 for a total of 80?  Currently that unit costs 80, so is this a decrease in total cost?  Or have I missed something?
  You're moving Adv. Eco. Engineering, Industrial Auto., and Bio. Engineering which means their pre-reqs and what they are pre-reqs for will change.   I know detailing all of it for me will be a time sink you might not want at the moment and I can wait if that's true.   If I did have that info, my current tests of the Yitzi/AIG merge could also be evaluating those changes.

  Yitzi's patch/mod seems to make fewer fundamental changes to the game mechanics than Thinker and practically none to the AI.  This has been viewed as a sad lack by many folks, but it does mean the AI behaves pretty much as in the basic unmodded game.  The Merge tests therefore have some relevance to how AIG works with an un-modded alphax.exe.  Or so my non-expert opinion currently is.

[bvanevery]

  So I guess the real question is if there are any that are specific to the Thinker testing and won't make it to the next AIG version.

There's no such thing as me designing anything for Thinker mod's benefit.  I currently have no plan to recommend the use of Thinker mod with my mod at all. 

Rather, Thinker mod is a form of stress put on the game.  It validates whether various things in the game are exploitable or out of balance.   It does a fair  job of approximating an expert minimaxing powergamer.  Or as I called someone once, "an indefatigable calculator".  I've learned that some popular strategies are overpowered, and that some of my factions are overpowered.  Those things haven't been previously apparent to me, because they are outside my own play style, and nobody's offered up a demo game showing just how much Borehole stomping one can do in SMACX AI Growth mod.

Ergo, all adjustments are for what I feel will ultimately make a better game, regardless of who's binary is being used.  Assuming the binary merely implements AI or bugfixes and doesn't change how the game works.  I'm trying to harden my tech tree against these major exploits.  Yes in the upcoming version 1.29 you can still do all the stuff, but it will cost more and you won't be able to get started on it until much later in the game.  That may tick some people off, but I think it's a good thing.  It means there isn't a One True Borehole Strategy to get through the game.  While someone is waiting around to get enough tech to stomp everyone with Condensers and Boreholes, someone else might just build cheap Hovertanks and pillage the crap out of someone's terraforming fantasy.  Or someone might build Super Former rail heads just before someone else was going to get cranking on the Boreholes and Supply Crawlers.

I wouldn't necessarily call such midgame tactics a "rush" but I'm trying to give players more of a window for a military solution.  This whole "Thinker AI crushes you with productivity" thing kinda sucks, and it seems in the cutthroat multiplayer arena, some human beings do exactly the same sorts of things.

Perimeter Defense -  *which* tier 3 Tech is it in now?

I moved it back to C3 Advanced Subatomic Theory, same place you get Plasma armor and the Citizens' Defense Force.

Supply Crawler cost - crawler chassis costs 10,

To be clear and exact, an Infantry chassis costs 1.  In version 1.29 a "Supply Transport non-combat package" will cost 30.  By some formula that I absolutely do not understand, this results in a Fission Infantry version that costs 80 and a Fusion Infantry version that costs 50.  The latter is the same as an Artifact and that's the design goal, to keep that "by 50" method of counting.  Of course that only works if INDUSTRY is 0, so I probably shouldn't get too hung up on it.

You're moving Adv. Eco. Engineering, Industrial Auto., and Bio. Engineering which means their pre-reqs and what they are pre-reqs for will change.   I know detailing all of it for me will be a time sink you might not want at the moment and I can wait if that's true.   If I did have that info, my current tests of the Yitzi/AIG merge could also be evaluating those changes.

I never, ever specify prereq changes in detail.     Not even in my CHANGELOG.  What I can do is attach draft 1.29 files for your contemplation.  You can diff them against 1.28 versions if you like.  They are not final, as I'm still playtesting and conceivably, I could find out I did something wrong / bad.  However most of it is considered ready to release and unlikely to change.  I just can't say it's tested until it's tested.

  Yitzi's patch/mod seems to make fewer fundamental changes to the game mechanics than Thinker and practically none to the AI.  This has been viewed as a sad lack by many folks, but it does mean the AI behaves pretty much as in the basic unmodded game.  The Merge tests therefore have some relevance to how AIG works with an un-modded alphax.exe.  Or so my non-expert opinion currently is.

My memory is Yitzi changed 3 encodings in alphax.txt.  Conventional Missile strengths, stockpiling energy, and there's some 3rd thing I have trouble remembering off the top of my head.  So the consequences are what, overpowered CMs?  Game crashing because stockpiling is messed up?  And what about Naomi?

The most significant change still needing a test, is the new -1 INDUSTRY -1 SUPPORT for the Pirates.  Both whether stock AI does a reasonable job with it (I expect it will), and whether it's fine or sucks for a human player.

I'm good with the Peacekeepers losing their +2 GROWTH without any testing.  There were many versions of my mod where they didn't have that ability.  Maybe it'll turn out they need a further boost, but it's not going to become unplayable or suck.

I don't think the Hive or Believer changes require testing, but it'll be interesting to know how human players feel about those factions.

I just realized that giving the Hive any kind of MORALE bonus is really against the lore of the game.  They should all be miserable and depressed!  I'm not including my hive.txt, I need to reconsider.  There's a 5 file limit anyways.

EDIT: found a bug in alphax.txt.  Super Slider Former was supposed to be given by EcoEng2.  Reuploaded, fixed now.

[Vidsek]

       Thank you thank you!!   This is just what I needed, and more than I expected.

   The draft alphax.txt is fine for my purposes, since I'm theoretically evaluating AI Growth in a (somewhat feeble) attempt to assist you with your testing.
   About a third of our community here claims they are using Yitzi's patch so I thought it might be worthwhile to see how it performs in that environment, imperfect as it is.
   So far it seems to be co-habitating very well, and I'm very pleased with the superior AI performance as well as the overall gameplay.  Making the "one best way to win" tactics just one way to win, no better or worse than others, is possibly the aspect that I approve of the most.  There is no strategy if you have no choices.

   With my Hive faction file I've changed a lot of the "flavor" parts (base names, diplomacy text, etc.) using primarily North Korea as an inspiration with a touch of Mao.  They're now the Peoples Hoard with a Great or Beloved Leader.
And so I must agree with you that +MORAL is simply unthinkable.  I might shy away from FANATIC as well.
More fitting for them is the comment attributed to Joe Stalin: "Quantity has a quality all it's own".  Tho precisely how to implement that effectively I'm vague on at the moment.

[bvanevery]

I'm unwilling to give the Hive an INDUSTRY bonus as I think those are overpowered.  Also you can get 3 such bonuses from my SE table anyways: for Capitalist, for Wealth, and for Eudaimonic.  The Free Drones are the only faction in my mod that gets an INDUSTRY bonus and I want to keep it that way.

I just gave the Believers a SUPPORT bonus.  That's true to the original material.  I don't really want the Hive to have that as well, as it starts to feel generic.

In theory, +3 SUPPORT lets you support units up to the base size.  In practice, if your bases are big, you don't care about SUPPORT anymore because you have lots of minerals.

I could give the Hive free Recycling Tanks.  From a lore standpoint, one might wonder why.  Yes one of my favorite quotes is about that, leading to the phrase "Into the tanks!" and I don't mean the kind you drive.  But why should the Hive be more into recycling than anyone else?

[Vidsek]

   

I could give the Hive free Recycling Tanks.  From a lore standpoint, one might wonder why.  Yes one of my favorite quotes is about that, leading to the phrase "Into the tanks!" and I don't mean the kind you drive.  But why should the Hive be more into recycling than anyone else?

  Rename them Soylent Green Kitchens?

  Between the limited number of options, and being somewhat faithful to the lore, most of the better choices are gone.
  Perhaps some IMMUNITY/IMPUNITY to one or more categories or something similar could let them have bonuses in the same ones as other factions without being exactly the same.  If they can be rationalized with the lore.

  Balancing them is the goal, I wouldn't be terribly put out if that required bending logic into odd shapes.

[bvanevery]

I was actually proud of getting away from IMMUNITY / IMPUNITY and just giving them so much POLICE that they outright repress all those problems.  The original Hive has a serious -2 ECONOMY penalty for instance.  In my mod, you just end up with that by choosing Police State and Socialist.  The original Hive had EFFICIENCY that never drops below zero.  However now I call that JUSTICE, and I sure as heck am willing to have it drop below zero!  Also buffing that up isn't going to make any noticeable difference.  The POLICE already solve the problem of people becoming unhappy due to JUSTICE and Bureaucracy.

I guess there are actually 2 questions that aren't necessarily the same:

• Is the AI underperforming with this faction?  Is there a specific reason for that?
• Is the faction actually weak in the hands of a human player, compared to other factions?

I strongly feel like (1) is true from my playtesting.  Yet the Hive has Explore, Conquer as its research foci.  Explore is the basic stimulus for forming more cities and covering more land.

There was a time when I suspected the stock AI wanted to play the Hive and the Believers as though they had +2 SUPPORT, when in fact they did not.  I never figured out whether that was a real problem or not.  I remember versions of my mod where I thought the Hive was doing quite well.  Also the Spartans did quite well at some point.  Then they somehow became less impressive.

I thought it would be seriously funny to give the Hive a Punishment Sphere at every base.  However game mechanically it doesn't work.  It defeats the purpose of the +3 POLICE play style, and it forces the player to play with a research retarded faction.  Making Yang bad at research is also totally against his lore.

I could give the Hive a free Children's Creche.  That does fit the lore, particularly as I now give it with E2 Biogenetics.  Yang has the quote for that: "We hold life to be sacred, but we also know the foundation of life consists of a stream of codes not so different from the successive frames of a watchvid."  That said, completing your 1st Children's Creche gets a quote from Santiago.  Won't usually happen with the Hive since they would start with them, but they could capture a base and need to build one.

Giving a Children's Creche shouldn't empower the Hive the way +2 GROWTH empowered the Peacekeepers, because it's pretty much a given that you must build a Children's Creche for pop booming.  You still have to come up with +4 GROWTH from somewhere else.  That could be Socialist and a Golden Age, but Police State gives -1 ECONOMY -2 JUSTICE.  Socialist gives -1 ECONOMY +2 JUSTICE, so you're dealing with a net of -2 ECONOMY.  That's not so easy to get a Golden Age out of.

A Children's Creche also solves some of the Hive's JUSTICE problems for a human player.  You'll still get the same Bureaucracy warnings because the JUSTICE is so bad, but the actual effect won't be as severe.

The defensive MORALE bonuses, and whatever resistance to probe teams one gains, are in keeping with the defensive nature of the Hive and its lore.  I think I'm going to put this into test.  If it overpowers them, I could give them -1 MORALE.

[vonbach]

The hive was designed for defensive infinite city spam. In the original game you couldn't run police state planned without immunity to effic.
The social engineering  was quite unbalanced really. Democracy was frankly overpowered and everything else weak. Green wasn't much better.
Honestly I don't like the chance to "justice". I also like the old choppers.

[bvanevery]

Well you could, you just had to get the Ascetic Virtues and crank yourself to +3 POLICE.  Yang just has a much easier time getting to +3 POLICE than anyone else.  Granted another big difference in my mod is Socialist increases your JUSTICE as you grow.  Planned economy in the original game gives you -2 EFFIC, the opposite of the growth imperative.

The term JUSTICE is here to stay.  As my design notes say in readme_mod.txt:

Planned has been renamed to Socialist.  In addition, Efficiency has been renamed to Justice.  Stopping corruption by building a Courthouse was the original play mechanic in Civ II, and that's what SMAC is based on.  I dislike the term efficiency because it begs the question, efficient at what?  How about efficient at meeting people's needs and spending on social welfare programs?  The main player visible effect upon the game, is people becoming unhappy due to the Bureaucracy penalty as one's empire gets bigger.  In addition to stopping graft, why not have this also be about social justice, about providing essentials for a burgeoning population?  SMAC's Planned economy clearly refers to socialism in the diplomatic dialogues, but it insists on a kind of socialism found in historical totalitarian Communist regimes, that is not allowed to work.  It completely ignores the experience of socialism in Western democracies when a socialist party comes to power.  I have modeled Socialist economy as something that doesn't give you anything 'except' more, happier people, and it costs money.

Bold added.  How about being efficient at deploying death squads and disappearing people?  How about being efficient at throwing corpses into a hopper to utilize their fats and other organic products?  No, I want JUSTICE.

As for Choppers, I hate their multiple attacks, and I like having Gravships as a playable integrated part of the late game.  Play with the new Gravships for awhile, then get back to me about whether Choppers are missed.  The only thing I miss about Choppers is the goofy artwork, the crazy propeller thing.

[bvanevery]

I think I'm going to put this into test.  If it overpowers them, I could give them -1 MORALE.

Under Thinker mod this is a disaster.  Giving a faction +2 GROWTH in every city at the start of the game, what could possibly go wrong?  Thinker mod coughed up 20 cities in the time it took other leading factions to produce 10.

This points out a big problem in trying to please 2 masters.  What works ok for the stock binary, can be seriously amplified and twisted by Thinker mod.  Sometimes a middle ground can be found, such as only bumping Condensers and Thermal Boreholes 2X, not 3X.  But Thinker mod's middle ground threatens to change at any time.  One can never be sure how something will be amplified in the future.  Still, it's clear enough that a free Children's Creche is not a good idea, it's way more powerful than I expected.

I'm not going to give Yang his INDUSTRY bonus back.  I'm quite settled on that.  Firm.

The original game gave him +1 GROWTH, but is it really going to help?  I suppose it could offset the GROWTH penalties of Green and Power.  It's not what I'd call a "big whoop" for a human player though.  Well, maybe it'll help the AI and maybe that's all it has to do.  I'll test it.