Re-Balancing SMAX Factions

[Marid Audran]

Hello fellow Talents; I recently had the good fortune to aquire Alien Crossfire in a used game shop a fortnight ago and have been familarizing myself with its accessories and new factions thoroughly. I love the majority of the additions and can't fathom going back to vanilla SMAC now (expect perhaps for correctly functioning ambient music :( ). However, outside of the Progenitors whom I know are supposed to be formidable AI opponents, three of the new human factions seem out of whack with the system of checks and balances of the original 7. Namely, the Consciousness (Overpowered), the Data Angels (Slightly Overpowered), and the Cult of Planet (Underpowered). So I've recently altered the text files for these factions to re-balance their pros and cons a bit more, and I'd like your thoughts on the matter.

I first started with the Consciousness, whom strike me as the University on steroids. Just a SE comparison shows that the Consciousness can run, say, Planned and suffer the inefficiency loss in stride and overcome her lone growth set-back whereas the University would get "appalling" inefficiency (hurting their research) and be none the better for handling their problem drones. I thought about giving the Consciousness a -1 support penalty, but thought that'd probably stagnate their early game beyond the call of balance; -1 support coupled with -1 growth would probably tie up their mineral production too much, especially in the hands of the AI. So instead I opted for a -1 Morale penalty {Cyborgs would rather peacefully build towards perfection than wage war}, taming that Spoils of War ability of their's a bit (they must *take* the base with their moral-impaired troops first) and rendering them as a Gaian/University Hybrid.

Next I fiddled with the Data Angels, whom struck me as a bit overpowered. Their lone drawback is -1 to police, which only difference to a 0 police rating is the inability to use nerve stapling, which most players of this faction aren't planning to use anyway. In compensation, they can wage a probe war without peer, got their own personal HSA going with the discovery of Pre-Sentient Algorithims {scaled back probe rating to +1 to overcome that hideous +4 bug}, *as well* as a personal Planetary Datalinks with widespread infiltration, being in the best position for a game of catch-up if they are falling behind in tech. Players will probably be drawn to use Demo/FM with the Angels to finance all their probe activities, making their police rating even more insignifigant. So, what I did was increase their police rating from -1 to -2 {anarchistic followers will not be subjugated by military}, making use of police impossible while still not prohibiting military outside of bases. I also prohibited them from using Police State instead of Power, and changed their SE agenda to the "Wealth" values (remaining true to their Morganic roots, plus no other faction strives for this agenda). Now, to avoid stumping their unruly base growth without the benefit of police, and that revered non-lethal methods ability, I awarded them the Peacekeeper Talent ability. Except, instead of a Talent every four citizens, I made it every six so it's not quite the drone eliminator the PKs have, so its primary use will be in the early stages of the game, where on Transcend you can have a base sized three before requiring a Rec Commonds (just like most other factions).

Lastly, there is the Cult of Planet, which just from a glance struck me as almost hopeless and in serious need of beefing up to compete on Chiron--psi harmony be damned. Sure, their psi attack bonus is great, *if* you can get an army of native life up and running first. Centauri Genetics and brood pits is a long way off. Their personal economy is bad and won't get much better for those future techs, ditto the industry needed for lifecycle-inducing instructure *and* an army, mind worms or otherwise. So, what I did was give them a +1 support bonus {Cultists lead life of asceticisim}, and Impunity to the negative effect of the Fundamentalist SE choice {Zeal for the Will of Planet doesn't contravert scientific research}, turning them into a more momentum-worthy, sort of mind worm-based Believer-esque faction.

So, what do you think of my alterations? An acceptable compromise? I got the feeling that the SMAX factions weren't as playtested as thoroughly as SMAC's ones so were in need of a little tinkering. Has anyone else done any other alterations to faction text files?

[Solver]

Some (very) short comments.
Cult is not underpowered, by any means. Industry and Economy hurts, but it's very easy to achieve conquest in 100 years. My first game I played with the Cult, not really knowing what to do, I only built 2 colony pods, and achieved global conquest in less than 100 years, with IoDs and Worms.
Cybers are good, yes, but they are tough at pop-boom.
And, the Angels are only good at probing. Snag the Hunter-Seeker, plus create

[Solver]

...plus create lots of defensive probes, better on Infantry chasis with ECM, those are cheaper and will have a bonus vs. Roze's Speeder Probe Teams. Polymorphic encrypt the troops, and send in an army.

[cbn]

Your fixes seem reasonable based on the assumptions you have made, and many players have come to similar views BUT you will find a whole bunch of diference of opinion on this. There have been many many suggested fixes to the Smax factions but they are often at cross purposes to one another.

For example there are players that believe the Pirates to be overpowered or the Drones. I have seen arguments for both. The most popular argument you make is that the Cult is underpowered but even on that you will find a fair bit of stiff opposition. I acknowledge the builder difficulties of the cult but find that a mindworm army is a given. Even if not used for conquest the early exploration means more pod pops with good results. The early worm is the best explorer in the game and can pop pods with immunity.

The only test is MP and you will find that there are advocates of the Cult out there. I am playing one game as them and am the world leader in some categories, including overall standing (around 2170) despite not fighting anyone. The worms

1. found other factions early so I could play tech broker between unmet factions (great to obtain 3 techs for 1 of mine)
2. Found the best base locations including some landmarks and ZOC others away from them
3. Popped a lot of pods-- I really lucked out by getting a couple of just-started facilities done but casgh and worms are good too
4. Provided cash galore-- I love it when a pod spawns 4-5 worms--attack at leisure and bring the worms morale up

I will say that the AI plays the Cult quite stupidly in most cases (they go to war with everyone). While I would never say the Cult is a powerhouse, I find them quite playable as they are, given even average fungus/native life. Played as a "pure builder" they do have significant handicaps but there are compensating though very different advantages.

As the cult I very rarely lose a former or colony pod to a "suprise worm" out of the fungus. To achieve this type of security with other factions would require sensors or escorts. With the cult you simply troll the fungus, invariably seem to capture the first worm you encounter, and then send one worm off exploring while the local worm defends against its wild bretheren. The exploring worm using fungus and rivers can often get very far afield in very short order and this early map information is oh so valuable.

Your opinions are shared by many. The Smax factions seem less "vanilla", their advantages and disadvantages more extreme. some of them require different playstyles to be effective

[Earwicker]

I have found that the Cult's high planet rating more than offsets its penalties when played by a human. Early capture of an IoD (usually the first one your gun foil encounters) reaps huge dividends for controlling the seas (cbn's list for worms goes equally well for finding factions separated by water).

What they need in the hands of the AI is reduced aggression. A more patient cult that actually took time to build, explore, and develop would be a more formidable opponent than one which comes after you with hordes of empty transports or larval mass mindworms.

I find the Angels to be significantly underpowered when played by the AI. Not nearly devious enough. I usually find her hunkered down on some sizable island, enjoying the sun and not engaging in any of the hacking for which her faction was designed. If she does share your landmass, find a chokepoint along the border with her and stack two units there - threat eliminated. If it comes to vendetta, use the buddy system when counterattacking (which is an effective military strategy in any case). I've never seen her reach pre-sentient algorithms.

I'm a big builder, so for me her strengths are tough to exploit except in momentum style. There's only so many times you can frame other factions for your misdeeds before someone catches on, or sooner or later you'll wind up in vendetta with the world. That's not so fun. Offsetting that is the fact that her weaknesses don't hinder a builder style, so I do enjoy playing as Angels (but for their dumb-looking bases).

[Morganstern]

Marid Audran, in my admittedly brief experience with SMAX, I have perceived no need to "re-balance" the SMAX human factions. I had considerable trouble with the overwhelming power of the alien factions in my first Thinker level game, even to the point of thinking the imbalance to be silly, but after further experience realize this is more a function of altering playstyle from SMAC. The power of the Alien Factions can be accepted as simply a feature of the game, a challenge to be overcome. I've played both Drones and the Consciousness, and find they can readily prevail.

If you like to tinker with the factions, good for you, that's part of the game too. But one must wonder whether this is derived from a desire to tinker, and doubt that there is any necessity to do so for others not so inclined.

[Marid Audran]

Thanks for your input thus far folks, though I've been lurking here for several months and played SMAC off and on since it was released, that was my first post.

Morganstein, I don't dispute the *intentional* power of the Aliens, so left them alone. The Caretakers gave me quite a scare my first game. With the other factions, I just (IMHO) felt they needed a little tweaking to balance them from a human standpoint. You have to admit that one has on average an easier game with the Cyborgs because of their bonuses than with most other factions (as Vel has also pointed out, any Faction can also pop-boom through psych-allocation if Demo-Planned-Creches aren't enough).

Cbn and Earwicker, those are some deft observations you made on the Cult. Admittedly they aren't a weak faction per se, but nor is any in the hands of a competent adaptive player. Their strengths aren't as readily apparent to me; or maybe after that magic first worm I just had trouble capturing anymore in spite of my Planet rating (sounds like a hell of a start you got in that MP game Cbn). I don't think I messed with their faction chemistry that much, as Fundy morale benefits don't apply to native lifecycles. On an average I'm still more comfortable starting a momentum game as the Believers rather than the Cult.

I found it surprising that it's suggested that the Angels have less than ideal Builder prospects. I found they were great Builders with a home-probe defence system both for counter-espionage and bribing, where others have to augment it with a home-rover defence as well. In the hands of a human those -25% probes are murderous; it's just the AI who parks them square outside your base at the end of its turn.

I don't find the Pirates under or overpowered, but a lot of fun, a fairly different game from the landlubbers. Though its the faction I've found the AI plays the worst, even more than Morgan.

[Ned]

Madrid, Rather than argue about whether factions are over or under-powered, you can actually run tests to see which is consistently better. This is what the game designers actually did to balance the factions.

What you do is set up a game, then turn it to full automatic and see who wins. The game would have 3 of one faction and 4 of another. This way, accidents of location minimized.

I have run several tests, faction to faction. Just to test your thinking on this, which faction is more powerful

Lal vs. Zakarov?

Domai vs. Morgan?

BTW, the Cult is one of my favorite factions primarily due to its ability to explore the fungus and pop pods. However, take away pods, and minimize fungus, and playing the Cult is a lot harder.

[Marid Audran]

I actually have run total AI sims Ned; it's something I've recently took up on before retiring for the night. Insofar I've found that the Cult fairs slightly better than it did before in the hands of the AI; still no Hive by a stretch of the imagination but it can hold it's own and even become a real threat if it has decent starting terrain. Part of my adjustments where made with the AI in mind: AI Cult seems to have a love affair with Fundy though I think it's the SE choice that benefits them the least, since morale doesn't apply to native lifecycles, and their research, usually subpar anyway, suffers. So I gave them the immunity so it doesn't hurt them as much. The support is there to alleviate their poor industry so it takes a little bit longer for the AI to run out of minerals with all the units it cranks out. One thing AI Cult really needs to warm up to is running Green; it seems to avoid it like plague (AI Gaians is guilty of the same) even though many circumstances would greatly benefit from it, least of all giving them a monster planet rating for psi assault. I think the AI algorithm takes one look at the growth penalty and steers clear (since focus on Growth is one of the Cult's priorities in their text file).

My strategy with the Cult is probably different from most; I don't opt for an early mind worm rush but play an altered Builder strategy until mid-game, focusing on lifecycle-boosting facilities in my bases: Biology Labs, Centauri Preserves, Bioenchancement Centers, and so forth, until I get Centauri Genetics, build up a native army and let it loose. It's true they're best at pod popping, and I like nothing better than sending an Isle of the Deep into New Saragossa and coming back with two or three alien artifacts in its hold, though there are still hazards, and I make a point to never pop pods on my last two movement points.

Anyhow, your AI vs. AI queries:
1) Lal vs. Zakarov: In my experience Lal will win out. With their additional Talents they experience the least amount of drone problems, less AI citizens wasted by becoming doctors. AI Lal seems to run Demo/Planned most often, and is big on building a fair amount of facilities in their bases, including Creches, so he unintentially pop booms the most. It also benefits Lal the most because of their loosened Hab limits. His AI choices are probably the most flexible, and Demo is an almost universal good choice so it's doesn't hurt that its AI's Lal's only political choice.

Zakarov on the other hand, in the AI's hands, typically has a good start and then stagnates. He leaps forward in research early on, until his "problem drones" catch up to him and start to cripple his productivity. He doesn't put those drones in a colony pod enough to alleviate this, and seems to often become a mid-game target for a momentum faction.

2) Domai vs. Morgan: In my experience Domai will do better, not great, but better than AI Morgan. More by AI playing Morgan about as worst as possible than anything else though. You have to work to overcome his support problems in the early game but AI Morgan just aggravates it by building three Synthimetal Garrisons in his bases, essentially eating up all his minerals. He also does the least terraforming I've seen of nearly any AI faction, when he needs those mines and forests the most. He tends to meet a swift end when next to any AI faction in the early game; at best I've seen him be isolated and become a middling presence mid-game. Sadly, never a leader in the power charts though.

Domai does better than Morgan simply because his Industry bonus supports the AI's unit and facility habits better. Though he's often short on SP's I've seen AI Domai's bases with a fair amount of facilities in them.

What have been your findings in your sims?

[Ned]

Marid, You got it right. Lal beats Zakarov and Domai beat Morgan.

I am now running a Diedre vs. Aki sim. I stopped the sim to write this. But when I left it, there were two Diedre's and two Aki's left. Both Diedre's were ahead on the power meter. One was double either AKI. However, both Aki's were right there in research and one had the majority of SPs.

It is quite clear that the game designers balanced three of the "research" factions, Morgan, Aki and Zakarov, with growth restrictions. The AI has difficulty overcome these, otherwise these factions would clearly dominate. Humans, however, find ways to grow, so in the end, these three factions are probably the best for a human player.

In my experience, Diedre always gives me a good Single Player challenge. Ditto Aki. However, Morgan and Dr. Z are rarely a factor. You identified why Morgan and Dr. Z don't grow their base sizes. What I have also seen is that they also stop expanding. In contrast, Aki does expand, even though each base's growth is slower.

Until my current game, I have never known Domai to be anything but pathetic. However, in this game, where I am playing the Pirates, Domai is a real problem.

BTW, I really have a hard time playing the Pirates - even in SP. No game is easy. Always a real challege. I have a hard time understanding why anyone would say the Pirates are overpowered.

Since you are experimenting with changing the factions profiles, is there any way of forcing the AI to preferentially build Tree Farms and Centauri Preserves. These facilities have significant positive effects on reducing eco-damage. The lack these facilities forces the AI to limit mineral production. This seems to be a major limiting factor on mid to late AI performance.

[Blake]

One of the main strengths human players have over AI's is the tendency to build tree farms as soon as practical. Because tree farms are arguably the best facility in the game this has a huge impact on the AI's game. Therfore if you wanted to make an improved "cheating" AI the thing to do would be to give them free tree farms (with tech, or without for extra challenge, as often AI's don't even get enviromental eco).

Another huge improvement to AI playstyle is given them a free former at the start of the game, several factions, but particuallary Believers, Spartans, Uni often completely fail to build even one former early in the game. One former is a lot better than no formers, especially on worlds with adverse conditions.

The final factor which makes human players so much stronger than AI's is Clean Reactors.

I would estimate the AI would give a stronger fight if all factions got a free former, free tree farms and free clean reactors. Because these are things which human players make a real effort to get, it can be assumed that any decent human player will have tree farms in most of his bases within ten years of getting Env.Eco. And almost all of his units will have clean reactors within 10 years of getting Bio-engineering. Because AI's will likely NEVER build tree farms or equip Clean Reactors this gives the human a huge advantage, and this advantage would be lessened if all factions got these advantages automatically with tech.

There is also some alphax.txt modifications you can make to improve the AI's play. One is to push all terraforming options but forest further back in the tech tree, so they are forced to plant some forest early in the game, and forest spreads... another great choice if making free market police penalty -3 or -2, human players can completey nullify the "outside terrority" penalty by clever use of probes and specialist / punishment sphere bases. AI's actually cope better with reduced FM penalty because they have the -3 drones per base AI bonus.
(So many times I have seen AI Lal or Zak or Morgan crippled by running FM and having a huge airforce)

And you can also gear maps towards rapid AI expansion. How you ask? Open a map, any map in scenerio editor and dot the map with little mini-forests, just pretend that the unity pods WORKED and actually planted forest automatically. The forest spreads and the AI doesn't have to worry about being killed by thier own terraforming choices. This is by far the simplest and most effective "fair" way I have found to make the AI play better (watch out for AI's which like running FM).
(also the thick forest covering creates interesting strategic challenges, with no rover rushes and greatly boosted defense for infantry when attacking bases...)

So the simplest fixes to AI incompentence is playing on pre-forested maps, and reducing the FM police penalty to -2. While both of these also benefit the human player the AI gets a lot more benefit.

[Ned]

Good thinking Blake. No, if we could only get those AI factions to actually BUILD the TFs, rather than just get them, now we would have an equal game.

Also, your idea about formers and forests is right on.

On the FM front, I have tried taking over an AI faction that was killing itself running FM, setting the SE to something really attractive, and then returning control. The AI just goes back to FM.

I'm beginning to think that removing FM as an SE option might be what is needed.

Ned

[Marid Audran]

Yes, Blake is spot on with all his suggestions and observations.

The problem is that there is no way to make the AI preferentially build tree farms, hybrid forests, or preserves, (and I would add genejack factories); the most tinkering you can do to faction AI strategy, outside inherient factional benefits and gifting them free units, techs, and facilities, is to alter these vague priorities for War, Tech, Wealth, Power, and Growth (measured in "Yea" and "Nay" terms pretty much), and their SE preference in the text files. I have seen factions ahead in power charts eventually get around to building tree farms and even hybrid forests, but will build them haphazardly and without perspective on the benefits; they won't re-terraform their base squares into forest for example.

This whole terraforming issue is one in which I thought made AI strength and resources worse from Civ II. In Civ II the Settler handled all base-founding and square-improving responsibilities and was readily available from the start of the game; in SMAC the two are seperated and a tech is required for a former. In Civ II the AI could improve as expand as it went along, in SMAC you may have a faction wait 50 years into planetfall before getting Centauri Ecology. Additionally, there are so many more terraforming options in SMAC than in Civ II that the AI doesn't know how to exploit them and mix-terraforms badly, where in Civ II it was more straightforward and easy to get right.

I wish there was a way to directly control the AI's guildlines for terraforming, square-by-square. Additionally, it seems to participate in tech beelining, now it just needs to understand facility beelining (getting tree farms up and running). Third, unit workshop exploitation to keep their army in the field up-to-date.

Another good way to get a challenge out of the AI is to play the "Switch Sides Challenge," controlling a faction for 50 years or so then switching to some runt in the power charts, spend another 50 years beefing them up, then changing again, etc. Your worst enemy here happens to be yourself.

I think what we're all itching for is the discovery of Pre-Sentient Algorithms in computer gaming.

[Blake]

Ned, the clean mineral benefit for TF's is moot for AI's - they are too incompetent to actually get decent #minerals per base. However, if they get a decent mineral count the clean threshold will be increased by pops - and hopefully the AI can handle the worms, after all they usually have heaps of units sitting around taking up support.

Also having TF's in bases would put them in a better position to build HF', CP's.

It's annoying how easy it would be to improve the AI terraforming/build choices with access to the code. After all there is fairly rigorous criteria for good terraforming, the simplist being "plant forest (no other optionn)". It's a pity that improving AI is such a low priority in many games. The AI should have been improved much in SMAX over SMAC.. prehaps in a different universe...

Back to FM - the main penalty a human player gets from running FM is the inability to use police. With FM police penality reduced to -2 human players still wouldn't be able to use police, but AI's wouldn't be able to kill there game by running FM. Removing FM completey is another option, but FM is the only way that AI's seem to be able to make decent amounts of energy. (as AI's don't understand specialists either)

While trying to improve the AI may look hopeless minmising it's ability to hurt itself is suprisingly effective. Playing on a forest-seeded map with a reduced FM penalty and all factions have "build" priority added could almost give a challenge (atleast make for a more interesting builder game).

[Basil]

Re: the human SMAC-X factions. I've always thought they started with too many techs. The Pirates and the Cult are the only two for which I like both the concept and the technical implementation. I like the concept of the Drones, but the implementation is a bit extreme: Adam Smith used to say they were unbeatable played properly, and I believe him. I'd lean towards reducing both their industry bonus and their research penalty. I will also mention that I find the Aliens badly designed: they almost always run Fundy/Planned in the mid-game, which just kills them, and that the faction technicalities don't seem to match the faction stories (if Marr is trying to achieve Godhood through technology he knows about but needs to rediscover, why is his emphasis conquest? shouldn't it be research?). But that's a whole 'nother thread.

Re: giving the Angels Avoidance: Police State. Marid, I've never seen an AI faction except Yang run Police State, and he has it as both an agenda and something he has special bonuses with. So I don't think giving the AI an avoidance of Police State will change the way it plays, except that it won't diplomatically like factions running Police State. I don't know if that's what you wanted (in your proposed changes to the Angels) or not.

Interesting simulation results .

I agree about the AI-Market problem. I've seen AI's societies crash under the strain of maintaining needlejets while running Market too . But I don't like assigning only -2 Police to Market: not only does it make the game considerably easier, and make a SE choice I already like too much yet more powerful, but if you (the player) get the Ascetic Virtues that would take you down to -1 Police, and now the player can use military units for police under Market!

One solution I thought of was altering the Power SE choice to give it a +1 Police modifier. It makes sense considering what Power values are about, and, since I think Power is undervalued, it makes for a more balanced game, especially since the AI seems to like Power a lot. (Of course, the AI will like Power even more with this mod! )

The other way I'd tweak this is to modify the faction files to give the AI factions (whoever they may be in this particular game) a +1 bonus to police. That would also benefit them when they're not using Market, of course, but I think that only helps balance for the fact that they don't know how to use police units.

[Ned]

Hey Basil, I think you got something there. Add +1 Police to Fundy as well as to Power, and the AI is set.

In the Diedre vs. Aki sim, above, one of the Aki factions pulled far ahead in tech and took all the SPs after a point. I call that a victory for Aki.

I just got through playing a game as the Pirates. As always, I had a rough time. Lal pulled ahead in tech, not by much, but just enough so that he got virtually all the SPs. He also built Tree Farms and Hybrid Forests in virtually every base. It was pleasant? to see AI bases producing 25-30 minerals without ED.

Even so, I will also modify the game to give the AI a starting terraformer, "build" agenda, and seeded forests. Most probaby, I am going to get killed.

Ned

[Marid Audran]

Basil, I agree with you about the Aliens. They SE themselves to death more than any other faction; I've tried changing their social agendas but it's difficult when they only have one choice as an agenda and one aversion. Basically, they find a new ways to botch their government up. Additionally, their model of play is still too similar to the human factions to really give an alien vibe. After I got past their horrific novelty I'm more annoyed by their presence in a game than anything else. Their honorless diplomancy, lack of Council participation, and necessity for elimination for a Diplomatic Victory is yet another grievance since one ends up gassing them to death anyway.

Anyhow, about the Angels, my tinkering was as much for ideological reasons since I couldn't see the Angels ever functioning as a Police State, but could perhaps sponsor "Power." She's still a Democracy most of the time; mostly my agenda was to give her a proclivity to the Valies choice I think makes the most sense for the Angels, Wealth (and keeping her out of anti-probe Knowledge).

Factions by the way don't care if you run their Aversion, only the polar opposite SE choice of their Agenda. Zak for instance doesn't care if you run Fundy, his aversion, but gets in a huff over Wealth or Power, the polar opposites of his Knowledge agenda. Same with running Knowledge in Miriam's presence; she only cares if you're running Police State or Demo opposite her Fundy. Ditto Domai if you're running Green, he only cares about Future Society choices (he cancelled a pact with me for running Cybernetic). So far the altered profiles for the Cult and Angels have worked out but I think I'll renege the morale penalty for the Consciousness, whom in the AI's have trouble enough overcoming their growth penalty to become a presence.

On the Police State SE choice, I have seen other factions besides Yang use it. Santiago, Sven, and Aki (interestingly enough) all drag it out from time to time, though they typically don't remain in it as long as Fundy or Demo. In that regard it's similar to Wealth, which also has a short half-life among factions. The SE choice I've really seen rarely used is Green, even among the eco-friendly Gaians and Cult. Fundy and Planned are waaaaay overused.

Speaking of police in free market, that's another thing the AI has going against it; it NEVER uses non-lethal methods with its home troops to help suppress drones. If it learned to maybe it'd warm up to Police State more.

That's an interesting game you have going Ned; it's interesting that you mention a rough time handling the Pirates but ease with the Cult, since with me it's the opposite. Anyhow, I agree with you about Gaian strength in most Single-player games; I find alongside the Hive those are the AI's best factions. I think Dee's extra nutrient production in fungus forgives sloppy terraforming, and even plays into her hands if she gets Centauri Genetics. Additionally, starting with Centauri Ecology allows her to pump out formers immediately. It's also a breeze for her to a get a Paradigm economy. Those little edges adding up...

[Ned]

Guys, I'm having a game like I never had before in CIV, CIV II or SMACX. I rejigged Alpahax to add +1 police to Fundy and Power, as the AI usually likes to run one or both with Free Market. In my current game, all but Diedre are of course running FM. Domai is on the warpath. He is Demo, FM, Power and with the +1 police, he hasn't had drone problems yet.

I also started with forests everywhere and gave the AI a free starting former. It still terraformed farms. But when it came time to build TF's and HF's, it did - everywhere - due to the forests, I believe, but also due to the following:

I rejigged all AI agenda's to interest in technology, growth and building and set every faction to Pacifist. Aki still declared war on me at the drop of the hat. So did Domai. However, the AI is not fighting itself. Pacts and treaties are the rule.

The result is that AI is playing just like a human player. Expanding rapidly and investing in infrastructure - even Dr. Z and Domai. I am very lucky to be Rose. I have the Empath Guild. Otherwise, just as I suspected, I would have no chance at all.

I don't think I am going to win this game. All the AI factions are just as large as mine, with equal or better infrastructure. Every AI faction's research rate is greater than my own - mostly because they are running 60-80% labs. I however have enormous income, and am slowly catching up in infrastructure.

Wow!

Clearly our complaints about the AI being a poor builder are largely related to its overly hostile settings. This can be fixed by editing their profiles along the lines of the above.

[Blake]

Well, our complaints about the AI being a poor builder is more due to the poor AI programming, it is unable to minimise it's weaknesses, so the weakeness's have to be nullified in other ways. I did my own experimenting and changed the SE settings as follows:

Police State : -2 -> -1 Effic
Demo N/C
Fundy: +1 Police

Free Market: -5 -> -2 Police (but I think -3 would suffice)
Planned: -2 -> -1 Effic
Green: -2 -> -1 Growth

And then played with completey standard SMAC factions on preforested Standard MoP (Approximately 2 tiles of forest per 8x8 patch of land). But forest spreads VERY quickly. Also this is Thinker difficulty, because I want to compare them without the Transcend bonuses.

Of particular interest was Morgan, usually weak he built a decent number of bases, first tree farm in 2213, and many hab complexes & treefarms by 2225. When I stopped the game at 2163 he had a few Hybrid Forests, and was getting tech at 1 every 3 turns (!). So AI's DO know how to build if they get the terraforming done for them, also Morgan avoided the support trap AI's fall into due to his low agression. Even more interesting he had a couple of condensors, and all his farms were in Rainy spots.

The biggest remaining problem the AI's had was eating up there minerals on vast numbers of useless units. Low agression/no conquer emphasis helps, but free clean reactors would help even more, the AI would then have a real late-game punch, instead of fading away. The human player can either use an unmodified faction, or accept the free clean reactors too

Ever noticed that AI's like to build bases really close together in the jungle, and often squeze 3 bases onto the crater? You can exploit this further by laying trails of Uranium all over the map, the AI's see everything as a landmark and build heaps of bases packed in quite close.

Anyone else think it would be a good idea to make a "AI enhanced SMAX" mod with specially modified Alphax.txt, factions, and maps? I would prefer to make it "fair" in the sense that no special AI factions are used.

[Ned]

Blake, I am near the end of the game I mentioned above. What a tech/infrastructure horserace! I was able to stay with the AI because of Roze's Datalinks abilities and because my pact Bro, Dr. Z, gave me up to 15 techs. I could have stayed with the AI, tech for tech, but I too would have had to set research to 60-80%. However, instead, I set it to 40% and spent my money on infrastructure and relied heavily on Rose's talents.

I cannot even imagine playing Miriam against AI mussle such as this (Diedre, Aki and Dr. Z) and Domai having the largest continent all to himself.

What I no believe is that the lack of emphasis on building and more emphasis on war was probably a game design decision to give the Human a chance.

Next game, I am going to put back some aggression in the AI and see what happens.

FYI, None of the AI factions overspent on military. They all invested in infrastructure. At some point though, they had to stop investing in factories because of ED (at around 60 minerals per base) - all except Diedre. Diedre alone invested in CP's and TOP's. We know what benefits these bring.

[Blake]

I just played an interesting test game. The settings as follow:
Police State: -2 -> -1 Effic
Demo: -2 -> -1 Support
Fundy: +1 Police (but this didn't seem to show up ?!)

FreeMarket: -5 -> -3 Police
Planned: -2 -> -1 Effic
Green: -2 -> -1 Growth

Power: +1 Police

SMAC (not SMAX, because SMAC runs faster)
All factions changed to all priorities except Conquer
Agression = 0 for Hive, Believe, Spartan (others -1)

Transcend Difficulty
Huge MoP seeded with forest, kelp and dotted with uranium.

I went as Lal and got the whole of the jungle continent to myself (I started on the coast above the dunes). I only built 12 bases.

I stopped the game at 2212 because I was having serious cashflow problems, 10000 cash (800 per turn) and no good way to spend it And getting 2 techs per turn (shows how much more powerfull you are with preseeded forest) I was about 20% ahead of the next strongest AI, had govoner and trade pact. Latest tech was Adv.Eco.Engineering and SilkSteel Alloys.

The AI's in order of strength: This is year 2212(!) :

Miriam: Miriam started near the crater and I attribute that to her strength.

Yang: North of FW sea.

Others equal strength:
Morgan: Started below FW sea.

Santiago: Same continent as Miriam, got a smaller chunk of the land and neither Crater nor Flats.

Zak: On that northen Isle.

Dee: East of FW sea, weakest Faction.

All factions except Gaian's had income of about 100 per turn. (gaian's about 50/turn)
Morgan, Zak were getting techs every three turns, the others at about every 7 turns (still pretty good).
Every faction except Gaian's had atleast one Pact, and combined with the Trade Pact and emphasis on commerce techs huge amounts of income was generated by commerce.

Miriam won on infrastrucutre, with several Hybrid Forests built (this is by 2214!). I attribute this to the crater, which gave a boost to several of her cities. She also got the best land (other than me).

All AI factions had a decent number of hab complexes.

All AI factions had a good number of Tree Farms.

Morgan lagged slightly in facilties such as CC's, but was AI tech leader. He also had a condensor+farm+enricher combo

Of particular interest is Yang and Miriam, both of them had a very good number of bases, and high populations too. They also both have a support bonus, when running power they had +4 support which meant support base size for free, and with the large bases meant they didn't fall into the support trap.

The SP's were evenly distributed, with no faction having none, and no faction having a clear monopoly (I had 3, the VW, AV, and Neural Amp).

All factions except Dee, Yang were running FM and none had drone problems, then again D:AP hadn't arrived.

Overall despite the fact the game was still a pushover for me the AI had done a lot better for themselves than usual. I may continue playing to see what happens once Space Flight and other high tech arrives, they might even launch some satellites.

[Ned]

Blake, Your game and mine were very similar. The AI was broadly treatied and pacted (as was I). When it came time to vote for the Global Trade Pact, the vote was 7-0! All benefited from huge amounts of commerce income.

The tweaks to the SE settings are less important than setting the aggression levels lower. It has been my impression that the single biggest fault of the AI is that it is constantly at war and is therefor constantly building units - even if the enemy is nowhere in sight.

To test the above, I rejigged the agression levels to put them back to "sometimes" with Yang, Santiago, Miriam, and Sven. I am playing that game now. I am playing Rose again. Yang, Santiago and Sven have all declared war. So I suspect its back to normal for these folks with overinvestment in units.

The presence of forests speeds the game up. But most important, it seems to provide, just as you suggested, the AI the incentive to build Tree Farms and Hybrid Forests when the time comes. However, with the tech speedup, Tree Farms may come before the AI has built up enough mineral production to expreince its first pop. As a result, they may not be getting the "full" clean mineral benefit of these units. So it might be advisable to not preseeed planet with forests to see if the AI will build Tree Farms and HF's all by itself on reduced aggression.

BTW, +1 police on Fundy, and Power keep the AI going for a long time in FM - the only exception was Dr. Z, who kept to Demo, FM, Knowledge and still build up his air force and conducted agressive wars. I believe the only reason I caught him was because he was having significant drone problems.

Also, all the other factions were right back into it when Cyber came along. It has -3 police. All were running it and Free Market, so police was at -7 or -8, depending on whether they were running Power. (In case you never had an appreciation for it, the Network SP really is important b/c it eliminates this -3.) None switched to Police State, even though the combination of Police State and Cyber looks like a very attractive setting.

Even though it makes Yang more of a monster, you might think to keep Police State at -2 eff, but boost "police" to +3. This would make Police State, Cyber even more attractive. It would also make Police State a very attractive setting earlier, in combo with Green and Non Lethal Methods - especially if one is getting energy and labs from specialists.

[Basil]

Interesting experiments...

One idea I got from SNAC is to technologically restrict the building of mines until the discovery of ecological engineering. This prevents the AI from building mines in the early game, and the theory is that it makes the AI build forests early instead. If this works, it would make a much more "general" solution: you wouldn't have to create a map and preseed it with forest, but could play randomly generated maps with only a modified alphax file.

Just something you may want to add to your experiments...

[Ned]

Basil, Thanks for the hint. I'll give it a shot. Ned

[Blake]

Another way is to play on Arid, high ersosion maps. Practically everything is arid, and most is flat. Even the AI can figure out it is best to plant forest.

The main problem is the AI can't figure out that it is meant to build formers, so unless they start with one it could be 50 years before they start on terraforming (for Miriam especially)

[Blake]

Well, I spoke too soon on the game being a pushover because I failled to take into account the AI transcend bonuses.

I lost the govoner to Miriam when she bribed morgan (and after all I did for him!). Then most of the AI's declared vendetta and I ended up with only 1 pact (morgan) and 1 treaty (yang).

Morgan ran away with the tech, because he had 1 pact and 4 treaties, and combined with the trade pact and his inherint trade bonus... he was getting 2 turns per tech and with my self imposed no expansion rule I was reduced to 4 turns per tech and begging for tech from Morgan and stealing it from Yang (framing ). Altough my final pop-boom to size 18 helped remedy some of my problems.

I rememedied the cash situation by building an army of 8-4-2, Clean, ECM tanks and assorted other goodies for defense. I kinda don't want to get caught with my paints down should an AI decided to invade my territory, because it will probably be with well equiped troops, backed up by missile bombardment, possibly a joint attack too.

I was researching to Hab Domes, and then before I even had MMI someone was building the CV SP. I launched a fleet of food & energy satellites which bought my techs back down to 3 turns, researched to the CV tech and immediately rush buyed it. The CV was something I had hoped to let the AI's play with, but I was concerned at the concept of Miriam or Morgan clones. All my bases had been maxed out by pop-booms so I'm waiting for hab-domes to use it.

so, now later in the game:
I'm still on top, getting 3/turns per tech, my bases are producing 40-60 minerals, and I'm getting 1200 cash per turn.
Morgan is the true AI powerhouse, getting techs every 2 turns and 200 cash per turn. He just started running Thought Control (I'm nowhere near that :-\) which made all his units cost 2 to support, but also completey nullified the police problems.
Miriam is technically 2nd after me, but thanks to -2 research + fundy is nowhere near Morgans sheer researching power, but she is still doing pretty good for herself.

[Ned]

At least you got the CV. I didn't. It was the very first game EVER that I did not get it - had no chance to get it - since I was living on fumes from my pact-mate Dr. Z, who did get the CV.

A non aggressive AI and a preforested MAP is a real challenge! It might be best to tone down the challenge by not preforesting and just setting the aggression settings to -1.

The very high commerce income for the AI reminds me of CIV I where trade routes between AI Civilizations magically appeared - across oceans and continents - without having to have a caravan make the trip. The AI always seemed to have the advantage on trade income. The took that away in CIV II, which made the game a lot easier than CIV I. SMACX with these revised settings restores a key AI advantage that we have not seen since CIV I.

Ned

[Basil]

Marid: dropping the morale penalty for the CyC makes sense, since (in my experience) they're one of the weakest factions when controlled by the AI.

I've started a test game with minimal changes from the standard game: Free Market -4 police, Power +1 police, and no mining until Environmental Economics. I haven't gotten far enough yet to find out how the changes to Police effect the game, but the restriction on mining hasn't caused the AI to build forests; it's just caused the AI to build lots of farms (even on flat arid squares!) and solar panels. So that didn't work.

I also did some checking of old games to see how the different AI priorities (Build/ Conquer/ Explore/ Discover) effect the AI's construction of facilities. I was rather surprised by the results. The AI's broke down into 3 groups:
- Factions which had a priority of Building (Wealth), regardless of any other priorities, built many facilities of most types. These factions include Morgan, Hive, Drones, Data Angels, and Usurpers. Except for the Usurpers, all of these built a wide variety of facilities including the workhorses (recycling tanks, creches, energy banks, network nodes, and tree farms).
- Factions which had a priority of Conquer (Power) but not Build normally build few facilities beyond the 4 which seem to be built when needed. These factions include the Believers, the Pirates, the Cult, and the Caretakers.
- Factions which have neither Conquer nor Build as priorities built a reasonable number of facilities with some holes. This group includes the University, the Gaians, the Peacekeepers, and the Consciousness.
- I did not have any save games showing the AI-controlled Spartans as an advanced faction, but they should fall into the Conquer group.

Facilities also fell into categories:
A few facilities appear to be built on an as-needed basis: recreation commons, holo theatres, hab complexes, and of course pressure domes.
Most facilities were associated with the Build priority. An emphatic example is tree farms, which were built by every faction with the Build priority except the Usurpers and, in my sample games, only one faction without it (the Peacekeepers). (Even in an AI-only game where the University had discovered 71 technologies and built 5 fusion labs, a paradise garden, and a nanohospital, Zakharov did not build any tree farms, while both Morgan and Yang did, despite being far behind him in tech.)
Discover was associated with research hospitals, Explore with biology labs (emphatically) and with creches (weakly), and Conquer with morale-enhancing facilities like command centres and aerospace complexes (although Build factions also like aerospace complexes).

I only had records of one Alpha Centauri (non-X) game and it was a SNAC game. These rules did not hold there. For instance, Yang is Conquer/Explore in SNAC, (the same as Miriam in SMAC/X), but he builds lots of facilities anyway. I don't know whether the difference is due to SNAC (the mod pack) or whether the AI works significantly differently in SMAC than it does in SMACX.

These results were also somewhat dependent on circumstance. In one game with six human-controlled factions, the Cult (as the lone AI-controlled faction) built a lot of facilities despite its Conquer/Explore priorities. Perhaps this was because it was a much more peaceful game than usual for this faction?

All of this seems like a no-brainer in retrospect. But it does suggest some tweaks to make the factions stronger when run by the AI. For example, the Pirates default to Conquer/Explore, with the result that Svensgaard builds almost no facilities. If you made him Build/Conquer or Build/Explore you might significantly improve the play of the AI. The Caretakers also should be given a Build emphasis if they're to have any hope of creating size 10 cities for their subspace generators. Since they also have a Planet emphasis they need an Explore priority, leaving them Build/Explore or Build/Explore/Conquer. Miriam isn't interested in Planet, so take away Explore, and replace it with Build, making her Build/Conquer like Yang.

[Ned]

Basil, I am still holding out on reducing FM to -4 Police in favor making Fundy +1. Most factions in a serious war use Fundy, so the effect is the same. However, those who wish to stay in Demo + FM and fight a war should have a severe penalty, IMHO.

Your comment about Aki not being played well by the AI is the opposite of my experiene where Aki is typically the toughest opponent. I personally think she is overpowered, so giving her more "benefits" is the wrong approach.

I agree with your observation that the AI is extremely reluctant to build forests. Preseeding seems the only option.

I also agree that adding "build" to the AI is critical to getting them to build facilities such as Tree Farms. However, I still have observed the AI building endless numbers of units regardless, especially if they are at war. Making the factions "pacifist" seems to help most of them spend less on units and more on facilities. However, some factions, such as Santiago, still seen to build endless numbers of units. Perhaps we need tweaking elsewhere.

None of the factions, in my observation, build CP's or ToP's except Diedre. These facilities are critically important to building up clean mineral capacity. We should think more on this.

Ned

[Blake]

Basil, I am still holding out on reducing FM to -4 Police in favor making Fundy +1. Most factions in a serious war use Fundy, so the effect is the same. However, those who wish to stay in Demo + FM and fight a war should have a severe penalty, IMHO.

Ned, it is for exactly the reason you stated that reducing FM penalty is a good idea. Any smart human player is going to see that running both FM and fundie will result in lower police, enabling a serious war effort while still having FM cash. But the AI isn't smart enough to reason about things like that, and the factions which gets really screwed is the Peacekeepers and University (Roze also). Put simply AI's can't work around the FM penalty. Putting police in fundie to negate some of the FM penalty will only benefit human players and AI factions which like fundie.

I changed FM police to -3. Then the AV wonder reduced it to -2, meaning police still couldn't be used. It also made police state very viable.

I also reduced penalties on the other economies, because the potential exists for the AI's to screw themselves on planned also, the -2 effic can kill most income in a sprawling AI empire, mainly because the AI doesn't know to run Demo with planned, and wont use specialists. So I reduced planned penalty to -1. When Sven or Aki runs green it results in complete growth stagmentation (without Demo, or CC's), so again the AI can potentially trap themselves. This isn't a major problem but I thought I would halve the growth penalty anyway (also Dee needs a break with the enhanced FM)

Then I got carried away and reduced the PS and Demo penalties too. PS for the same logic as planned. Demo because AI's benefit greatly from support (if only because they continue expanding throughout the game, meaning the 10 free minerals helps).

I wasn't sure what to do with Fundie, reducing the penalty is tempting, but in the end I gave +1 support (That makes a lot of sense for fundie too, fighting for a cause and all). It also makes fundie + power rather deadly, gaining the magical +3 support, which otherwise only Yang, Miriam can achieve (or any other faction running PS).

The Values looked alright, AI's can easily cope with Industry, Probe, Morale penalties so no tweaking was required. Also power + fundie is now so powerfull that increasing power seems like overkill.

The major exploit from my changes is that PS + FM is now a real power setting, and ICS is made a bit easier. However seeing as I changed the settings for myself (making an enjoyable challenge for me, rather than a new balanced ruleset) that was quite acceptable, as I'm not an ICS'er and don't enjoy managing lots of bases.

For all AI's I then reduced the agression setting by 1 (if it wasn't already -1).

As far as priorties go, presumably BUILD encourages the construction of the economy boosting facilties (including TF's) while DISCOVER the labs facilties. This would seem to make them both important to encourage AI builders. EXPLORE may increase the chance of a few other facilties being built, but does encourage growth and expansion. CONQUER seems to put the AI into attack frenzy mode, where every mineral is put into new military units. Prehaps give all factions BUILD, DISCOVER, EXPLORE and then also give Yang, Miriam, Aliens, Cult the CONQUER priority (which would be diluted by the other 3 priorites). But even without conquer in my game Miriam easily crushed Santigo, so they hardly need conquer in order to conquer.

On a preforested map my changes mean the AI can't hurt themselves however bad there choices are, this means they just keep growing and building, the only thing stopping expansion being the borders of another faction. This made the game alot more interesting for me, it could be summarised by the time when Miriam took the position of govonership, depsite me being Lal and #1 on the power chart.

No doubt the easiest way to play against such AI would be NOT to be in first place, because of the AI's tendency to form global alliances against Player 1. Normally this is just something to laugh at, but with each AI having power comparable to yours their global alliance can easily take the tech lead (and get the majority of SP's, espically seeing they build crawlers). It would probably be easier to be in 2nd place and keep the AI's fighting between themselves (if only to reduce the trade income ).