What's the optimal setting to enjoy the game most?

From my year or so experience and the comments of others, I think one candidate set of conditions would be:

- medium galaxy (the AIs do not prioritize well in the extended colony rush phase in gigantic galaxies)

- tech trading off OR exclude minor races (minor races are otherwise money trees the AIs do not well exploit)

-not abundant anomalies, the rarer the better (the AIs never build survey ships)

- at least one level above Tough (and w/max CPU), to give the AI at least a little bonus to offset somewhat poorer research decisions

- at least 6 major AIs in medium, and more in larger galaxies

Thanks! Gotta try those. I took a look at the medium galaxy and it seemed somewhat small (I hate to have an empire that spans just a few turns of space flight :)) but it'll also make it a lot faster to play.

I'm surprised at how unbalanced the tech trading actually is. In GalCiv1 you could sell techs to AI for leases and cripple them. In GalCiv2 you can sell them to buy all their colonies which is even worse. In my current game the AI wouldn't trade barren world colonization for aquatic world colonization but sold its class 26 colony for that. Granted it was newly colonized with small population and no buildings but still... Likewise the AI didn't want to give me economy and research treaties even though I had more of both and was offering them back. Then a dozen or so turns later they gave both for just a few techs. And people rave on how great the AI in GalCiv is. Of course if you compare it to like Space Empires 4 it is good but that's not saying a lot. I noticed that you could give 100% CPU to the AI and of course did that right away but it doesn't seem to make it much smarter. Still its ability (or lack of thereof) to trade and attack are crippling. Going two levels up to suicidal wouldn't probably change anything else than the fact that because of the AI bonuses a balanced play is no longer good enough so that you need to use something like all labs/factories tricks to outresearch/produce the AI which I think isn't really strategy but just a trick to use the AI weak spots and thus not fun...

The best part of the game is mid game after the colony rush and before it is clear that you've won. I'm trying to maximize that part to actually *play* the game instead of just beating it asap.

YAQW!

If you go larger than Medium, adding more AIs might help.  Say, medium with 6, the next larger with 8 or 9.

I have not yet managed to beat the game at Suicidal in DA, but the game I'm in now might do it - gigantic, 9 AIs, 8 Minors, all abundant, scattered, normal research  I'm crushing what was the #2 AI but just got the Research Victory warning for my humongous ally (#1) on the other side of the map.

I need to get some combo of victory in the next 5 - 10 turns.  However, that ally is stronger than moi by some, and I've been fighting a war against a gun shooting, missile defending AI and the ally is a beamer with armor.  Nor could I have gone beamer and beam defense, as the AI that has teetered on the edge of war with me all game long is missile toting and beam defense.  Only recently has that AI dropped down from #2, and that was due to my keeping it constantly at war with my #1 ally and the next AI below it in the rankings.

DA Suicidal is tough without some tricks.

That's what I've gathered by reading this forum and other stuff on the web. I'm wondering what is the level you can more often than not beat with the balanced build strategy. Well, obviously at least masochistic as it was no challenge at least on the biggest map. I guess I need to try the all labs thing on suicidal but it feels that the game wasn't meant to be played that way. It might be of course that on suicidal the game might actually be more fun but having to use tricks and starting over until you get significant bonus tiles to your homeworld etc. doesn't feel like strategy anymore, it's semicheating and tricks. Gotta try it though before judging it.

I'm wondering if the production sliders were meant to work as they seem to do, i.e. even if I've got more than big enough economy with a big surplus I still can't get all the production points from my plants using the balanced build strategy. That feels stupid. Forcing people to use something like all labs strategy is annoying.

I know this isn't chess which is comparatively simple to teach well to a computer but the biggest thing strategy games lack these days is definitely a smarter AI. The technology to make it better is there but I guess it's easier to concentrate on other parts of the game and make the AI "smarter" by giving it huge bonuses. Granted, GalCiv makes the good if not best job compared to other 4X games but it would be cool to have opponents that play the game almost as well as a computer can play chess.

In the early years, the way masters beat computers at chess was to get them out of the book as quickly as possible.  After that, it was easy, since human masters could prioritize which variations they would analyze while the software had to treat all possibilities equally when analyzing.  (Heh-eh!  You could send programmers back then into wide-eyed fits of stammering by mentioning "nodal pruning"!  I know this because I was guilty of that and similar sins!)

That's essentially what humans do in 4X games to programmed opponents, and it's much easier to do than at chess.

Suppose armor is the easiest to research, so you program the AI to go that route.  Well, the human reacts and does not research guns.  If you program the AI to research multiple defenses, the human goes deep in one, same with weapons. 

Any 4X game of any scope has a great many opportunities for human tweaks, and a truly strong AI would be a masterful achievement, and an unlikely one as the game will sunset, unlike chess.  Folk have spent their lives being chess players (Steinitz, Morphy, Capablanca, Fischer, etc.) and also simply trying to program chess.  IIRC, the first chess program was in 1957 and it represented several years work.  Thus, folk have been refining chess programs for well over 50 years!

Despite all the refinement in chess program software, I think that it was also the advances in computer hardware that let the chess programs get strong.  The original chess programs ran on mainframe computers capable of maybe 50,000 instructions a second (kIPS) with under 100K of memory.  Current consumer (not the Crays!) run at something like 70 billion instructions per second (70K MIPS) with over 100 Gig of memory!

Nonetheless, scientists and programmers of all sorts have been refining chess "AIs" for over 50 years!

Yeah that's very true and creating even something close to as good as a chess AI is very hard if not impossible. Knowing something about neural networks and computerized learning I'd guess that a 4X AI could be much better than it is today. Chess on the other hand has received a very large amount of resources for AI development as it has been some kind of a benchmark for how "bright" a computer can be. Perhaps after ten or twnety more years of development also 4X AIs will be significantly smarter.

In the meanwhile I'll go back to my obscene and massive galaxy.

Optimal?

Conquer Nine opponents and Eight minors without trading anything or using diplomacy & other victory conditions to escape abundant settings everywhere while defining maximal difficulty options such as boosted AI algorithms through suicidal capacity levels -- on a tiny map. Immense only delays the inevitable outcomes as epic as they may seem.

17 opponents on a tiny map? That sounds pretty tight. Just tried the all labs strategy on a tiny map with 2 opponents and still everything was colonized in the first 10 turns or so. Probably not much room for colonization with 1 opponents.

Btw I yesterday patched from 1.97 (which came with the package) to 2.02. As it corrupted my old saved games I had to start a new game. It seems that the new version is somewhat harder or at least requires more effort to play competitively. I'm guessing that it's because tourism revenue raises now so slowly that you just can't keep you economy in the black which was so easy in 1.97 after the first few dozen turns. Forces you to trade more and at least on the immense map I wasn't able to trade enough (didn't get to find enough civs soon enough) to get keep my production at 100% all the time. Now I do but it's all the time close falling apart as. So far I like the change, adds some difficulty and planning need to the game without just obscenely boosting the AI.