Here is part of my thoughts on the problems of civ, lengthy so people might not be interested, but I hope you read it and find thoughts interesting, and hopefully thought and discussion provoking. I also have other negative outlooks about civ, but for now this will do
Lets first just agree that alliances, and duopoly (or maybe tripoly) is stable/attracting state of games with diplomacy. I'll try to convince you.
So you start a game, you are doing relatively good, better than all your neighbors in fact! They are not happy about it, and couple of them form an alliance, and in unity they are stronger than you. You have to find a way out of it, so you recruit some other player to your alliance, likely it is not hard to convince them, as you are quit strong, and they can expect some protection and cooperation with stronger player, there is little reason to not accept it. So the first alliance is now again at disadvantage, they go looking for more members. Until the world is divided into two more or less equal forces.
There are some problems with such scenario, but if allied victory is enabled, unless you are very very strong or have smaller alliance which is very strong, joining in is likely best for you. If you are that good, and do not want to share score, you (or your small elite alliance) can try to play on its own. But if you are doing great (not needed in fact, only perception that you are overly strong is enough), expect all guns to be targeted at you, as others might see no chance unless they band up. So you will still likely end up in two alliance state, maybe taking a bit longer. So such scenarios have examples in my previous experiences.
If allied victory is disabled, you would expect alliances to collapse some time. But even though I have no experience with complete game without allied victory, I have a strong suspicion that duopoly will exist for most of the time, as again being left alone when all the the forces are consolidate is not good place to be in.
I would call all this power concentration by accretion, power attracts new players, creates some opposition, which also attracts players, until merges are complete. And here is the most important part, and biggest problem with alliances in my eyes: once power is accreated there is very little incentive to leave the power camp. I would even dare to say that likely there is negative incentive to leaving. If you are not switching sides, you are not making friends, but leaving an alliance will most likely make you look like traitor, that you kinda are, and you can expect your alliance on your back, unless they are already in dire situation, but then why would you be leaving? If something is crushing alliance, surely you can't stand against it on your own.
But why there is no reason to leave alliances? Well first of if allied victory is enabled, there will never be such conditions in my opinion, so lets forged allied victory. The reason is simple in my eyes: victory condition. There is single victory condition, which is very very far most of the time, and it is risky to break of if there is that much time ahead as bad things surely will happen in such a long time! Now more observant of you might say wait wait, what single victory condition? There is spaceship victory, and other types of victories in other civ games! And I say they are bullshit and stupid, as they change nature of game little. Especially in freeciv; only alternative is spaceship, that come so far in a future that by the time it is reached you can expect one player or alliance to hold most of everything. That is for most of the game you can't scare nether your enemies nor friends with completion of ship. You can't count on wining that way, so you must stick to grand plan, the alliance.
This comes from the fact that civ is game of largeish scale, and takes ton of time, so the end of game is far, I would also say that this also promotes more passive play (not only diplomacy & alliances, some more specific rules contribute significantly too), at least it seems from my couple of multi-player games, sitting and "developing" everything until you build your "awesome army" (TM) and crush everything. Such style of play in other strategy games, namely rts'es is hallmark of noobines. Here it is more difficult to say that, as game almost actively promotes it. And I do not think it is a good thing, I'm starting to find this insanely depressing, especially on gt where turns take day, if you have to wait months for any real actions or progress, that seems just like a waste of time. And for those of you who think that his is more strategic and stuff, well I'll likely address it in future posts.
Now I will digress a bit. There is this board game, havannah(http://en.wikipedia.org/wiki/Havannah
). It is abstract strategy game, like chess or go, and definitely a match for them in depth and in how interesting it is (well at least for me
). The game is plaeyd on hexagon of hexagonal cells. Most of the games are won by having a chain of stones connecting three sides of hexagon. The important word in previous sentence is most. As such connection (called fork) is just one of three ways to win the game, also the one taking most stones, but it is the most common. (others are ring, surrounding any number of cells, even one; and fork connecting two corners). Even though other two ways in ideal case take much less stones than fork, they are rear, since they are mostly easy to foil. But the existence of those win conditions makes game insanely more rich, as they are often utilized as threats, making your opponent defend while providing you opportunities to extend your influence.
After introducing you to havannah, lets get back to civ. We (well I in this case) can try draw certain analogies between civ and havannah. You have a moderately lengthy game that is won by moderately complex goal. The difference is that havannah provides opportunities to threaten your opponent and use those threats to shape the game making it much more "alive" and interesting. And I think that civ lacks precisely that! You can't just break of alliance and complete some feat to quickly win and end the game, you know that you must drag yourself through all of it, fighting likely superior foes if you are left alone, so you stay, and help build the "awesome army" nothing that much happens in a meantime. But if there was something that you could reasonably expect to accomplish in secrecy from your alliance that would win you the game you might do it! Your allies will likely find out about it eventually and attack you, but if you get just enough time, you win, and fuck em'! So that could end game too soon... which wouldn't be all that good! But remember havannah, most (I think above 90% of times) games are won by completing most complicated objective! That is because goals interact so well. I would expect that something like that could be found and implemented in civ games, GT in particular for our interested.
Some possibilities for that could be: build certain number of special "win wonders" or maybe just improvements, have city of certain size, build one super win wonder, build special win "base"/fortress maybe combination of such stuff, or some other practical sillynes like that. The point is to have something that would keep your allies on their toes about your actions, and that you could reasonably expect to pull off so that you would try to do it, while still having some way to stop it.
Lets analyze some proposed conditions: suppose you can win by having wonder that cost lets say 1000 shields, quite hefty, but of course build-able in one turn with enough caravans, so you can try covering it up with trade route build, but your allies will not be happy if they see tons of caravans in your lands, expect them to eliminate such threat (but maybe you can hide it, route them in unclear ways, and leave just enough time, and backup gold to complete it!), if you disable vision sharing, expect even more suspicion! Alas alliance will start to destabilize at least somewhat (I would hope significantly). If your allies are not lame they should manage to prevent it, then you pack your caravans and probably go with your bushiness as your former (or even repeated) allies still have major threat of other alliance, to deal with.
Or say you can win by building "victory base", that needs something like 200 worker turns or similar, you have too much workers? well you'r a suspect of trying to pull it off! Same pattern like before. You again can expect to trick people with that, but if they are not blind they should catch you most of the time. And so on.
Well if you read this far, all I can say is thanks