Quote Originally Posted by econ21
Ganging up on the leader may make for a more competitive game (because you are playing to the same objectives and now are one player against N others colluding). But it can't be a realistic as a rule because it assumes each AI faction has the goal is to stop other factions "winning the game" (or becoming too powerful). In reality, of course, countries have their own interests and these may or may not coincide with another power being dominant. For example, much of the world is currently content with a Pax Americana while others don't see it as in their interest.
I disagree. I think ganging up is a realistic mechanic, as I already stated. And it's certainly as realistic as the alternative. There are many examples in history where a bunch of powers got together to stop one power from growing too powerful. Heck, just look at the history of classical Greece - it's an object lesson in the strategy. The reason a Greek city-state never came to dominate the world like Rome was because Greece was a constantly changing flux of different alliances ensuring that one city could never come to dominate. Look, for example, at what happened to Athens when it tried to create an Athenian Empire.

The history of Europe is much the same, and although I'm less familiar with other parts of the world I'm sure you'd find similar patterns over and over.

The fact that some powers come to dominate the world in any case is not proof that most countries like it this way - Empires usually come about for the simple reason that the other powers are not able to stop one great power from growing stronger.

Quote Originally Posted by econ21
Again, a realism vs playability thing. Personally, I'd find it more interesting to be able to cultivate a dependable ally, browbeat a weak faction or cut a nefarious deal (Molotov-Ribbentrop style) with the enemy of my enemy. Having the AI suddenly collectively turn pyscho on me if I get too big just breaks the immersion. But then I've always preferred turtling and going for limited GA goals to the exhausting (and ahistorical) goal of conquering the entire map (or 50 provinces etc).
I take your point regarding a GA game, but let's face it, it doesn't look like M2TW is going to have a GA campaign. And since it's going to be primarily about conquest, a "ganging up" mechanic is important to maintaining balance and challenge in my opinion.

Quote Originally Posted by econ21
I agree it is important to maintain the mid-game challenge. But perhaps this is better done by programming the game so that AI factions can rise in power and reach - just as the human does. The best TW game I ever played was when I stepped into a mid-game Almohad PBM, with half the map orange and the other half purple. The conflict with a powerful Byzantine was epic, especially when added to loyalty problems and re-emergent factions including the terminator style "I'll be back" Papacy
I'd be very much against the idea of the AI being scripted to allow one power to rise above the rest along with the human player. I'd much rather have random effects from game to game, it would be pretty boring to know that one power or another is always going to be rising in power in tandem with you. It would also make it easy to focus your attentions on that one power in order to beat it.

Arguably though there could be some concessions made to diplomacy. So for example, if you had cultivated good relations with a power, its trigger for declaring war on you might be higher than for factions with whom you had a neutral or bad relationship.

It's admittedly a complicated issue, but for me the fundamental issue is gameplay and I'm more than willing to sacrifice some diplomatic nuances in order to ensure the game remains a real challenge right to the end.