Improving the AI

[LDiCesare]

Yes the map size plays a part, but do a little experiment -

Play the first 100 or so turns with 2 civs on "Tiny", and then play the same 100 or so turns with 2 civs on "Huge". The huge map certainly does not take exponentially more time per turn, it takes marginally more time.

But then compare the Huge map with 2 civs versus 16 civs, and you'll find that the time does go up significantly.

Which supports the idea that it is the units, not the map which is the largest factor.

Now make the same experiment with the 100 last turns and tell me if that's still the same thing. Yes, early on, number of civs is important because in this time the number of units is proportional (more or less) to the number of civs. Later in the game, it is not.
There is a very big point in checking the ai performance on the late game and not the early game: Civ3 is unfinishable for me because late game it takes forever and a half to compute its moves whereas it's ok in the early game.

You can say that early on the ai will be civ-intensive and later on it will be map-intensive. But this 'later on' is when most of the time is spent anyway.

Just to deemphasize the processing power point of view, I want to say that MrBaggins's post was spot on. The difficult thing is to think of everything you have to manage, assess every danger and opportunity, and make bids as to which course of action is the best and what resources should be allocated to it.

[edit for spelling]

[MrBaggins]

Originally posted by LDiCesare

Now make the same experiment with the 100 last turns and tell me if that's still the same thing. Yes, early on, number of civs is important because in this time the number of units is proportional (more or less) to the number of civs. Later in the game, it is not.
There is a very big point in checking the ai performance on the late game and not the early game: Civ3 is unfinishable for me because late game it takes forever and a half to compute its moves whereas it's ok in the early game.

You can say that early on the ai will be civ-intensive and later on it will be map-intensive. But this 'later on' is when most of the time is spent anyway.

Just to deemphasize the processing power point of view, I want to say that MrBaggins's post was spot on. The difficult thing is to think of everything you have to manage, assess every danger and opportunity, and make bids as to which course of action is the best and what resources should be allocated to it.

[edit for spelling]

The original point I was making was a very simple one, responding to an oversimplified idea... that civilization numbers effect AI work in a linear fashion, and that the number of opponents in Chess are fixed, but are variable in Civ... therefore whats common practice there shouldn't necessarily be common practice here.

I believe that the linear scaling of civilization numbers does hold up to the point of tile expansion conflict (by which time you're relatively late into game anyway,) unless someone has come up with a method of dealing with every strategy for every AI simultaneously, in one pass.

As for the difficulty of modelling these situations... I've investigated these for a couple of years now. The issues are obvious, and the solutions incomplete. There is a general lack of understanding of the nuts and bolts of AI by the general gaming public... its viewed from a black box mentality... it just works... and they can always improve it.

Personally, I think that since the problem is fuzzy logic (almost every decision is partially different and partially interellated)... that the solution should be approached with at least a psuedo-fussy approach. There's no substitute for initial information though, and sadly theres no mathematical data structure which can perfectly handle spacial influence systems including a time element, which are viable for the situation. We are stuck with using rough approximations of human perception... particularly brute force methods when involving the map.