Re: Turn Based Games

On Sun, 23 Nov 2003 21:36:00 GMT, Jason Reposa wrote:
> The algorithm() code should be something a second year undergraduate CS
> major could complete in one to two weeks.

Ok, I think I got it.  What about dominoes?  True, it's an imperfect
information game, but I don't think that changes the complexity like it
does for, say, poker.

I think the imperfection is comparable to the randomness of dice in
backgammon, and AIs are fairly simple (and effective) for that game.

It is very well suited for minimax, and should definitely be doable in two
weeks.

Whatever you decide, I'd like the protocol specs so I can participate.  I'd
be doing mine in Delphi, though. :)

> In particular I was looking at minimax. Could you suggest other algorithms I
> should be researching for games? I know a* for path finding. But what else?

For second year programmers with no AI experience, I would not go beyond
minimax.  It is simple, and depending on your game, the heuristic algorithm
can be tough enough.

> I own an autographed copy of the book

Oh, now I'm jealous. <g>
-- 
Vik

Food + Heat = Cooking