You are viewing a single comment's thread from:
RE: Knight tour - graph theory problem related to chess.
Thanks. I can write more now, I have cancelled exercises and lectures on studies until 14th April. Do You have any mathematical questions which You want me to write article about?
Yes, as a matter of fact I wrote a chess program a few years ago, haven’t updated it though. And it was always my intention to find some way to use calculus (rate of change) equations to be a part of the chess engine. For many years I believe it could be done as oppose to ten or more iterations (loops). I’ve been lead to believe very early on that calculus could be applied to everything.
Can one apply calculus to a chess program?
Hmm. I think that the future in computer in games is localised in machine learning. I think that it is too hard to measure precisely how arbitrary move will change the strengths. One different postiion of one figure can turn move which is at 1000 situations very good move into a very bad move. I think that chess are too complex to even write such equations, not saying about implementating them. Machine learning computers win in Chess, in Go, in Starcraft 2 easily with human. This is the future I think.
Maybe for checkers it would be possible, but it wouldnt be worth to do it I think.
Thanks for your reply and I agree.