Hey guys,

So I wanted to finally learn how to use interrupts, so I made a Reversi program for kicks. Some of you might know this game as Othello, but whatever. So this is pretty different from most other reversi programs for calc out there, because it's going to have a built-in AI that I'm trying to make as strong as possible (my goal is to beat builder's dad, who's apparently a reversi champ!)

So far for the AI, I'm utilizing a combination of minimax and strong evaluation (for corners/edges), but I'm still ironing that out so it doesn't destroy the VAT like it does now Here's a screenshot of a very early version, at any rate.

What exactly does your evaluation function look like, and are you doing any sort of quiescence searching?

So what I'm currently doing is using a evaluation edge table for the first 46 moves or so, with some special cases, then for the last 18 moves or so use minimax with alpha-beta pruning to a depth of 4 or 5, using the same evaluation algos. I'm still looking for stronger and stronger AI methods online, so expect this to change in the future!

I haven't really looked at quiescence searching, but it certainly sounds interesting! I'll be sure to check it out.

EDIT: Yes, it looks like I'll have to implement this! The horizon effect is exactly what I've been having trouble with, and noise reduction seems like a pretty simple way to alleviate those effects (although to what degree, I'm still not certain).

Yeah, a chess AI with that looks 2 moves deep and then quiescence searches will always beat a chess AI that looks 4 moves deep without it.

I'm glad you both are being so scientific with your AIs; you'll be much better coders for it (not to mention good algorithmicists). From a user-facing point of view, the game looks great so far, and if you haven't released a beta anywhere yet, I still urge you to make it a contest entry.