News for hackers. Stuff that matters.
Re:Why Go Still Foils the Computers by Someone on Friday January 01, 2016 @08:22AM (#10821812)
I'm not convinced proofs for arbitrarily large boards say much about solving the 19x19 board. There may be a generalization of tic-tac-toe for n-dimensional boards that are hard to analyze, but that does not imply that tic-tac-toe is particularly hard to analyze (to be fair, tic-tac-toe is an example where one basically has to do an exhaustive search to show it is a draw)

A full analysis of go will have to include some "if N is less than X" clause for an X at least equal to 20, making it less beautiful, but I don't see that as a big hindrance (and X need not be optimal in a proof for a 19x19 board)

In particular, the proof I know of (http://www.cs.bu.edu/faculty/gacs/courses/cs535/papers/Licht...) uses building blocks so large that you can, maybe, fit 2 on a 19x19 board.

Also, I think it is clear white doesn't have a winning strategy, as black can pass at the start of the game.

There are no comments on this comment yet.