AI

Wednesday, March 03, 2004

Hey man. Things are ok?

I want to clarify what I mean by solved here - when I talk about solving chess, I mean solve as in solving an algebra equation, or solving Tic Tac Toe. I don't care if you put the Japanese Earth Similator supercomputer on it, I will remain undefeated in 100 human-speed tic tac toe games. There is an algorithm you can follow that will ensure you never lose in that game. That's what I want, or at least part of what I want. In my definition, chess will be solved when:

A: An algorithm is laid out, no matter how complex, wherein white (or black) always wins, and this is borne out by extensive testing. I.E., a set of practical theorems of playing chess which may not be PROVEN (in an academic, theoretical, book-l'arnin way) to be true, but for which no counterexample exists.

B. Conversely, an algorithm is laid out by which the game inevitably ends in a draw, which I think is likely. (Call it the Tic Tac Toe solution.)

Or, separate from the practical solutions above...

C. A solution is arrived at by mathematical methods, whereby it is demonstrably proved that white (or black) must win, or which proves that the problem is unsolveable - that is, a THEORETICAL solution.

I would far prefer the latter. There are more ways than you would think to pin down academically "there is no way to determine the answer to this question in the universe we live in - it is unsolveable." I would be satisfied (and even comforted) by that.

I want to clarify what I mean by solved here - when I talk about solving chess, I mean solve as in solving an algebra equation, or solving Tic Tac Toe. I don't care if you put the Japanese Earth Similator supercomputer on it, I will remain undefeated in 100 human-speed tic tac toe games. There is an algorithm you can follow that will ensure you never lose in that game. That's what I want, or at least part of what I want. In my definition, chess will be solved when:

A: An algorithm is laid out, no matter how complex, wherein white (or black) always wins, and this is borne out by extensive testing. I.E., a set of practical theorems of playing chess which may not be PROVEN (in an academic, theoretical, book-l'arnin way) to be true, but for which no counterexample exists.

B. Conversely, an algorithm is laid out by which the game inevitably ends in a draw, which I think is likely. (Call it the Tic Tac Toe solution.)

Or, separate from the practical solutions above...

C. A solution is arrived at by mathematical methods, whereby it is demonstrably proved that white (or black) must win, or which proves that the problem is unsolveable - that is, a THEORETICAL solution.

I would far prefer the latter. There are more ways than you would think to pin down academically "there is no way to determine the answer to this question in the universe we live in - it is unsolveable." I would be satisfied (and even comforted) by that.