Author: Albert Silver
Date: 09:15:17 12/29/01
Go up one level in this thread
On December 29, 2001 at 05:11:26, Tom Kerrigan wrote:
It would help if you gave a clear example of such a position. I'll add that the
3rd point of Uri is dead on IMO (gets position A but not B anymore) and so is
Martin's. Martin's shows the complexity in what you hope to achieve, as it isn't
a question of one move, but all the moves that lead to it as well. Another
problem is that you may be choosing a position that also requires a certain
depth to be achieved even with the correct criteria and weights, and that depth
would have to be very well measured. If you force the eval to find the move
earlier, you are unquestionably hurting your eval and not improving it since the
basis to find the move is insufficient. It would be like my trying to find a the
key to a 5 move combination even though I absolutely see only 3 moves ahead. If
I can't see the real reason by the necessary means (sufficiently deep
calculations), then based on what did I find the move?
Albert
>Here's my idea.
>
>You have a position and you want your program to play a certain move (which it
>presumably isn't playing). You run this algorithm:
>
>1. Search the position, get a PV. The evaluation of the last position of the PV
>is eval(1).
>2. Search only the move that you want your program to make, get a PV. This
>end-point evaluation is eval(2).
>3. Figure out which eval terms are different between eval(1) and eval(2).
>Decrease the weights of all the different eval(1) terms slightly. Increase the
>eval(2) terms slightly.
>4. Repeat until the program plays the move you want.
>
>You could run this on lots of positions from GM games, to get your program to
>play like a GM. (At least in some positions, heh.)
>
>Has this been done before? Are there any glaring problems with this idea? Does
>anybody want to try this? If so, I'd like some credit for it. If not, I'll
>probably get around to trying it sometime...
>
>-Tom
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.