Author: jonathan Baxter
Date: 00:31:12 06/14/98
Go up one level in this thread
On June 14, 1998 at 00:13:47, Steven J. Edwards wrote: >On June 13, 1998 at 23:47:15, jonathan Baxter wrote: > >>>A position is folded via one or more of X, Y, and x=y rotations until >>>the non-default flavor subscript man is in the restricted (32 or 10) >>>square field. >> >>Do you mean *reflections* rather than rotations? > >A rotation though 3-space. Assume that the x axis runs between ranks >four and five and the y axis runs between files d and e. The three >transformations are: > >1) x' = -x > >2) y' = -y > >3) x' = y; y' = x ok. i would call those reflections about a line in 2-d. If you rotate in 3-d the pieces fall off :-) (maybe they are magnetic?) >>If you are interested, I am trying to train an evaluation function >>directly >>using the tablebases. > >Please post your results. Others have tried this before, but all >attempts to come up with a completely accurate (if imprecise) function >have generated functions more complicated (time wise) than a simple file >probe. I find the file probes to be pretty expensive---in KnightCap the disk seeks can cost up to 2/3 of the processing time. But I have two reasons for doing this: speed is one and the other is if you can do it accurately then you may be able to generalize to more pieces without having to compute all positions. I'll certainly post the results. Do you have some references for previous attempts? Cheers, Jon
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.