Author: William H Rogers
Date: 13:26:56 02/05/02
Go up one level in this thread
On February 03, 2002 at 17:25:48, Wylie Garvin wrote: >On February 03, 2002 at 13:32:42, William H Rogers wrote: >>As you can see in the lower ply numbers the program gained the most strenght, >>but as the ply level got higher the rating increase became smaller and smaller. >>It would be nice to see some math on a curve to estimate the over all effects. >>Bill > >Hi, > There's a 1997 paper by Schaeffer et. al. that refutes the idea that the >increase in strength is constant per ply at high search depths. They suggest >that there are diminishing returns for deeper search, and that previous research >didn't reveal it simply because chess programs make lots of evaluation mistakes. > >http://citeseer.nj.nec.com/junghanns97diminishing.html > >http://citeseer.nj.nec.com/rd/0%2C17407%2C1%2C0.25%2CDownload/http%253A%252F%252Fciteseer.nj.nec.com/compress/0/papers/cs/1444/http%253AzSzzSzwww.cs.ualberta.cazSz%257EandreaszSzPaperszSzdim.ps.gz/junghanns97diminishing.ps > >regards, >Wylie Your post just agrees with what I posted above. I stated that the deeper a program searches the less its playing streght increases, but it does still tend to increase even if at a smaller rate. Now as far as the author states that "increases in ply depth becomes more inaccurate because of program mistakes" I can not say that I agree with that. I would like to know on what he based his opinion on, ie. what programs, etc. If you ask any of the successful programmers here if they think that their programs grow weaker with increased plys searched, I do not think that they would agree with you or the author of that statement. Many of the best programs now search to 12 or plys and even more. I do not think that todays programs have any openly defined errors in logic that would cause errors in their search depth. In my opinion, you would be better to poll the top programmers here and then state some facts based upon what they have to say rather than some professor who most likely has not written a chess program himself or if he even tested some, probably tested some inferior programs at best. I will give you one example: Say a program was designed to search for captures only as material was its only function. You can not say that the program was flawed or made errors just because it lacked other logical functions. In fact, it should and would probably functions perfectly within it designed parameters. Missing better moves based upon center control or other ideas is an ongoing exploration in the chess field if it wern't then chess would have been solved years ago. Bill
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.