Author: Vincent Diepeveen
Date: 15:03:09 08/11/99
Go up one level in this thread
On August 11, 1999 at 15:10:04, Christopher R. Dorr wrote: >This is an idea that has interested me for a while, and I was interested in what >others thought of it, and of how it might be implemented. > >CM6K (among others) has various 'personalities' that are supposed to play like >various players. Generally, they don't seem to. The changes in playing >parameters are often too coarse, and often seem to be ideas that someone had >about what a player values, rather than being based upon something objective. > >What I'd like to do would be to take the entire corpus of a players games (most >great GM's didn't play more than a few thousand recorded games at most), and >analyze them with a very strong program. Have the program determine what it >would play in every given position that the player had to face, and score the >correlation. A score of 1.0 would indicate that the computer chose every move >the player did, a score of .5 would mean that it chose 1/2 the moves the player >did, and 0 would mean that it chose none of the moves that the player did. > >Then, we modify a or some variables that the computer uses to evaluate the >position, i.e. center control bonuses, open file bonuses, pawn structure >penalties, piece/pawn values, depths and extensions, and do it again, possibly >using a genetic algorythm approach. Eventually, we will have better and worse >approximations of the player's style. While we may never have an exact >electronic Petrosian, the set of variables that scores a .90 correlation is a >better representative of his style than one that scored .75 (over the body of >his professional games). This would of course take many iterations, and a ton of >computing time to evaluate 3000 games at a reasonable time control, but I feel >that it would generate some useful information. the computing power is not the problem. I have about 14 computers to my avail to do that if needed for my program, and my 450PII dual can run for as long as needed. The main problem is a human who's willing to dedicate half a year of his life to sort out the problems, as the whole credibility of this thought is defined by the quality of the testset. Programs play simply too well now to directly judge them. Many moves in difficult positions they will outperform most GMs, especially tactically. In time trouble they'll perform superb. A lot of how a GM plays is defined by the openings he plays. The passiveness of Karpov is for a big deal defined by his black openingsrepertoire. I'm very sure Karpov is tactical not afraid to get into troubles IF HE HAS A POSITION THAT ALLOWS him to do that. Now as i play myselve caro-kann i simply KNOW that i must not get into tactical trouble cuz in 99% of the cases you simply lack development tempi to get a good outcome. Just grabbing a pawn is no good. Talking of grabbing pawns, look at this blitz game i played today at the server: Diepeveen (2423) vs. Hossa (2474) --- 1999.08.11 08:43:43 Rated blitz match, initial time: 5 minutes, increment: 3 seconds Move Diepeveen Hossa ---- ---------------- ---------------- 1. e4 (0:02) c6 (0:00) 2. d4 (0:02) d5 (0:00) 3. Nc3 (0:03) dxe4 (0:00) 4. Nxe4 (0:01) Qd5 (0:34) 5. Bd3 (2:07) Qxd4 (0:53) 6. Be3 (0:06) Qxb2 (0:36) 7. Ne2 (0:07) e5 (0:00) 8. a3 (0:13) Bxa3 (0:33) 9. O-O (0:12) a5 (0:27) 10. Rxa3 (1:34) Qxa3 (0:48) 11. Bc5 (0:03) Qxc5 (0:28) 12. Nxc5 (0:04) b6 (0:03) 13. Ne4 (0:07) Be6 (0:01) 14. f4 (0:07) Nd7 (0:17) 15. fxe5 (0:06) Nxe5 (0:10) 16. Nd4 (0:02) Nxd3 (0:04) 17. Nxe6 (0:05) fxe6 (0:12) 18. Qxd3 (0:01) e5 (0:07) 19. Nd6+ (0:09) Ke7 (0:10) 20. Rf7+ (0:08) Ke6 (0:08) 21. Qf5+ (0:02) Kxd6 (0:03) 22. Rd7+ (0:01) Kc5 (0:01) 23. Qxe5+ (0:02) Kc4 (0:03) 24. Rd4+ (0:02) Kc3 (0:08) 25. Qe3+ (0:01) Kb2 (0:06) 26. Qb3+ (0:06) Kc1 (0:05) 27. c4 (0:10) Rd8 (0:05) 28. Rxd8 (0:01) Nf6 (0:03) 29. Rd1# (0:02) {Black checkmated} 1-0 >Once we have some good 'proxies' of great players, we can evaluate some things >about playing strengh. One of my favorites was the old "Morphy was a 'better' >player than Tal" debate, that could never be answered because of a lack of a >common frame of reference. > >If my "95% Morphy" plays my "95% Tal", then we can get some objective judgement >about the validity of their respective approaches and philosophies of chess. > >At some point, this could have some instructive value as well. When I was >actively teaching chess, I often had an intuitional feel about where a student >needed to improve, but sometimes, I couldn't get a handle on their issues. If I >could 'pollute' my model to play 90% like my 1600 student, I could determine >such things as whether it was consistency, an undervaluation of positional >considerations, or an overvaluation of the attack relative to material, and >hence give some objective targets to attack. > >I'd be interested to hear whether anyone else feels that this is feasible, or >just a bad idea, or actually has some promise. > >Thanks, > >Chris
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.