Author: James T. Walker
Date: 12:00:37 03/25/02
Go up one level in this thread
On March 25, 2002 at 13:42:50, Gian-Carlo Pascutto wrote: >On March 25, 2002 at 13:35:33, James T. Walker wrote: > >>I just did a quick test to see if there is any gain through book learning. I >>loaded One computer with Fritz 7 and let it learn from the 3 databases I have >>(more than 1700 games played by Fritz 7. In the other computer (both AMD 1.4G) >>I cleared the book learning in Fritz 7 and played 100 games at G/1 minute (for >>quick results of course). The final score: Fritz without previous book >>learning won by 52-48. >>Comments?/Conclusions?/Insults? >>Jim > >In matches it certainly helps. I remember a Yace clone >that had book learning disabled on FICS. After losing 2 games to >it, Fritz picked up a winning line in the Marshall Gambit. >Every game with black thereafter, Yace lost chancelessy. > >-- >GCP That's funny since I still remember when Fritz book learning was broke for one side. I can't remember which side but one side was not learning. I was playing a match vs Junior and Junior found a winning line and then won every other game untill I stopped the match (about 15 straight games from one side). So I know that book learning prevents that kind of thing. But what about in normal operation where each program has many thousands of lines to play from. How much book learning does it take to start making a diffenence in rating? It also depends on how it is implemented. I think Fritz has to lose at least 2 games for it to take effect (default setting). So in 1700 games Fritz probably lost only 400. Of those how many were duplicates? So Probably there is very litttle effect so far (except for the case noted above). Jim
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.