Author: martin fierz
Date: 02:22:05 09/10/02
Go up one level in this thread
On September 10, 2002 at 04:14:08, Ingo Althofer wrote: >Hello Martin, > > >>>... my main philosophy: To have done a thing first is worth more than >>>to repeat it in improved or refined ways. >> >>i think the misunderstanding is that when i say "nemesis' title is worth more >>than chinook's" i mean that nemesis plays checkers better, not that nemesis >>achieved more than chinook did. > >Thanks for this clarification. > >And with my limited background knowledge in Checkers I also believe that Nemesis >of today is probably better than the Chinook from 1996 or 1997. Of course, >"better" is a problematic term in the drawish world of Checkers. What I mean is >that I would happily accept the following bet, assuming a 100-game match between >Chinook and Nemesis: >I get 10,000 Euro when Nemesis achieves an overall win. >I lose 10,000 Euro when Chinook achieves an overall win. >Nothing is paid in case of an overall draw. you can be very happy with this bet :-) the main problem of chinook '94 is probably not that it is really worse as engine than nemesis, but that nemesis has an incredibly good opening book - much better than that of chinook. BTW, this is where the programs which played in las vegas have been first at something: we are the first to build opening books by computer alone in checkers. they are of very high quality - better than "human" books. >By the way, when I were Jonathan Schaeffer I would not put new energy in some >updating of Chinook (there are other more important tasks and challenges) but >simply allow a match between Nemesis-2002 and Chinook-1997, without any >modernization of Chinook... Just to see what progress has been made during the >years. yes, that would be nice. perhaps it will happen, if schaeffer decides to defend his title. they are still trying to solve checkers though, computing the 10-piece db for that purpose. >Db-building was only one part of the achievement of Schaeffer's group. What also >counts for Jonathan are his management qualities: he organized the whole >project, for (a small) instance he toured to get computer sponsoring. Successful >management is part of good science. i guess you are right :-) the outstanding things from a programmer's viewpoint (well, mine) about chinook is the 8-piece database which they computed at a time where it was nearly impossible to do it. and the parallel search. the rest is probably just standard stuff. aloha martin
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.