Author: Georg v. Zimmermann
Date: 04:44:36 11/28/03
Disclaimer: the following is not legally binding. It is an offer to negotiate. Nevertheless, it is a serious proposal. It has been repeatetly stated that a 'clone' of crafty can not be 100 or more elo stronger, by many people here. I do not believe so. How about a bet ? I would state that I can improve Crafty by 100 points or more within 4 month. Since it would involve many sleepless nights I would want the opposition to "feel the pain" as well. I would want this bet to be *at least* $5000 worth for me. And since I am a student I dont think I can offer a ratio better than 1:5 in my favor. But keep in mind that I have to do all the work. I am willing to bet with multiple people on this, to raise this ammount. But to raise the ammount should be no problem, considering how many stupid rich folks with high GHz computers post here. :-) Testing proedure: I am very flexible here, make your suggestion. Something like 1000 games , 100 each against a selection of opponents of which 5 are stronger and 5 weaker than Crafty sounds fair. Maybe even more different opponents, so I can not "tune" against them. Important is that the exact procedure, including versions to play against and crafty version to compare against, and exact machine to run the tests on, his agreed beforehand. Note I consider the book part of the engine ! Actually 100 points seems a bit low on consideration. If thats a problem I might be willing to raise it a bit. Betting procedure: I suggest some arbiters who are universally respected, if they are willing. Prof. Hyatt comes to mind of course. I suggest all money is paid to them on completion of the bet. This avoids many possible problems. What is a clone: My clone would have to have significant Crafty code in it. We would need to discuss how significant. I would think using Crafty Move Generation should be seen as enough. Of course, I would have to submit code to the arbiter, and show I am not "cloning" any other program. -- This is a pretty good offer, people. Remember I am a pretty bad programmer. My "own" program (which i didnt even completely write by myself) is rated around 2100 only in the winboard lists. Also I am a law student. And my state examination is in one year, so I dont even really have time for this. Georg
This page took 0.01 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.