Author: Mogens Larsen
Date: 00:07:50 05/02/00
Go up one level in this thread
On May 02, 2000 at 02:24:59, Chessfun wrote: >No you are to busy reading Dr. Hyatt's posts. >You assume I assumed. When in fact I assumed something was wrong >based on two different results 15-5 and 11-9. Had Crafty the first >set 11-9 without the second set happening I wouldn't have questioned >it at all. I know that you assumed something was wrong based on two results, but how can you reach that assumption? Rerunning the test once isn't enough to validate your choice. >Unbiased? what does that mean that I would make false scores, >to support my opinion that Crafty cannot beat F6a in Nunn 1. >If it don't mean that then I don't see your point. As stated >elsewhere I check all games. No, I'm not talking about making false scores. Give me _some_ credit. I would like to know how you check the games. Crafty _did_ beat Fritz in Nunn 1, you posted the score yourself. You can't just throw that result out. >Your posts are always critical but you never address either >the content of the actual games themselves nor any facts >supporting what you say. The games are hard data and IMO that >is worth more than just hearing I am biased. No, my posts are not always critical. I approach this on two levels: 1) I ask questions about the actual testing procedure when I'm uncertain and I ask you about possible factors that may or may not skew the results. 2) I question your conclusion, whether they are a expression of opinion or not, when they are unfounded. Since I don't own Fritz or Hiarcs I can't check the games, which you already know. Even if I had the right software, I would need a very similar setup to make intelligent conclusions about the validity of the game results. I can only thrust what you supply of information. I trust the games when you make it clear how they were obtained and under which conditions. However, I don't trust your conclusions and opinions about their interpretation. I hope that clarifies my position. >There are NO problems with Nunn. In your opinion it don't answer >questions but in mine it does. The only problem is their non-uniqueness. They're just as subjective as any other set of positions. Therefore they mean nothing. There's no evidence that Nunn 1 is especially good at revealing anything about the strength of a program. Since the set doesn't represent an objective viewpoint it's no better than an opening book, who knows maybe even worse, rendering the positions useless in my opinion. I you have evidence that confirms the excellence of Nunn positions, please let me know. Which program is best at Nunn 1 positions? Sincerely, Mogens
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.