Author: Chris Carson
Date: 05:39:46 07/29/00
Go up one level in this thread
On July 29, 2000 at 03:28:19, Mogens Larsen wrote: >On July 28, 2000 at 20:25:06, Chris Carson wrote: > >>We disagree here. I do believe that Fritz on a single chip (AMD 450 or >>faster) could achieve a GM norm. > >Since it's impossible to know, it's pure speculation. I do find it hard to >believe that it could have performed like Deep Junior or PConners. That's pretty >close to a fact, but it probably wouldn't be impossible to achieve with lower >rated opponents. However, results from the Israeli league suggest that Fritz >will have difficulties there too. Fritz on a AMD k6-2 450 isn't a 2700+ rated >player. > I agree with you. Most likely score for 450 is between about 2500 and 2580, 2600 would be a surprise. >>I would not be surprised if other top programs on the list achieved >>a 2600 or higher TPR (or a much lower one) based on the ratings. > >Well, I would. > >>Ok, I can accept that you do not like the association, you mentioned >>educated guess, how would I do that if not from a testing and stats >>point of view? > >The normal approach is games played, but it should be possible to reach a >reasonable estimate by testsuites alone as long as you take into account that >they're probably inflated. > Some very good points and I agree with you. I have not talked about test suites, but they are another good source for strength estimates and so is analysis of games by experts. All three make better predictors together than any one by it self. :) >Best wishes... >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.