Author: Bruce Moreland
Date: 11:05:59 06/16/01
Go up one level in this thread
On June 16, 2001 at 05:17:38, Gian-Carlo Pascutto wrote: >On June 16, 2001 at 00:58:38, Bruce Moreland wrote: > >>When you play games you figure out how much these differences matter in real >>games. >> >>This works for me, I think. > >Yes but it doesn't help much in getting a definition for GM so >that you can decide 'computers are GM strength'. > >Say you set GM strength to 2500 Elo. That would be a good mark >nowadays. > >If you let the computer play vs mixed opposition you would be >able to calculate an Elo rating for it. And see < or > 2500 > >But that 'mark' will shift over the years. Aditionally you >have changed the pool which you compare to (the humans elo >ratings are based on a pool without computers). The elo >of computerkillers would raise and those of 'open' players >would drop. > >You have done comparisations from one pool of players to >another. The comparisation is meaningless. I'm not talking about games with computers, I'm talking about games with humans. The SSDF list is a restricted pool and therefore anything is possible. The only way that computers will get real ratings is if they regularly play in the same pool with a lot of other players. bruce > >-- >GCP
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.