Author: Mike S.
Date: 10:47:35 06/30/02
Go up one level in this thread
On June 30, 2002 at 13:11:35, John Reynolds wrote: >Someone told me 2450 is this even close? Tony Hedlund has a list of Man vs. Machine results: http://home.interact.se/~w100107/manmachine.htm The performances <2500 elo are mostly from very old engine versions on outdated CPUs. I think, a good example is the Tiger performance of that Argentine tournament (in comp-comp, Tiger 14 is very close to Fritz' performance). That was 2788 on 866 MHz, against "medium" GMs /IMs. If the opponents are skilled in anticomputer chess, computer performance will be somewhat less... but I wouldn't expect much less than 2600. In a recent match (but with shorter time controls, 60m+10s), GM Smirin beat some progs running on 2*1 GHz duals or on 2.2 GHz, 5-3. I didn't calculate the computer's elo performance exactly, but it is >2600 (Smirin is ~2700 elo). In a previous similar event, GM Gulko (2605 elo) lost 3-5. Probably humans can do a bit better in tournament time controls, but not very much better IMO. If I would have to guess, I'd say >2500 in each case, very likely above 2600. The CT 14 results indicates it. (It won't matter much, which of the top programs it is, against humans. For example, Smirin beat Shredder with a copy of a Fritz loss.) Regards, M.Scheidl
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.