Author: Peter McKenzie
Date: 14:47:23 08/08/02
Go up one level in this thread
On August 08, 2002 at 17:28:38, Mike S. wrote: >We've seen from that latest GM tournament in Argentinia, that current top >software can probably perform ~2600 "anytime" (single CPU, 1 GHz) under >difficult circumstances, too: The opponents had been warned by Tiger's >spectacular success in 2001... but still the program, Hiarcs 8 this time, was a >big danger for them. But the elo performance was ~2580 (only). > >Hiarcs 8 and Junior 7 aren't that much different in strength against humans IMO, >like all the top programs. In addition to SSDF (where Hiarcs is a bit behind), >there are rating lists where it is ahead of Junior. > >So, based on Hiarcs' performance in Argentinia, we can try to add estimations of >how much the latest Junior version *could* (not necessarily must be) better than >Hiarcs 8, and add another estimation for the faster hardware: > >2580 + 20 + 50 = 2650 (just rough estimations of course, but IMO reasonable). > >Now, I've taken a look into my database to see how Kasparov usually scores >against opponents between 2630...2670 elo. I found, that since 1990 he scored >77% with white, and 66% with black against such opposition. Average = 71,5%. > >71,5% of 6 = 4,29 > >Which means, predictions of 4-2 or 4,5-1,5 seem very likely. If Kasparov manages >to apply anticomputer methods better than the GMs did against Hiarcs (I don't >know if and how good they did that), which is not unlikely that he can do, 5-1 >also wouldn't be surprising. > >My own first guess was 6-0 in case DJ plays on dual pc (only), but that was >"from stomach"... After reconsidering, I still think that result is possible >though - but not very likely to happen. What was the last "...to zero" result in >a comparable match? Must have been long ago. Deep Thought vs David Levy > >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.