Author: blass uri
Date: 09:50:20 11/28/99
Go up one level in this thread
On November 28, 1999 at 09:56:52, Robert Hyatt wrote: >That doesn't matter. You release your program on September 1. I get it and >study it for 3 months and release my program on December 1. Since everyone is >on a "one release per year" sort of schedule, for 9 months I have an advantage. I do not think that it is so easy to do in 3 monthes a 25 elo improvement(the differenece between tiger and number 2). You can look at the improvement from Fritz5 to Fritz532 they had more than 3 monthes and they did only 13 elo improvement. Nimzo732 had also more than 3 monthes and they did only 19 elo improvement. Junior6 had more than one year and I guess it is not going to do more than 100 ssdf elo improvement. > >That's the way it has gone with commercial program on the SSDF for years... > >My only comment is that "he who releases last has a measurable advantage" and >nothing more than that. I agree but in this case there are also disadvanteges. The dos version is slower by 8-10% relative to the windows version because the windows version has a better compiler so I do not think that the final result is going to be clearly different. Tiger also does not use nalimov tablebases and it could probably earn 10-20 ssdf elo rating from these tablebases. Uri
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.