Author: Komputer Korner
Date: 10:49:29 09/30/98
Go up one level in this thread
On September 29, 1998 at 14:32:27, William H Rogers wrote: >On September 29, 1998 at 14:26:53, John Coffey wrote: > >>On September 29, 1998 at 14:00:40, William H Rogers wrote: >> >>>Has anyone done any research on the ratings that a program has at different >>>plys? >>>I once saw a test of chess 4.7 and they had some preliminary ratings on their >>>program at different ply levels. They could set the depth that their program >>>would search, not necessarily based on time. >>>bill >> >> >>I have seen in several sources that 1 extra ply = 1 rating class, but >>I don't know how true this really is. It is hard for me to imagine that >>it would be true at the top level. i.e. if 11 ply = an IM, then 12 ply = a GM? >> >>John Coffey > >The ratings I saw were large at the lower plys but got smaller at higher plys. >I think that Larry Kauffman? did an article on it once in his computer reports. >Chess 4.7 ratings jumped 200 to 300 points at lower plys but settled down to >about 100 or so at higher levels. >Bill Well now that the Pentium 450s are out it's speed over a Pentium 200 should be around 3 times, so it should be easy to measure the differences in its ratings against slower computers. It would be interesting whether the 75 point difference will hold. My guess is that the difference will drop to 50 points. Another very important difference and caveat is that computer vs computer at different hardware speeds exaggerates the difference. The real test would be games at 40/2 against humans. Failing that we need games in 30 at different hardware speeds against humans. I say games in 30 because by now, computer programs have surpassed even the best GMs at faster time controls. -- Komputer Korner
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.