Author: Jouni Uski
Date: 04:06:16 05/02/02
Go up one level in this thread
On May 02, 2002 at 06:57:35, Chris Carson wrote: >On May 02, 2002 at 06:10:55, Jouni Uski wrote: > >>It seems, that SSDF 40/2 ratings need again 50-100 points lowering. Hmm. how >>many times that has been done already? And will oldest computers soon have >>negative >>rating then? Or does this prove, that comp-comp play exaggerates differences... >> >>Jouni > >It has been calibrated twice and changed once that I know about. > >I hope the old computers do not go negative, that would be a huge adjustment. ;) > >Calibration between different pools of participants may need to be recalibrated, >this is true for humans and machines. I am not sure the lists are out of sync, >I will take a look and come back with the results. > >I think the SSDF cosiders human results and compares to current SSDF list, my >guess is that some of the comps SSDF rating will be above the human ELO >performance and some will be below. We will see. Sadly abovementioned games were not 40/2. I quess, that at this level humans are better - for that reason I quess we need 50 points lowering. Jouni
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.