Author: Detlef Pordzik
Date: 16:45:34 11/23/98
Go up one level in this thread
On November 23, 1998 at 18:57:15, Amir Ban wrote: >On November 23, 1998 at 18:47:05, Detlef Pordzik wrote: > > >>Do you try this the other way round, too ? >>REBEL on 200 MMX ? JUNIOR on P 90 ? >>If yes, that's better, >>if no, what do you assume, is the worthy of such games, at all ? Using a new >>prog with newer algorithms, larger book and sure more hash tables - against an >>elder combattant, maybe with a much smaller, clearly elder book and less hash >>tables ? >>Not to speak of the PC version. I don't mind, if you have these 2, but it would >>be only fair to change machine every xyz games, wouldn't it ? >>Note : I don't mean you - in person - but shouldn't you be aware, that this kind >>of ridiculius setups influence the ELO ranking - in a missleading way ? >> >>c/u >> >>ELVIS > >I think you have a misunderstanding on how the rating system works. Rebel8/P90 >has a rating of round 2440. Junior has to win about 2-to-1 to stay in its >current rank. > >The SSDF also ran a match against Comet32/P90. Comet has a rating of around >2200, so Junior won 17.5-2.5 but still lost points on this. It had to win 18-2. >Rather than thank the SSDF for this "easy" match, I asked them not to run >matches with such a big rating spread (they answered they have a 400-difference >limit). > >Amir Great thing, if it were this way - this would nivellate reasonable. Away from this, how would you estimate the influence of the other things I've mentioned - such as RAM and books ? I think I've seen quite a few results 18 : 2 and so on. And, be confident : I missunderstand nothing - at least not, as what this list is concerned. I know and watch it from it's VERY beginning ! And this is quite a time by now, ain't it ?
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.