Author: Chessfun
Date: 09:17:38 12/26/05
Go up one level in this thread
On December 26, 2005 at 11:50:52, Sune Larsson wrote: >On December 26, 2005 at 11:12:53, Chessfun wrote: > >>On December 26, 2005 at 08:50:21, Sune Larsson wrote: >>> >>> Why not settle this once and for all by making a *Ratinglist for Endings*? >>> >>> There are a lot of testers out there with good hardware so I think it will >>> be possible. Actually I mailed Heinz van Kempen about this. Such a ratinglist >>> would be very interesting, IMO. >> >> >>While a member of CEGT I proposed something along those lines. I think it is >>actually a good idea to make an endgame ratinglist, none currently that I'm >>aware of are doing such a thing. >> >> >>> As a start I suggest the Nunn positions. There are 10 of them but n:o 3 is not >>> valid since there are only 5 pieces in that position. Short description: >>> >>> 1+4 are rook endings >>> 2 is pawn ending >>> 5 is queen ending >>> 6 is knight ending >>> 7 is bishop vs knight ending >>> 8 is opposite coloured bishops ending >>> 9 is bishop ending (same colour) >>> 10 is rooks+opposite coloured bishops ending >>> >>> I suggest 40 moves in 40 minutes. Additional positions are of course >>> possible - like "good knight vs bad bishop" - themes like "rook behind the >>> pawn" - "activity in rookendings" etc. >>> >>> I have played several games from NunnEnding1 - and you'll be surprised of >>> how revealing this simple position is! >> >>I agree with Nunn as a start you would as you say want to add additional >>positions however it isn't so easy to pick them. I like the time controls on >>fast hardware all testers should have 5 piece tablebases. >> >>Personally I can't see myself doing it right now but it is an interesting >>subject. >> >>Sarah. > > > Hi Sarah, > > you wrote "While a member of CEGT I proposed something along those lines." > Does that mean that you are no longer a member of CEGT? Hi Sune, Correct as Heinz has previously mentioned there was a split in CEGT. > I mailed Heinz about this idea - but no answer yet. I think its a great idea however like anything there is only so much that can be done. > Anyway, glad that you find the idea interesting. That way we might also > follow the progress/added endgame knowledge in Rybka. Very true. It would also add to what has been the debate on the value of tablebases by using some engines that don't use them but have identical versions that do. Fruit 2.2 and 2.2.1 and Loop List come to mind. > There are 9 positions to start with. I have some additionals also. Both > the "good knight vs bad bishop" are won for white so the question there > is whether the program gets the full point or lets half slip away. > > We can keep this idea alive anyhow - as you say - noone is doing this... Can you post the fen of all the positions you think should be included? Sarah.
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.