Author: Richard Pijl
Date: 13:40:45 05/16/04
Go up one level in this thread
On May 16, 2004 at 15:01:52, Mike Byrne wrote:
>On May 16, 2004 at 11:59:21, Richard Pijl wrote:
>
>>On May 16, 2004 at 07:46:09, Dieter Buerssner wrote:
>>
>>>This is from yesterdays ICC tourney. I think, especially without TBs, the
>>>positions might be rather hard.
>>>
>>>[D] 8/1R6/8/7P/2k3P1/1p4K1/8/7r w - - 8 97
>>>
>>>Here Baron played Rb8, allowing a draw. Rb6 wins (I think, everything else
>>>draws).
>>
>> 10(22). 0:42.24 3273271 +1.26 Rb7-b6 Kc4-c3 Kg3-g2 Rh1-h4 Rb6-c6 Kc3-d2
>> Kg2-g3 Rh4-h1 Rc6-d6 Kd2-c3 Rd6-b6 b3-b2
>> (Kg3-f2)
>
>despite the score - the position reached at the end of the pv here comes to a
>dead draw position
>
agreed.
Baron's static eval knows that connected passers are very tough for a lone rook,
especially with the king far away. And alternatives may have already shown a
draw ...
The next ply though, it sees the win. If you look at the timings, those are
interesting as well:
The same machine (P4-2.4, 512MB), same hash settings (120 MB, 32Mb cache),
learning off, but the windows file cache had the relevant EGTB files stored:
10(22). 0:09.83 3273271 +1.26 Rb7-b6 Kc4-c3 Kg3-g2 Rh1-h4 Rb6-c6 Kc3-d2
Kg2-g3 Rh4-h1 Rc6-d6 Kd2-c3 Rd6-b6 b3-b2
(Kg3-f2)
11(21)+ 0:10.93 3676563 +1.61 Rb7-b6 (Kc4-c3) (h5-h6) (b3-b2) (g4-g5)
(Rh1-h4)
(Kg3xh4) (Kc3-c2) (Kh4-h5) (Kc2-c3)
(g5-g6) (Kc3-c2)
(h6-h7)
11(23)& 0:14.72 4968620 +2.45 Rb7-b6 Kc4-c3 h5-h6 b3-b2 g4-g5 Rh1-g1
Kg3-f4 Rg1-f1 Kf4-e5 Rf1-e1 Ke5-f5 Re1-f1
Kf5-g6 Rf1-a1 h6-h7 Ra1-h1 Kg6-g7
Richard.
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.