Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Endgame evaluation

Author: Anthony Cozzie

Date: 16:14:00 12/14/03

Go up one level in this thread


On December 14, 2003 at 19:00:39, Richard Pijl wrote:

>On December 14, 2003 at 18:33:49, Anthony Cozzie wrote:
>
>>On December 14, 2003 at 15:18:46, Richard Pijl wrote:
>>
>>>In a very recent game between Baron and Amateur (time control 45 10) on ICC, the
>>>following position occurred:
>>>
>>>[D]7k/2B2pp1/2n1b3/P2p2Pp/2p4n/4NP2/5K2/5B2 w - - 0 55
>>>
>>>White is two pawns down, both passers. As compensation, White has a pair of
>>>bishops. Understandably, Amateur showed a big plus score (+3.4), however The
>>>Baron thought is was almost drawn (-0.34). What do your engines think about
>>>this?
>>>
>>>The game ended in a draw after many more moves.
>>>Richard
>>
>>Zappa prefers black:
>>
>>1. a6 Kh7 2. Bb8 Kg6 3. f4 Nf5 4. a7 Nxa7 5. Bxa7 h4 6. Nxf5 Kxf5 7. Ke3 f6
>> = (-1.34)      Depth: 14/31    00:01:16.17     32035kN
>>
>
>I get the impression that this line is a draw line.
>Zappa sac's the knight at a7, let the remaining knight be exchanged. Black would
>have 3 pawns (all passers) against a bishop + the bishop pair advantage.
>But the connected passers are blocked by the king, no way to advance there
>without the assistance of the black king, and the h-pawn cannot advance without
>king assistance either.
>
>But it's good to see that also other engines see a significant 'compensation'
>for white for the two pawns, despite the connected passers.
>
>After seeing other engine's analysis, and a session with Dieter "Yace" Bürßner
>I'm starting to believe this is a drawn position.
>
>Richard.
>
>>anthony

Probably true.  Zappa evaluates its line as +1 because of the split passers.  I
think if (in the position at the end of its search) the knights were all
switched with bishops black would win.  Maybe I should add some code where "if
pieces remaining == bishops" -> "value passed pawns/split passed pawns less".
Still there is no way the position is +3 for black.

anthony



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.