Computer Chess Club Archives

Search

Terms
Messages

Subject: Re: Testposition for any new Engines

Author: Uri Blass
Date: 08:23:30 12/30/04
On December 30, 2004 at 06:51:31, Eduard Nemeth wrote:

>This Testposition was my idea from a game Savtschenko-Naumann, (Analyse Nemeth)
>in RE St.21 NR. 4/99. 0–1
>
>I have test any engines with this position and many programs have problems with
>it ( German: z.B. Spike, Fruit ) !
>
>Here an analysis by Fruit 2.0:
>
>[D]8/7P/8/8/K2b4/p7/1k6/1B6 b - - 0 1
>
>Analysis by Fruit 2.0:
>
>1...Kxb1 2.Kxa3
>  =  (0.00)   Tiefe: 1/2   00:00:00
>1...Kxb1 2.Kxa3
>  =  (0.00)   Tiefe: 2/4   00:00:00
>1...Kxb1 2.Kxa3 Ka1
>  =  (0.00)   Tiefe: 3/6   00:00:00
>1...Kxb1 2.Kxa3 Ka1 3.Kb3
>  =  (0.00)   Tiefe: 4/6   00:00:00
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1
>  =  (0.00)   Tiefe: 5/8   00:00:00
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4
>  =  (0.00)   Tiefe: 6/10   00:00:00  1kN
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1
>  =  (0.00)   Tiefe: 7/10   00:00:00  3kN
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
>  =  (0.00)   Tiefe: 8/12   00:00:00  8kN
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
>  =  (0.00)   Tiefe: 9/13   00:00:01  14kN
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
>  =  (0.00)   Tiefe: 10/14   00:00:01  28kN
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
>  =  (0.00)   Tiefe: 11/15   00:00:01  47kN
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
>  =  (0.00)   Tiefe: 12/15   00:00:01  71kN
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
>  =  (0.00)   Tiefe: 13/16   00:00:01  106kN
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
>  =  (0.00)   Tiefe: 14/18   00:00:02  156kN
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
>  =  (0.00)   Tiefe: 15/19   00:00:02  224kN
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
>  =  (0.00)   Tiefe: 16/20   00:00:03  292kN
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
>  =  (0.00)   Tiefe: 17/22   00:00:03  405kN
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
>  =  (0.00)   Tiefe: 18/22   00:00:04  552kN
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
>  =  (0.00)   Tiefe: 19/27   00:00:05  785kN
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
>  =  (0.00)   Tiefe: 20/28   00:00:06  1170kN
>1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
>  =  (0.00)   Tiefe: 21/31   00:00:08  1804kN
>
>(Nemeth, Privat 30.12.2004)
>
>Here I waiting now >5 minutes (!) on my P3 600, and I see that Fruit 2.0 looks
>the move Bc3! as move 5/16 in Deep 21/38, but Fruit 2.0 needs not use this move
>in his HV!
>
>Why Fruit 2.0 cannot find here Bc3?

Fruit can find it
It simply needs more time.

2:57 is enough on A3000(512 mbytes hash)

surprisingly movei can find it very fast and needs only 1512 nodes to fail high
and 833422 nodes to produce a winning score of 9.8 pawns at depth 12.

I guess that the problem is wrong pruning of fruit thanks to some knowledge
about KBP vs KB draws that encourage the pruning.

New game,
8/7P/8/8/K2b4/p7/1k6/1B6 b - - 0 1

Analysis by Fruit 2.0:

1...Kxb1 2.Kxa3
  =  (0.00)   Depth: 1/2   00:00:00
1...Kxb1 2.Kxa3
  =  (0.00)   Depth: 2/4   00:00:00
1...Kxb1 2.Kxa3 Ka1
  =  (0.00)   Depth: 3/6   00:00:00
1...Kxb1 2.Kxa3 Ka1 3.Kb3
  =  (0.00)   Depth: 4/6   00:00:00
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1
  =  (0.00)   Depth: 5/8   00:00:00
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4
  =  (0.00)   Depth: 6/10   00:00:00  1kN
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1
  =  (0.00)   Depth: 7/10   00:00:00  3kN
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
  =  (0.00)   Depth: 8/12   00:00:00  8kN
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
  =  (0.00)   Depth: 9/13   00:00:00  14kN
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
  =  (0.00)   Depth: 10/14   00:00:01  28kN
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
  =  (0.00)   Depth: 11/15   00:00:01  47kN
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
  =  (0.00)   Depth: 12/15   00:00:01  71kN
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
  =  (0.00)   Depth: 13/16   00:00:01  106kN
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
  =  (0.00)   Depth: 14/18   00:00:01  156kN
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
  =  (0.00)   Depth: 15/19   00:00:01  224kN
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
  =  (0.00)   Depth: 16/20   00:00:02  292kN
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
  =  (0.00)   Depth: 17/22   00:00:02  405kN
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
  =  (0.00)   Depth: 18/22   00:00:02  552kN
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
  =  (0.00)   Depth: 19/27   00:00:02  785kN
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
  =  (0.00)   Depth: 20/28   00:00:02  1170kN
1...Kxb1 2.Kxa3 Ka1 3.Kb3 Kb1 4.Kb4 Ka1 5.Kb3
  =  (0.00)   Depth: 21/31   00:00:02  1804kN
1...Bc3 2.h8Q Bxh8 3.Kb4 Be5 4.Ka4 Bd6 5.Be4 a2 6.Kb5 a1Q 7.Bd5 Kc3 8.Kc6 Bg3
9.Kb5 Qb2+ 10.Kc6 Qa3 11.Bh1 Qa8+ 12.Kb5 Qxh1
  -+  (-13.50)   Depth: 21/53   00:02:57  269212kN
1...Bc3 2.h8Q Bxh8 3.Kb4 Be5 4.Ka4 Bd6 5.Be4 a2 6.Kb5 a1Q 7.Bb7 Kc3 8.Kb6 Qg1+
9.Kc6 Qc5+ 10.Kd7 Qc7+ 11.Ke6 Kd4 12.Bh1 Qe7+ 13.Kf5 Qe5+ 14.Kg6 Qe6+ 15.Kg5
Qe3+ 16.Kg6 Qg1+ 17.Kf5
  -+  (-13.77)   Depth: 22/60   00:06:37  601826kN
1...Bc3 2.h8Q Bxh8 3.Kb4 Be5 4.Ka4 Bd6 5.Be4 a2 6.Kb5 a1Q 7.Bb7 Kc3 8.Kc6 Bf4
9.Kd7 Kd3 10.Bf3 Qa7+ 11.Kc6 Qa8+ 12.Kb5 Qxf3 13.Kc5 Be5
  -+  (-13.98)   Depth: 23/64   00:11:06  1018895kN
1...Bc3 2.h8Q Bxh8 3.Kb4 Be5 4.Ka4 Bd6 5.Be4 a2 6.Kb5 a1Q 7.Bb7 Kc3 8.Kc6 Bf4
9.Kd7 Qa7 10.Ke6 Qxb7 11.Kf5 Qf3 12.Ke6 Qe4+ 13.Kf6 Kd4
  -+  (-14.13)   Depth: 24/81   00:17:47  1635687kN

(, MyTown 30.12.2004)
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.