Endgame test position

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Rebel, chrisw

Arpad Rusz
Posts: 273
Joined: Sat Apr 17, 2010 2:34 pm
Location: Budapest

Endgame test position

Post by Arpad Rusz »

Is there an engine showing a draw score for this position?

[D]8/2p2p2/2P5/8/1p6/1P4p1/p3P1Pp/B1k4K w - - 0 1
Top
zullil
Posts: 6442
Joined: Tue Jan 09, 2007 12:31 am
Location: PA USA
Full name: Louis Zulli

Re: Endgame test position

Post by zullil »

Arpad Rusz wrote:Is there an engine showing a draw score for this position?

[D]8/2p2p2/2P5/8/1p6/1P4p1/p3P1Pp/B1k4K w - - 0 1
Apparently not Stockfish:

Code: Select all

info depth 58 seldepth 96 score cp -8924 nodes 1847035524 nps 7955222 time 232179 multipv 1 pv e2e4 c1c2 e4e5 c2d3 a1b2 d3d2 b2d4 d2c1 d4a1 c1c2 a1d4 c2b1 e5e6 f7e6 d4f6 b1c1 f6a1 c1d1 a1e5 d1c2 e5b2 c2d2 b2h8 d2d3 h8a1 d3e4 a1d4 e4d5 d4e5 d5c6 e5a1 c6d5 a1d4 e6e5 d4a1 e5e4 a1e5 c7c5 e5g3 a2a1q h1h2 e4e3 h2h3 e3e2 h3g4 a1g7 g4f3 g7g3 f3g3 e2e1q g3h3 e1c3 h3h4 c3b3 g2g4 b3f3 h4g5 f3e3 g5h5 b4b3 h5g6 b3b2
Top
peter
Posts: 3186
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: Endgame test position

Post by peter »

Arpad Rusz wrote:Is there an engine showing a draw score for this position?

[D]8/2p2p2/2P5/8/1p6/1P4p1/p3P1Pp/B1k4K w - - 0 1
Well, Arpad, not having tried long time computing with 6MOB on and nullmove off, but as far as I've seen till now, I'd say no.
:)
That's a real good one again, congrats to the author
Peter.
Top
zullil
Posts: 6442
Joined: Tue Jan 09, 2007 12:31 am
Location: PA USA
Full name: Louis Zulli

Re: Endgame test position

Post by zullil »

Arpad Rusz wrote:Is there an engine showing a draw score for this position?

[D]8/2p2p2/2P5/8/1p6/1P4p1/p3P1Pp/B1k4K w - - 0 1
Critter shows mate-in-21:

Code: Select all

36/89  313:16 242008272892 12875287   -M21 e4 Kc2 e5 Kd3 Bb2 Ke4 Ba1 Kd5 Bb2 Kxc6 Ba1 Kd5 Bb2 c5 Ba1 Ke4 Bb2 c4 e6 fxe6 bxc4 b3 Ba1 Kd3 Be5 Kc2 Bxg3 a1=Q+ Kxh2 Qf6 Kh3 b2 Bh4 Qf5+ Kg3 b1=Q Kh2 Qf4+ g3 Qf3 c5 Qbh1#
Top
User avatar
Mike S.
Posts: 1480
Joined: Thu Mar 09, 2006 5:33 am

Re: Endgame test position

Post by Mike S. »

WinCHEST Ver.3.19i+, 08-Mar-2011
Options = -M64 -Z13 -eL -U -0 -5 -rS
Input file: STDIN
Reading job:
% created by ChestUCI Ver.5.2
W: Kh1 Ba1 Pb3 Pc6 Pe2 Pg2 (6)
B: Kc1 Pa2 Pb4 Pc7 Pf7 Pg3 Ph2 (7)
FEN: 8/2p2p2/2P5/8/1p6/1P4p1/p3P1Pp/B1k4K w - -
analysing (selfstalemate in 13 moves):
No solution in 13 moves.
Time (virt) = 543.553 sec (ca. 9.1 min)

FEN: 8/2p2p2/2P5/8/1p6/1P4p1/p3P1Pp/B1k4K w - - (6+7)
Position-Analysis: C0/R0/K5/P2/X6 W:2/9
Search for SelfStalemate in 13 ... (Hash=64MB)
Search completed ... (Time=543.55s)
No SelfStalemate in 13 found ! (09:03)
best move:---- time: 9:03.773 min n/s: 505.742 nodes: 274.898.011

:mrgreen:
Regards, Mike
Top
User avatar
Houdini
Posts: 1471
Joined: Tue Mar 16, 2010 12:00 am

Re: Endgame test position

Post by Houdini »

zullil wrote:
Arpad Rusz wrote:Is there an engine showing a draw score for this position?

[D]8/2p2p2/2P5/8/1p6/1P4p1/p3P1Pp/B1k4K w - - 0 1
Critter shows mate-in-21:

Code: Select all

36/89  313:16 242008272892 12875287   -M21 e4 Kc2 e5 Kd3 Bb2 Ke4 Ba1 Kd5 Bb2 Kxc6 Ba1 Kd5 Bb2 c5 Ba1 Ke4 Bb2 c4 e6 fxe6 bxc4 b3 Ba1 Kd3 Be5 Kc2 Bxg3 a1=Q+ Kxh2 Qf6 Kh3 b2 Bh4 Qf5+ Kg3 b1=Q Kh2 Qf4+ g3 Qf3 c5 Qbh1#
Confirmed by Houdini:

Code: Select all

30/64  1:15   -9.34    1.e4 Kc2 2.e5 Kxb3 3.e6 fxe6 4.Bd4 Kc4 
                       5.Bb2 Kd5 6.Bd4 Kxc6 7.Ba1 Kd5 8.Bb2 c5 
                       9.Ba1 c4 10.Be5 b3 11.Ba1 Ke4 12.Bb2 Kd3 
                       13.Be5 Kd2 14.Bc3+ (647.237.067) 8551 
                       TB:13.830 

31/66  1:49   -9.73--  1.e4 Kc2 (941.716.488) 8614 
                       TB:27.980 

31/66  1:53   -10.51-- 1.e4 Kc2 (973.804.366) 8571 
                       TB:28.014 

31/66  2:00   -12.73-- 1.e4 Kc2 (1.031.636.857) 8528 
                       TB:28.071 

31/68  2:32   -19.31-- 1.e4 Kc2 (1.305.536.894) 8571 
                       TB:28.197 

31/72  3:44   -M36     1.e4 Kc2 2.e5 Kd3 3.e6 fxe6 4.Be5 Ke4 
                       5.Ba1 Kd5 6.Bd4 Kxc6 7.Be5 Kd5 
                       8.Bxg3 a1Q+ 9.Kxh2 Kc6 10.Bh4 Qa2 
                       11.Be7 Qxb3 12.Bf6 Qc2 13.Be5 b3 
                       14.Ba1 (1.970.703.928) 8777 
                       TB:28.990 

32/73  6:02   -M21     1.e4 Kc2 2.e5 Kd3 3.e6 fxe6 4.Be5 Ke4 
                       5.Ba1 Kd5 6.Bd4 Kxc6 7.Be5 Kd5 
                       8.Bxg3 a1Q+ 9.Kxh2 Qh8+ 10.Kg1 Qc3 
                       11.Kh2 Qxb3 12.Bh4 Qc3 13.Be7 Ke4 
                       14.Bf8 (3.254.589.467) 8979 
                       TB:39.601 

33/73  7:25   -M21     1.e4 Kc2 2.e5 Kd3 3.e6 fxe6 4.Be5 Ke4 
                       5.Ba1 Kd5 6.Bd4 Kxc6 7.Be5 Kd5 8.Bb2 c5 
                       9.Ba1 Ke4 10.Bb2 c4 11.bxc4 Kd3 
                       12.c5 b3 13.Be5 Kc2 14.Bxg3 (4.027.967.337) 9047 
                       TB:61.142
The black king can grab pawn c6, after which c7-c5-c4 removes the white stalemate.
Probably not the result Arpad expected...
Top
zullil
Posts: 6442
Joined: Tue Jan 09, 2007 12:31 am
Location: PA USA
Full name: Louis Zulli

Re: Endgame test position

Post by zullil »

Houdini wrote: Confirmed by Houdini:

Code: Select all

30/64  1:15   -9.34    1.e4 Kc2 2.e5 Kxb3 3.e6 fxe6 4.Bd4 Kc4 
                       5.Bb2 Kd5 6.Bd4 Kxc6 7.Ba1 Kd5 8.Bb2 c5 
                       9.Ba1 c4 10.Be5 b3 11.Ba1 Ke4 12.Bb2 Kd3 
                       13.Be5 Kd2 14.Bc3+ (647.237.067) 8551 
                       TB:13.830 

31/66  1:49   -9.73--  1.e4 Kc2 (941.716.488) 8614 
                       TB:27.980 

31/66  1:53   -10.51-- 1.e4 Kc2 (973.804.366) 8571 
                       TB:28.014 

31/66  2:00   -12.73-- 1.e4 Kc2 (1.031.636.857) 8528 
                       TB:28.071 

31/68  2:32   -19.31-- 1.e4 Kc2 (1.305.536.894) 8571 
                       TB:28.197 

31/72  3:44   -M36     1.e4 Kc2 2.e5 Kd3 3.e6 fxe6 4.Be5 Ke4 
                       5.Ba1 Kd5 6.Bd4 Kxc6 7.Be5 Kd5 
                       8.Bxg3 a1Q+ 9.Kxh2 Kc6 10.Bh4 Qa2 
                       11.Be7 Qxb3 12.Bf6 Qc2 13.Be5 b3 
                       14.Ba1 (1.970.703.928) 8777 
                       TB:28.990 

32/73  6:02   -M21     1.e4 Kc2 2.e5 Kd3 3.e6 fxe6 4.Be5 Ke4 
                       5.Ba1 Kd5 6.Bd4 Kxc6 7.Be5 Kd5 
                       8.Bxg3 a1Q+ 9.Kxh2 Qh8+ 10.Kg1 Qc3 
                       11.Kh2 Qxb3 12.Bh4 Qc3 13.Be7 Ke4 
                       14.Bf8 (3.254.589.467) 8979 
                       TB:39.601 

33/73  7:25   -M21     1.e4 Kc2 2.e5 Kd3 3.e6 fxe6 4.Be5 Ke4 
                       5.Ba1 Kd5 6.Bd4 Kxc6 7.Be5 Kd5 8.Bb2 c5 
                       9.Ba1 Ke4 10.Bb2 c4 11.bxc4 Kd3 
                       12.c5 b3 13.Be5 Kc2 14.Bxg3 (4.027.967.337) 9047 
                       TB:61.142
The black king can grab pawn c6, after which c7-c5-c4 removes the white stalemate.
Probably not the result Arpad expected...
To err is human.

Only six minutes to find the minimal mate? Very impressive. Critter took much longer.
Top
User avatar
Mike S.
Posts: 1480
Joined: Thu Mar 09, 2006 5:33 am

Re: Endgame test position

Post by Mike S. »

Maybe it was intended to test, if any engine would wrongly show 0.00 although it isn't. :mrgreen: Otherwise, the c-pawns seem useless anyway, so I removed them:

[D]8/5p2/8/8/1p6/1P4p1/p3P1Pp/B1k4K w - -
Using backwards analysis in a variation starting with 1.e4 Kc2 2.e5 Kxb3 (which makes of course more sense than the engine's pv in the end), I could get Stockfish to verify a draw:

Engine: Stockfish 2.1.1 GTB (128 MB)
by Tord Romstad, Marco Costalba and Joona

41/02 0:01 0.00 1.e4 Kb1 2.Bd4 f6 3.e5 fxe5 4.Bxe5 Kc2
5.Bh8 Kxb3 6.Ba1 Kc4 7.Bb2 Kc5 8.Bg7 Kd5
9.Bb2 b3 10.Bf6 Ke4 11.Bd4 Kd3
12.Bf6 Ke2 13.Bb2 Kf1 14.Bg7 (842.099) 428 TB:1

(...)

54/63 2:54 0.00 1.e4 Kb1 2.Bd4 f6 3.e5 fxe5 4.Bxe5 Kc2
5.Bh8 Kxb3 6.Ba1 Kc4 7.Bb2 Kc5 8.Bg7 Kd5
9.Bb2 b3 10.Bf6 Ke4 11.Bd4 Kd3
12.Ba1 b2 13.Bxb2 Kc4 14.Bd4 (67.400.448) 387 TB:1

best move: e2-e4 time: 2:54.097 min n/s: 387.143 nodes: 67.400.448 TB: 1

(Intel Atom N455/1.66 GHz only, 128 MB hash)
Regards, Mike
Top
Arpad Rusz
Posts: 273
Joined: Sat Apr 17, 2010 2:34 pm
Location: Budapest

Re: Endgame test position

Post by Arpad Rusz »

Houdini wrote:The black king can grab pawn c6, after which c7-c5-c4 removes the white stalemate.
Probably not the result Arpad expected...
Yes, that's a surprise to me! :lol:
(I was adding two pawns in the last minute before I have posted the position...)

Here the corrected position:

[D]8/3p1p2/3P4/8/1p6/1P4p1/p3P1Pp/B1k4K w - - 0 1
Top
Arpad Rusz
Posts: 273
Joined: Sat Apr 17, 2010 2:34 pm
Location: Budapest

Re: Endgame test position

Post by Arpad Rusz »

A little study:

[D]8/8/5p2/4p3/1p2P3/1P4p1/p5Pp/2B2k1K w - - 0 1

White's third move is surprising:
1.Bb2 Ke2 2.Bxe5! Kd3 (2...fxe5 stalemate) 3.Ba1! (3.Bxf6? Kc2! -+) 3...Kc2 (3...Kxe4 4.Bxf6 =) 4.e5 fxe5 5.Bxe5 =
Top

Return to “Computer Chess Club: General Topics”