Computer fail

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

Moderators: hgm, Rebel, chrisw

Uri Blass
Posts: 10360
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: Computer fail

Post by Uri Blass »

M ANSARI wrote:Engines don't see that as equal, for example Rybka gives a +.45 at depth 21 and this kind of evaluation in a position where black is totally passive can only mean an easy win. There are way too many iterations and possible black replies for the engine to look through every single line and give an extremely high evaluation. I am pretty sure most of the top engines would win this as white pretty easily.
The question is not if they can win it with white
The problem is that many programs see positive evaluations for black because of their poor mobility evaluation and they estimate the value of the black passed pawns as more than the value of the bad mobility of black.

Rybka3 starts by seeing it as advantage for black but
stockfish1.9 is even worse than Rybka3.

Stockfish1.9 see a positive score for black
and needs a significant time to see only a draw score(I used only single cpu but a good program should evaluate the position as better for white)

It seems clear to me that stockfish1.9 has the potential to improve because it seems to underevaluate big scores of mobility.

New game
r1b2rnk/pp3p1q/2pp1Pp1/4p1Pp/2B1P2Q/2NP2R1/PPP5/5R1K b - - 0 1

Analysis by Stockfish 1.9 JA:

1...Bc8-g4
³ (-0.48) Depth: 1 00:00:00
1...Bc8-g4 2.a2-a3
³ (-0.56) Depth: 2 00:00:00
1...Bc8-g4 2.a2-a4 Ra8-d8
³ (-0.56) Depth: 3 00:00:00
1...b7-b5 2.Bc4-b3 Bc8-g4
³ (-0.64) Depth: 3 00:00:00
1...b7-b5 2.Bc4-b3 Bc8-g4 3.Kh1-g2
µ (-0.72) Depth: 4 00:00:00
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-e2 Bc8-g4
µ (-0.76) Depth: 5 00:00:00
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-e2 Bc8-g4
µ (-0.84) Depth: 6 00:00:00 4kN
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-e2 Bc8-g4 4.Rg3-e3
µ (-0.88) Depth: 6 00:00:00 4kN
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-e2 Bc8-g4 4.Rg3-e3
µ (-0.72) Depth: 7 00:00:00 5kN
1...b7-b5 2.Bc4-b3 Bc8-g4 3.Kh1-g2 b5-b4 4.Nc3-d1 Bg4-e2 5.Rf1-e1
µ (-0.84) Depth: 7 00:00:00 9kN
1...b7-b5 2.Bc4-b3 a7-a5 3.a2-a3 Bc8-g4 4.Kh1-g2 b5-b4 5.a3xb4 a5xb4
µ (-0.84) Depth: 8 00:00:00 15kN
1...b7-b5 2.Bc4-b3 a7-a5 3.a2-a4 b5-b4 4.Nc3-d1 Bc8-g4 5.Nd1-e3
µ (-0.76) Depth: 9 00:00:00 22kN
1...b7-b5 2.Bc4-b3 a7-a5 3.a2-a4 b5-b4 4.Nc3-d1 Bc8-g4 5.Nd1-e3
³ (-0.68) Depth: 9 00:00:00 26kN
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-d1 Bc8-g4 4.Nd1-f2 Bg4-d7 5.Kh1-g2 Ra8-b8 6.Rg3-e3
³ (-0.60) Depth: 9 00:00:00 38kN
1...b7-b5 2.Bc4-b3 Ra8-b8 3.a2-a3 a7-a5 4.a3-a4 b5-b4 5.Nc3-d1 Bc8-g4 6.Nd1-e3 Bg4-e2
³ (-0.64) Depth: 10 00:00:01 81kN
1...b7-b5 2.Bc4-b3 Ra8-b8 3.a2-a3 Bc8-d7 4.Rg3-g1 a7-a5 5.Kh1-g2 Bd7-g4 6.Rg1-h1 a5-a4 7.Bb3-a2
³ (-0.60) Depth: 11 00:00:01 132kN
1...b7-b5 2.Bc4-b3 Ra8-b8 3.Kh1-g2 Rf8-d8 4.a2-a4 b5-b4 5.Nc3-d1 d6-d5 6.e4xd5 c6xd5 7.Rg3-e3
³ (-0.52) Depth: 12 00:00:01 190kN
1...b7-b5 2.Bc4-b3 Ra8-b8 3.Kh1-g2 Rf8-d8 4.a2-a4 b5-b4 5.Nc3-d1 d6-d5 6.e4xd5 c6xd5 7.Rg3-e3
³ (-0.44) Depth: 12 00:00:01 279kN
1...b7-b5 2.Bc4-b3 Ra8-b8 3.Kh1-g2 Rf8-d8 4.a2-a4 a7-a6 5.a4-a5 Bc8-d7 6.Nc3-e2 Rd8-c8 7.Rf1-h1 d6-d5 8.d3-d4
³ (-0.48) Depth: 12 00:00:01 337kN
1...b7-b5 2.Bc4-b3 Ra8-b8 3.Kh1-g2 Bc8-d7 4.Rf1-h1 Rf8-d8 5.a2-a3 a7-a5 6.a3-a4 b5-b4 7.Nc3-d1 d6-d5 8.e4xd5 c6xd5 9.Bb3xd5 Bd7xa4
³ (-0.52) Depth: 13 00:00:01 398kN
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-d1 a7-a5 4.Bb3-a4 Ra8-a6 5.Nd1-e3 Bc8-e6 6.Ba4-b3 Be6xb3 7.a2xb3 a5-a4 8.b3xa4 Ra6xa4 9.Ne3-c4
³ (-0.52) Depth: 14 00:00:01 500kN
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-d1 a7-a5 4.Bb3-a4 Ra8-a6 5.Nd1-e3 Rf8-d8 6.c2-c3 b4xc3 7.b2xc3 Bc8-e6 8.Ba4-b3 Be6xb3 9.a2xb3 Rd8-b8
³ (-0.60) Depth: 15 00:00:01 663kN
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-d1 a7-a5 4.Bb3-a4 Ra8-a6 5.Nd1-e3 Rf8-d8 6.Kh1-g2 Bc8-e6 7.Ne3-c4 Be6xc4 8.d3xc4
³ (-0.44) Depth: 15 00:00:01 733kN
1...b7-b5 2.Bc4-b3 Bc8-d7 3.Kh1-g2 Ra8-b8 4.Rf1-h1 Rf8-d8 5.a2-a3 a7-a5 6.a3-a4 b5-b4 7.Nc3-d1 d6-d5 8.e4xd5 c6xd5 9.Bb3xd5 Bd7xa4
³ (-0.52) Depth: 15 00:00:02 985kN
1...b7-b5 2.Bc4-b3 Bc8-d7 3.Kh1-g2 Ra8-b8 4.Rf1-h1 Rf8-d8 5.a2-a3 a7-a5 6.Rg3-e3 b5-b4 7.Nc3-e2 b4xa3 8.b2xa3
³ (-0.44) Depth: 16 00:00:04 1374kN
1...b7-b5 2.Bc4-b3 Bc8-d7 3.Kh1-g2 a7-a6 4.Rf1-h1 Rf8-b8 5.a2-a4 Ra8-a7 6.Rh1-a1 Ra7-c7 7.a4xb5 a6xb5 8.Nc3-d1 d6-d5 9.Nd1-e3 d5xe4 10.Qh4xe4
³ (-0.44) Depth: 16 00:00:05 1683kN
1...b7-b5 2.Bc4-b3 Bc8-d7 3.Kh1-g2 a7-a6 4.a2-a4 Rf8-b8 5.a4-a5 b5-b4 6.Nc3-d1 Bd7-g4 7.Nd1-e3
³ (-0.36) Depth: 17 00:00:05 2401kN
1...b7-b5 2.Bc4-b3 Bc8-d7 3.Kh1-g2 a7-a6 4.a2-a4 Ra8-c8 5.Rf1-a1 Rf8-d8 6.Nc3-b1 Rc8-a8 7.Nb1-c3 b5-b4 8.Nc3-d1 Ra8-a7 9.Nd1-e3 Ra7-b7 10.Rg3-f3
³ (-0.36) Depth: 17 00:00:06 2675kN
1...b7-b5 2.Bc4-b3 Bc8-b7 3.a2-a4 a7-a6 4.Kh1-g2 Rf8-d8 5.Rf1-h1 Ra8-b8 6.Rg3-f3 Bb7-c8 7.Rf3-g3 Bc8-d7 8.Rh1-a1 Rd8-c8 9.Nc3-d1 Rc8-c7 10.Nd1-e3 b5xa4 11.Ra1xa4
³ (-0.32) Depth: 18 00:00:08 4589kN
1...b7-b5 2.Bc4-b3 Bc8-b7 3.a2-a4 a7-a6 4.Kh1-g2 Rf8-d8 5.Rf1-h1 Ra8-b8 6.Rg3-f3 Bb7-c8 7.Rf3-f1 b5-b4 8.Nc3-d1
= (-0.24) Depth: 19 00:00:13 6381kN
1...b7-b5 2.Bc4-b3 Bc8-b7 3.a2-a4 a7-a6 4.Kh1-g2 Rf8-d8 5.Rg3-e3 Bb7-c8 6.Re3-g3 Bc8-d7 7.a4-a5 Ra8-b8 8.Nc3-d1 Bd7-c8 9.Nd1-e3 Rb8-b7 10.Rf1-h1 Bc8-d7 11.Kg2-f2 Rb7-b8 12.c2-c3
³ (-0.32) Depth: 19 00:00:24 9755kN
1...b7-b5 2.Bc4-b3 Bc8-b7 3.a2-a4 a7-a6 4.Kh1-g2 Rf8-d8 5.Rg3-e3 Bb7-c8 6.Re3-g3 Bc8-d7 7.Nc3-d1 b5xa4 8.Bb3xa4 Ra8-b8 9.Ba4-b3 Rb8-b4 10.Nd1-c3 d6-d5 11.Nc3-a2 Rb4-b8 12.e4xd5 c6xd5 13.Rg3-e3
= (-0.24) Depth: 20 00:00:27 11841kN
1...b7-b5 2.Bc4-b3 Bc8-b7 3.a2-a4 a7-a6 4.Kh1-g2 Rf8-d8 5.Rg3-e3 Bb7-c8 6.Re3-g3 Bc8-d7 7.Nc3-d1 b5xa4 8.Bb3xa4 Ra8-b8 9.Ba4-b3 Rb8-b4 10.Nd1-c3 d6-d5 11.Nc3-a2 Rb4-b8 12.e4xd5 c6xd5 13.Rg3-e3
³ (-0.32) Depth: 20 00:00:27 12196kN
1...b7-b5 2.Bc4-b3 Bc8-b7 3.a2-a4 a7-a6 4.Kh1-g2 Rf8-d8 5.Rg3-e3 Bb7-c8 6.Re3-g3 Bc8-d7 7.Nc3-d1 Ra8-c8 8.Nd1-e3 b5xa4 9.Bb3xa4 Bd7-e6 10.Rf1-a1 Rc8-c7 11.Ne3-c4 Be6xc4 12.d3xc4
= (-0.24) Depth: 21 00:00:31 14450kN
1...b7-b5 2.Bc4-b3 Bc8-b7 3.a2-a4 a7-a6 4.Kh1-g2 Rf8-d8 5.Rf1-a1 Rd8-b8 6.Ra1-h1 Rb8-e8 7.Rh1-a1 Re8-d8 8.Rg3-e3 Rd8-b8 9.Re3-f3 b5-b4 10.Nc3-d1 Bb7-c8 11.Nd1-e3 Bc8-d7 12.Qh4-e1
= (-0.24) Depth: 21 00:00:33 16002kN
1...b7-b5 2.Bc4-b3 Bc8-b7 3.a2-a4 a7-a6 4.Kh1-g2 Rf8-d8 5.Rf1-a1 Rd8-b8 6.Rg3-e3 b5-b4 7.Nc3-d1 a6-a5 8.Re3-e1 Rb8-d8 9.Nd1-e3 d6-d5 10.Re1-h1 Ra8-b8 11.Qh4-g3
= (-0.16) Depth: 22 00:00:44 23866kN
1...b7-b6 2.a2-a4 a7-a6 3.Bc4-b3 Ra8-b8 4.Kh1-g2 Bc8-d7 5.Nc3-d1 b6-b5 6.a4-a5 Rb8-d8 7.Nd1-e3 Bd7-c8 8.Rf1-h1 Rf8-e8 9.Rh1-e1 Bc8-b7 10.c2-c4 b5xc4 11.Ne3xc4 Re8-f8 12.Rg3-f3
= (-0.20) Depth: 22 00:00:58 32394kN
1...b7-b6 2.a2-a4 a7-a6 3.Bc4-b3 Ra8-b8 4.Kh1-g2 Bc8-d7 5.Nc3-d1 b6-b5 6.a4-a5 Rb8-d8 7.Nd1-e3 Bd7-c8 8.Rf1-h1 Rf8-e8 9.Rh1-e1 Bc8-e6 10.Bb3xe6 Re8xe6 11.b2-b4 d6-d5 12.c2-c3 d5-d4 13.c3xd4
= (-0.16) Depth: 23 00:01:02 35387kN
1...b7-b6 2.a2-a4 a7-a6 3.Bc4-b3 Ra8-b8 4.Kh1-g2 Bc8-d7 5.Rf1-a1 Rf8-e8 6.Nc3-d1 b6-b5 7.Nd1-e3 Rb8-a8 8.Ra1-h1 Re8-b8 9.a4-a5 Rb8-d8 10.Rh1-a1 Rd8-f8 11.Ra1-h1 Ra8-d8 12.Rh1-e1 b5-b4 13.c2-c3
= (-0.16) Depth: 24 00:01:08 39083kN
1...b7-b6 2.a2-a4 a7-a6 3.Bc4-b3 Ra8-b8 4.Kh1-g2 Bc8-d7 5.Rf1-a1 Rf8-e8 6.Nc3-d1 b6-b5 7.Nd1-e3 Rb8-a8 8.Ra1-h1 Re8-b8 9.a4-a5 Rb8-d8 10.Rg3-f3 Ra8-b8 11.Rf3-f1 Rd8-c8 12.Rf1-e1 Rc8-c7 13.Qh4-f2 Bd7-c8
= (-0.08) Depth: 25 00:01:36 55635kN
1...b7-b6 2.a2-a4 Rf8-e8 3.b2-b4 Bc8-e6 4.Bc4xe6 Re8xe6 5.Rg3-g2 Ra8-c8 6.b4-b5 c6xb5 7.Nc3xb5 a7-a6 8.Nb5-a3 Rc8-c5 9.Na3-c4 d6-d5 10.e4xd5 Rc5xd5 11.Qh4-e4 Rd5-d4 12.Qe4-e3
= (0.00) Depth: 25 00:02:54 103201kN
1...b7-b6 2.a2-a4 a7-a6 3.Bc4-b3 Ra8-b8 4.Kh1-g2 Bc8-d7 5.Rf1-a1 Rb8-a8 6.Nc3-d1 Rf8-b8 7.Nd1-e3 b6-b5 8.Rg3-f3 Rb8-c8 9.Rf3-f1 Rc8-b8 10.Rf1-h1 Rb8-d8 11.Qh4-f2 Rd8-e8 12.Qf2-h4 Re8-d8
= (0.00) Depth: 25 00:04:41 168709kN
1...b7-b6 2.a2-a4 a7-a6 3.Bc4-b3 Ra8-b8 4.Kh1-g2 Bc8-d7 5.Rf1-a1 Rb8-a8 6.Nc3-d1 Rf8-d8 7.Nd1-e3 Bd7-c8 8.Ne3-c4 Ra8-b8 9.Nc4-e3 Bc8-b7 10.Ne3-c4 Bb7-c8
= (0.00) Depth: 26 00:05:03 181189kN
1...b7-b6 2.a2-a4 a7-a6 3.Bc4-b3 Ra8-b8 4.Kh1-g2 Bc8-d7 5.Rf1-a1 Rb8-a8 6.Nc3-d1 Rf8-d8 7.Nd1-e3 Bd7-c8 8.Rg3-f3 Ra8-b8 9.Rf3-f1 Bc8-d7 10.Rf1-h1 Rb8-a8 11.Ne3-c4
= (0.08) Depth: 27 00:06:38 236917kN
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-d1 a7-a5 4.Bb3-a4 Ra8-a6 5.Nd1-e3 Rf8-e8 6.Ne3-c4 Re8-d8 7.Kh1-g2 Bc8-e6 8.Rf1-h1 Rd8-d7 9.Rh1-f1 Rd7-d8
= (0.00) Depth: 27 00:08:03 285432kN
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-d1 a7-a5 4.Bb3-a4 Ra8-a6 5.Nd1-e3 Bc8-e6 6.Ne3-c4 Be6xc4 7.d3xc4 Ra6-b6 8.Rg3-d3 Rf8-d8
= (0.00) Depth: 28 00:09:24 334679kN

(, 18.10.2010)
Top
De Vos W
Posts: 431
Joined: Tue Dec 01, 2009 11:59 am

Re: Computer fail

Post by De Vos W »

M ANSARI wrote:Engines don't see that as equal, for example Rybka gives a +.45 at depth 21 and this kind of evaluation in a position where black is totally passive can only mean an easy win. There are way too many iterations and possible black replies for the engine to look through every single line and give an extremely high evaluation. I am pretty sure most of the top engines would win this as white pretty easily.
Dear M ANSARI,

My Rybka 4 thinks this is equal! (bad mobility evaluation)
Holy Father, i see, You have "Rybka 4.1" and no bugs.....
Oh well! Cheers.
Gods are fragile things; they may be killed by a whiff of science or a dose of common sense.
Top
User avatar
M ANSARI
Posts: 3707
Joined: Thu Mar 16, 2006 7:10 pm

Re: Computer fail

Post by M ANSARI »

Dear Mr. De Vos W (DE VOS Walter):

What makes you think that my Rybka is Rybka 4.1 ??? The Rybka I tried with is RC 2 of Rybka 4 which is pretty much the same as the released version. The Rybka 4.1 or the "bug fixed" version is still not being tested, but hopefully will be very soon :wink:
Top
De Vos W
Posts: 431
Joined: Tue Dec 01, 2009 11:59 am

Re: Computer fail

Post by De Vos W »

M ANSARI wrote:Dear Mr. De Vos W (DE VOS Walter):

What makes you think that my Rybka is Rybka 4.1 ??? The Rybka I tried with is RC 2 of Rybka 4 which is pretty much the same as the released version. The Rybka 4.1 or the "bug fixed" version is still not being tested, but hopefully will be very soon :wink:
Come on Mr MANSARI, keep it cool i was only joking and we all know for the
"bug fixed version of Rybka 4 we wait and wait and wait ..... just like Rybka 3... congratulate Mr. Vasik Rajlich on that record in your next Email!!! :lol:
Gods are fragile things; they may be killed by a whiff of science or a dose of common sense.
Top
Robert Flesher
Posts: 1280
Joined: Tue Aug 18, 2009 3:06 am

Re: Computer fail

Post by Robert Flesher »

Uri Blass wrote:
M ANSARI wrote:Engines don't see that as equal, for example Rybka gives a +.45 at depth 21 and this kind of evaluation in a position where black is totally passive can only mean an easy win. There are way too many iterations and possible black replies for the engine to look through every single line and give an extremely high evaluation. I am pretty sure most of the top engines would win this as white pretty easily.
The question is not if they can win it with white
The problem is that many programs see positive evaluations for black because of their poor mobility evaluation and they estimate the value of the black passed pawns as more than the value of the bad mobility of black.

Rybka3 starts by seeing it as advantage for black but
stockfish1.9 is even worse than Rybka3.

Stockfish1.9 see a positive score for black
and needs a significant time to see only a draw score(I used only single cpu but a good program should evaluate the position as better for white)

It seems clear to me that stockfish1.9 has the potential to improve because it seems to underevaluate big scores of mobility.

New game
r1b2rnk/pp3p1q/2pp1Pp1/4p1Pp/2B1P2Q/2NP2R1/PPP5/5R1K b - - 0 1

Analysis by Stockfish 1.9 JA:

1...Bc8-g4
³ (-0.48) Depth: 1 00:00:00
1...Bc8-g4 2.a2-a3
³ (-0.56) Depth: 2 00:00:00
1...Bc8-g4 2.a2-a4 Ra8-d8
³ (-0.56) Depth: 3 00:00:00
1...b7-b5 2.Bc4-b3 Bc8-g4
³ (-0.64) Depth: 3 00:00:00
1...b7-b5 2.Bc4-b3 Bc8-g4 3.Kh1-g2
µ (-0.72) Depth: 4 00:00:00
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-e2 Bc8-g4
µ (-0.76) Depth: 5 00:00:00
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-e2 Bc8-g4
µ (-0.84) Depth: 6 00:00:00 4kN
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-e2 Bc8-g4 4.Rg3-e3
µ (-0.88) Depth: 6 00:00:00 4kN
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-e2 Bc8-g4 4.Rg3-e3
µ (-0.72) Depth: 7 00:00:00 5kN
1...b7-b5 2.Bc4-b3 Bc8-g4 3.Kh1-g2 b5-b4 4.Nc3-d1 Bg4-e2 5.Rf1-e1
µ (-0.84) Depth: 7 00:00:00 9kN
1...b7-b5 2.Bc4-b3 a7-a5 3.a2-a3 Bc8-g4 4.Kh1-g2 b5-b4 5.a3xb4 a5xb4
µ (-0.84) Depth: 8 00:00:00 15kN
1...b7-b5 2.Bc4-b3 a7-a5 3.a2-a4 b5-b4 4.Nc3-d1 Bc8-g4 5.Nd1-e3
µ (-0.76) Depth: 9 00:00:00 22kN
1...b7-b5 2.Bc4-b3 a7-a5 3.a2-a4 b5-b4 4.Nc3-d1 Bc8-g4 5.Nd1-e3
³ (-0.68) Depth: 9 00:00:00 26kN
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-d1 Bc8-g4 4.Nd1-f2 Bg4-d7 5.Kh1-g2 Ra8-b8 6.Rg3-e3
³ (-0.60) Depth: 9 00:00:00 38kN
1...b7-b5 2.Bc4-b3 Ra8-b8 3.a2-a3 a7-a5 4.a3-a4 b5-b4 5.Nc3-d1 Bc8-g4 6.Nd1-e3 Bg4-e2
³ (-0.64) Depth: 10 00:00:01 81kN
1...b7-b5 2.Bc4-b3 Ra8-b8 3.a2-a3 Bc8-d7 4.Rg3-g1 a7-a5 5.Kh1-g2 Bd7-g4 6.Rg1-h1 a5-a4 7.Bb3-a2
³ (-0.60) Depth: 11 00:00:01 132kN
1...b7-b5 2.Bc4-b3 Ra8-b8 3.Kh1-g2 Rf8-d8 4.a2-a4 b5-b4 5.Nc3-d1 d6-d5 6.e4xd5 c6xd5 7.Rg3-e3
³ (-0.52) Depth: 12 00:00:01 190kN
1...b7-b5 2.Bc4-b3 Ra8-b8 3.Kh1-g2 Rf8-d8 4.a2-a4 b5-b4 5.Nc3-d1 d6-d5 6.e4xd5 c6xd5 7.Rg3-e3
³ (-0.44) Depth: 12 00:00:01 279kN
1...b7-b5 2.Bc4-b3 Ra8-b8 3.Kh1-g2 Rf8-d8 4.a2-a4 a7-a6 5.a4-a5 Bc8-d7 6.Nc3-e2 Rd8-c8 7.Rf1-h1 d6-d5 8.d3-d4
³ (-0.48) Depth: 12 00:00:01 337kN
1...b7-b5 2.Bc4-b3 Ra8-b8 3.Kh1-g2 Bc8-d7 4.Rf1-h1 Rf8-d8 5.a2-a3 a7-a5 6.a3-a4 b5-b4 7.Nc3-d1 d6-d5 8.e4xd5 c6xd5 9.Bb3xd5 Bd7xa4
³ (-0.52) Depth: 13 00:00:01 398kN
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-d1 a7-a5 4.Bb3-a4 Ra8-a6 5.Nd1-e3 Bc8-e6 6.Ba4-b3 Be6xb3 7.a2xb3 a5-a4 8.b3xa4 Ra6xa4 9.Ne3-c4
³ (-0.52) Depth: 14 00:00:01 500kN
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-d1 a7-a5 4.Bb3-a4 Ra8-a6 5.Nd1-e3 Rf8-d8 6.c2-c3 b4xc3 7.b2xc3 Bc8-e6 8.Ba4-b3 Be6xb3 9.a2xb3 Rd8-b8
³ (-0.60) Depth: 15 00:00:01 663kN
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-d1 a7-a5 4.Bb3-a4 Ra8-a6 5.Nd1-e3 Rf8-d8 6.Kh1-g2 Bc8-e6 7.Ne3-c4 Be6xc4 8.d3xc4
³ (-0.44) Depth: 15 00:00:01 733kN
1...b7-b5 2.Bc4-b3 Bc8-d7 3.Kh1-g2 Ra8-b8 4.Rf1-h1 Rf8-d8 5.a2-a3 a7-a5 6.a3-a4 b5-b4 7.Nc3-d1 d6-d5 8.e4xd5 c6xd5 9.Bb3xd5 Bd7xa4
³ (-0.52) Depth: 15 00:00:02 985kN
1...b7-b5 2.Bc4-b3 Bc8-d7 3.Kh1-g2 Ra8-b8 4.Rf1-h1 Rf8-d8 5.a2-a3 a7-a5 6.Rg3-e3 b5-b4 7.Nc3-e2 b4xa3 8.b2xa3
³ (-0.44) Depth: 16 00:00:04 1374kN
1...b7-b5 2.Bc4-b3 Bc8-d7 3.Kh1-g2 a7-a6 4.Rf1-h1 Rf8-b8 5.a2-a4 Ra8-a7 6.Rh1-a1 Ra7-c7 7.a4xb5 a6xb5 8.Nc3-d1 d6-d5 9.Nd1-e3 d5xe4 10.Qh4xe4
³ (-0.44) Depth: 16 00:00:05 1683kN
1...b7-b5 2.Bc4-b3 Bc8-d7 3.Kh1-g2 a7-a6 4.a2-a4 Rf8-b8 5.a4-a5 b5-b4 6.Nc3-d1 Bd7-g4 7.Nd1-e3
³ (-0.36) Depth: 17 00:00:05 2401kN
1...b7-b5 2.Bc4-b3 Bc8-d7 3.Kh1-g2 a7-a6 4.a2-a4 Ra8-c8 5.Rf1-a1 Rf8-d8 6.Nc3-b1 Rc8-a8 7.Nb1-c3 b5-b4 8.Nc3-d1 Ra8-a7 9.Nd1-e3 Ra7-b7 10.Rg3-f3
³ (-0.36) Depth: 17 00:00:06 2675kN
1...b7-b5 2.Bc4-b3 Bc8-b7 3.a2-a4 a7-a6 4.Kh1-g2 Rf8-d8 5.Rf1-h1 Ra8-b8 6.Rg3-f3 Bb7-c8 7.Rf3-g3 Bc8-d7 8.Rh1-a1 Rd8-c8 9.Nc3-d1 Rc8-c7 10.Nd1-e3 b5xa4 11.Ra1xa4
³ (-0.32) Depth: 18 00:00:08 4589kN
1...b7-b5 2.Bc4-b3 Bc8-b7 3.a2-a4 a7-a6 4.Kh1-g2 Rf8-d8 5.Rf1-h1 Ra8-b8 6.Rg3-f3 Bb7-c8 7.Rf3-f1 b5-b4 8.Nc3-d1
= (-0.24) Depth: 19 00:00:13 6381kN
1...b7-b5 2.Bc4-b3 Bc8-b7 3.a2-a4 a7-a6 4.Kh1-g2 Rf8-d8 5.Rg3-e3 Bb7-c8 6.Re3-g3 Bc8-d7 7.a4-a5 Ra8-b8 8.Nc3-d1 Bd7-c8 9.Nd1-e3 Rb8-b7 10.Rf1-h1 Bc8-d7 11.Kg2-f2 Rb7-b8 12.c2-c3
³ (-0.32) Depth: 19 00:00:24 9755kN
1...b7-b5 2.Bc4-b3 Bc8-b7 3.a2-a4 a7-a6 4.Kh1-g2 Rf8-d8 5.Rg3-e3 Bb7-c8 6.Re3-g3 Bc8-d7 7.Nc3-d1 b5xa4 8.Bb3xa4 Ra8-b8 9.Ba4-b3 Rb8-b4 10.Nd1-c3 d6-d5 11.Nc3-a2 Rb4-b8 12.e4xd5 c6xd5 13.Rg3-e3
= (-0.24) Depth: 20 00:00:27 11841kN
1...b7-b5 2.Bc4-b3 Bc8-b7 3.a2-a4 a7-a6 4.Kh1-g2 Rf8-d8 5.Rg3-e3 Bb7-c8 6.Re3-g3 Bc8-d7 7.Nc3-d1 b5xa4 8.Bb3xa4 Ra8-b8 9.Ba4-b3 Rb8-b4 10.Nd1-c3 d6-d5 11.Nc3-a2 Rb4-b8 12.e4xd5 c6xd5 13.Rg3-e3
³ (-0.32) Depth: 20 00:00:27 12196kN
1...b7-b5 2.Bc4-b3 Bc8-b7 3.a2-a4 a7-a6 4.Kh1-g2 Rf8-d8 5.Rg3-e3 Bb7-c8 6.Re3-g3 Bc8-d7 7.Nc3-d1 Ra8-c8 8.Nd1-e3 b5xa4 9.Bb3xa4 Bd7-e6 10.Rf1-a1 Rc8-c7 11.Ne3-c4 Be6xc4 12.d3xc4
= (-0.24) Depth: 21 00:00:31 14450kN
1...b7-b5 2.Bc4-b3 Bc8-b7 3.a2-a4 a7-a6 4.Kh1-g2 Rf8-d8 5.Rf1-a1 Rd8-b8 6.Ra1-h1 Rb8-e8 7.Rh1-a1 Re8-d8 8.Rg3-e3 Rd8-b8 9.Re3-f3 b5-b4 10.Nc3-d1 Bb7-c8 11.Nd1-e3 Bc8-d7 12.Qh4-e1
= (-0.24) Depth: 21 00:00:33 16002kN
1...b7-b5 2.Bc4-b3 Bc8-b7 3.a2-a4 a7-a6 4.Kh1-g2 Rf8-d8 5.Rf1-a1 Rd8-b8 6.Rg3-e3 b5-b4 7.Nc3-d1 a6-a5 8.Re3-e1 Rb8-d8 9.Nd1-e3 d6-d5 10.Re1-h1 Ra8-b8 11.Qh4-g3
= (-0.16) Depth: 22 00:00:44 23866kN
1...b7-b6 2.a2-a4 a7-a6 3.Bc4-b3 Ra8-b8 4.Kh1-g2 Bc8-d7 5.Nc3-d1 b6-b5 6.a4-a5 Rb8-d8 7.Nd1-e3 Bd7-c8 8.Rf1-h1 Rf8-e8 9.Rh1-e1 Bc8-b7 10.c2-c4 b5xc4 11.Ne3xc4 Re8-f8 12.Rg3-f3
= (-0.20) Depth: 22 00:00:58 32394kN
1...b7-b6 2.a2-a4 a7-a6 3.Bc4-b3 Ra8-b8 4.Kh1-g2 Bc8-d7 5.Nc3-d1 b6-b5 6.a4-a5 Rb8-d8 7.Nd1-e3 Bd7-c8 8.Rf1-h1 Rf8-e8 9.Rh1-e1 Bc8-e6 10.Bb3xe6 Re8xe6 11.b2-b4 d6-d5 12.c2-c3 d5-d4 13.c3xd4
= (-0.16) Depth: 23 00:01:02 35387kN
1...b7-b6 2.a2-a4 a7-a6 3.Bc4-b3 Ra8-b8 4.Kh1-g2 Bc8-d7 5.Rf1-a1 Rf8-e8 6.Nc3-d1 b6-b5 7.Nd1-e3 Rb8-a8 8.Ra1-h1 Re8-b8 9.a4-a5 Rb8-d8 10.Rh1-a1 Rd8-f8 11.Ra1-h1 Ra8-d8 12.Rh1-e1 b5-b4 13.c2-c3
= (-0.16) Depth: 24 00:01:08 39083kN
1...b7-b6 2.a2-a4 a7-a6 3.Bc4-b3 Ra8-b8 4.Kh1-g2 Bc8-d7 5.Rf1-a1 Rf8-e8 6.Nc3-d1 b6-b5 7.Nd1-e3 Rb8-a8 8.Ra1-h1 Re8-b8 9.a4-a5 Rb8-d8 10.Rg3-f3 Ra8-b8 11.Rf3-f1 Rd8-c8 12.Rf1-e1 Rc8-c7 13.Qh4-f2 Bd7-c8
= (-0.08) Depth: 25 00:01:36 55635kN
1...b7-b6 2.a2-a4 Rf8-e8 3.b2-b4 Bc8-e6 4.Bc4xe6 Re8xe6 5.Rg3-g2 Ra8-c8 6.b4-b5 c6xb5 7.Nc3xb5 a7-a6 8.Nb5-a3 Rc8-c5 9.Na3-c4 d6-d5 10.e4xd5 Rc5xd5 11.Qh4-e4 Rd5-d4 12.Qe4-e3
= (0.00) Depth: 25 00:02:54 103201kN
1...b7-b6 2.a2-a4 a7-a6 3.Bc4-b3 Ra8-b8 4.Kh1-g2 Bc8-d7 5.Rf1-a1 Rb8-a8 6.Nc3-d1 Rf8-b8 7.Nd1-e3 b6-b5 8.Rg3-f3 Rb8-c8 9.Rf3-f1 Rc8-b8 10.Rf1-h1 Rb8-d8 11.Qh4-f2 Rd8-e8 12.Qf2-h4 Re8-d8
= (0.00) Depth: 25 00:04:41 168709kN
1...b7-b6 2.a2-a4 a7-a6 3.Bc4-b3 Ra8-b8 4.Kh1-g2 Bc8-d7 5.Rf1-a1 Rb8-a8 6.Nc3-d1 Rf8-d8 7.Nd1-e3 Bd7-c8 8.Ne3-c4 Ra8-b8 9.Nc4-e3 Bc8-b7 10.Ne3-c4 Bb7-c8
= (0.00) Depth: 26 00:05:03 181189kN
1...b7-b6 2.a2-a4 a7-a6 3.Bc4-b3 Ra8-b8 4.Kh1-g2 Bc8-d7 5.Rf1-a1 Rb8-a8 6.Nc3-d1 Rf8-d8 7.Nd1-e3 Bd7-c8 8.Rg3-f3 Ra8-b8 9.Rf3-f1 Bc8-d7 10.Rf1-h1 Rb8-a8 11.Ne3-c4
= (0.08) Depth: 27 00:06:38 236917kN
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-d1 a7-a5 4.Bb3-a4 Ra8-a6 5.Nd1-e3 Rf8-e8 6.Ne3-c4 Re8-d8 7.Kh1-g2 Bc8-e6 8.Rf1-h1 Rd8-d7 9.Rh1-f1 Rd7-d8
= (0.00) Depth: 27 00:08:03 285432kN
1...b7-b5 2.Bc4-b3 b5-b4 3.Nc3-d1 a7-a5 4.Bb3-a4 Ra8-a6 5.Nd1-e3 Bc8-e6 6.Ne3-c4 Be6xc4 7.d3xc4 Ra6-b6 8.Rg3-d3 Rf8-d8
= (0.00) Depth: 28 00:09:24 334679kN

(, 18.10.2010)


There is a clear winning plan ( I showed this method in my previous post) that all engines fail to see . I completely destroy any engine from the given position, as would any human who sees the idea. Computers still have a few weaknesses, this is clearly one of them!

This is the point I believe the original poster was trying to make, perhaps not ? White wins, and engines do not see it. Positions like this are real gems, and when engines no longer have problems with them, yikes!
Top
peter
Posts: 3187
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: Computer fail

Post by peter »

Hi!
Thanks for the great position!
Do you know the origin and if so, could you give it, please?
Thanks in advance
Peter.
Top
Uri Blass
Posts: 10360
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: Computer fail

Post by Uri Blass »

You are right about the winning plan but even without the winning plan I believe that computers can win it with white assuming that they evaluate the position as positive for white so they do not allow repetition
(and part of the chess programs have positive score for white)

The reason is that black is playing without a queen and a knight and it seems to me that white can win also by playing on the queenside and slowly improving the position(I admit that I did not try comp-comp games to test my theory).
Top
EroSennin
Posts: 133
Joined: Fri Apr 09, 2010 3:26 am

Re: Computer fail

Post by EroSennin »

I was black and this was the position I thought about going into. Instead I played the second best move which rybka gives as +2 for white.
Top
Carlos Ylich
Posts: 175
Joined: Wed Apr 28, 2010 9:31 pm
Location: Brazil

Re: Computer fail

Post by Carlos Ylich »

Any engine can understand that position?
:(
Top
User avatar
M ANSARI
Posts: 3707
Joined: Thu Mar 16, 2006 7:10 pm

Re: Computer fail

Post by M ANSARI »

A good test would be to see if you can draw this position using take backs and computer assistance against R4 with say 20 seconds a move. I am pretty sure that R4 or most of the other top engines will get a totally winning position very quickly. The evaluation of an engine is more to steer play into a certain direction rather than giving an exact final outcome. I would worry about this if one of the top engines would not be able to convert this with best defense on the other side. I tried a few lines but was unable to hold as black, and within a few moves things became decisive. I guess it will be a while before you get an exact evaluation output, and if you watch many engine games, you will sometimes know that a +.45 advantage in a middle game where the other side is passive has a much higher probability of a win than a + 1.5 advantage with 2 rooks and a pawn or with opposite bishops and minus a pawn.
Top

Return to “Computer Chess Club: General Topics”