TalkChess.com

I saw this played by erdos(C) on ICC, supposedly Houdini 1.5a with 3-5 Gaviota EGTBs on board.

[D]8/7p/6p1/6P1/5K1k/8/8/8 b - -

Instead if winning with Kh5 it played h5?? and got mated not too much later.
What could be the cause and are other programs falling for it also?

Even a 1200 Elo player will be ashamed to make such a stupid blunder.....
Dr.D

marcelk wrote:Instead if winning with Kh5 it played h5??

Indeed, seems to be a problem with the Gaviota TBs (en passant bug?)

Besides Houdini gtb, Critter gtb also plays h5??

(of course, in Fritz GUI you must forbid Nalimov access)

Dr.Wael Deeb wrote:Even a 1200 Elo player will be ashamed to make such a stupid blunder.....
Dr.D

A very clever and adequate post...
Too bad, lately you seemed to have recovered... Get well, Doctor!

This is a known bug that was fixed in the 0.4 version of the Gaviota TB probing code.

http://sites.google.com/site/gaviotache ... tledpost-1

tmokonen wrote:This is a known bug that was fixed in the 0.4 version of the Gaviota TB probing code.

Hi Tony,

Yes, I had seen that, but what is meant by "the probing code"?
I thought that the TBs themselves did not need a modification (is there a new version of the TBs, a new download?)
But maybe it is Houdini and Critter that need to be modified, concerning their TB access?
So what is the final correction ?

Hi Tony,

Just saw your other post, to Richard, about the gtb32.dll and gtb64.dll files.

Could you also tell Robert for Houdini (but probably he already knows...)

ernest wrote:

tmokonen wrote:This is a known bug that was fixed in the 0.4 version of the Gaviota TB probing code.

Hi Tony,

Yes, I had seen that, but what is meant by "the probing code"?

Hi Ernest, I meant the standard code that is used by all programs that make use of the Gaviota tablebases, the code written by the author of Gaviota Miguel Ballicora.

I noticed that Critter 1.0 used the exact same gtb32.dll and gtb64.dll files as Critter 0.90. The latest version of the tablebase code, version 0.4, came out after Critter 0.90, which means Critter 1.0 is still using the old tablebase code with the en passant bug. I was hoping that Richard Vida could remake these two dll files easily, without trying to rush him into releasing a new version of Critter.

tmokonen wrote:This is a known bug that was fixed in the 0.4 version of the Gaviota TB probing code.

http://sites.google.com/site/gaviotache ... tledpost-1

In addition, it means that Houdini in this game was used in "probe only at root" mode. If you probe in search, it is almost impossible that this bug will have a significant effect. I suggest that all engines that have a GTB probing code previous to 0.4 to be used with "probing in search" or whatever name the author gave for that feature.

Sorry for the screw up, it went undetected for almost a year.

Miguel
PS: Why this post is here in this subforum?

Dr.Wael Deeb wrote:Even a 1200 Elo player will be ashamed to make such a stupid blunder.....
Dr.D

[D]8/7p/6p1/6P1/5K1k/8/8/8 b - - 0 1

Analysis by Houdini 1.5a w32:

1...Kh5 2.Kf3 Kxg5 3.Kg3 h5 4.Kh3 Kf4
-+ (-2.99) Depth: 6/9 00:00:00 1kN
1...Kh5 2.Kf3 Kxg5 3.Kg2 h5 4.Kg3 h4+ 5.Kh3 Kf5
-+ (-3.06) Depth: 7/11 00:00:00 1kN
1...Kh5 2.Kf3 Kxg5 3.Kf2 h5 4.Kg3 h4+ 5.Kh3 Kf4
-+ (-3.12) Depth: 8/11 00:00:00 1kN
1...Kh5
-+ (-3.28 ++) Depth: 9/11 00:00:00 2kN
1...Kh5 2.Kf3 Kxg5 3.Kf2 h5 4.Kg3 h4+ 5.Kf3 Kf5 6.Kg2
-+ (-3.33) Depth: 9/12 00:00:00 2kN
1...Kh5
-+ (-3.48 ++) Depth: 10/13 00:00:00 3kN
1...Kh5 2.Kf3 Kxg5 3.Kf2 h5 4.Kg3 h4+ 5.Kf3 Kf5 6.Kf2 g5 7.Kg2
-+ (-3.69) Depth: 10/16 00:00:00 3kN
1...Kh5
-+ (-3.86 ++) Depth: 11/16 00:00:00 5kN
1...Kh5 2.Kf3 Kxg5 3.Kf2 Kg4 4.Ke3 h5 5.Ke2 h4 6.Kf2 h3 7.Kg1 Kg3
-+ (-3.99) Depth: 11/16 00:00:00 6kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 h5 4.Kg3 h4+ 5.Kf3 Kf5 6.Kg2 Kg4 7.Kf2 h3 8.Kg1 Kg3
-+ (-3.99) Depth: 12/18 00:00:00 7kN
1...Kh5
-+ (-4.15 ++) Depth: 13/18 00:00:00 8kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 h5 4.Kg3 h4+ 5.Kf3 Kf5 6.Kg2 Kg4 7.Kh1 Kg3 8.Kg1 g5
-+ (-4.13) Depth: 13/18 00:00:00 8kN
1...Kh5
-+ (-4.30 ++) Depth: 14/19 00:00:00 10kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 h5 4.Kg3 h4+ 5.Kf3 Kf5 6.Kg2 Kf4 7.Kh3 g5 8.Kg2 g4 9.Kh2 h3
-+ (-4.44) Depth: 14/19 00:00:00 11kN
1...Kh5
-+ (-4.62 ++) Depth: 15/21 00:00:00 14kN
1...Kh5
-+ (-5.00 ++) Depth: 15/21 00:00:00 15kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 h5 4.Kg3 h4+ 5.Kh3 Kf5 6.Kh2 g5 7.Kh3 g4+ 8.Kg2 Kf4 9.Kh2 h3 10.Kg1 Kg3 11.Kh1
-+ (-5.11) Depth: 15/22 00:00:00 18kN
1...Kh5
-+ (-5.32 ++) Depth: 16/22 00:00:00 21kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 h5 4.Kg3 h4+ 5.Kf3 h3 6.Kg3 h2 7.Kxh2 Kg4 8.Kg2 Kf4 9.Kf2 g5 10.Kg2 Kg4 11.Kf2 Kh3 12.Kf3 g4+ 13.Kf4 g3
-+ (-5.62) Depth: 16/23 00:00:00 25kN
1...Kh5
-+ (-5.85 ++) Depth: 17/26 00:00:00 29kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 h5 4.Kg3 h4+ 5.Kf3 h3 6.Kg3 h2 7.Kxh2 Kg4 8.Kg2 g5 9.Kf2 Kh3 10.Kf3 g4+ 11.Kf2 g3+ 12.Kf1 Kh2
-+ (-5.92) Depth: 17/27 00:00:00 33kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 h5 4.Kg3 h4+ 5.Kf3 h3 6.Kg3 h2 7.Kxh2 Kg4 8.Kg2 g5 9.Kf2 Kh3 10.Kf3 g4+ 11.Kf2 g3+ 12.Kf1 Kh2 13.Ke2 g2
-+ (-5.83) Depth: 18/28 00:00:00 44kN
1...Kh5
-+ (-6.08 ++) Depth: 19/28 00:00:00 53kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 h5 4.Kg3 h4+ 5.Kh3 Kf4 6.Kxh4 g5+ 7.Kh3 Kf3 8.Kh2 g4 9.Kg1 Kg3 10.Kf1 Kh2 11.Kf2 Kh3 12.Kg1 Kg3 13.Kf1
-+ (-6.02) Depth: 19/28 00:00:00 64kN
1...Kh5
-+ (-6.27 ++) Depth: 20/28 00:00:00 72kN
1...Kh5
-+ (-6.77 ++) Depth: 20/28 00:00:00 77kN
1...Kh5
-+ (-8.16 ++) Depth: 20/28 00:00:00 84kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 h5 4.Kg3 h4+ 5.Kh3 Kf4 6.Kxh4 g5+ 7.Kh3 Kf3 8.Kh2 g4 9.Kg1 Kg3 10.Kf1 Kh2 11.Kf2 g3+ 12.Kf3 g2 13.Kf4 g1Q
-+ (-8.95) Depth: 20/28 00:00:00 92kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 h5 4.Kg3 h4+ 5.Kh3 Kf4 6.Kxh4 g5+ 7.Kh3 Kf3 8.Kh2 g4 9.Kg1 Kg3 10.Kf1 Kh2 11.Kf2 g3+ 12.Kf3 g2 13.Kf4 g1Q 14.Ke4 Qg6+ 15.Ke5 Qg5+ 16.Ke4
-+ (-9.13) Depth: 21/31 00:00:00 111kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 h5 4.Kg3 h4+ 5.Kh3 Kf4 6.Kxh4 g5+ 7.Kh3 Kf3 8.Kh2 g4 9.Kg1 Kg3 10.Kf1 Kh2 11.Kf2 g3+ 12.Kf3 g2 13.Kf4 g1Q 14.Ke4 Qg6+ 15.Ke5 Qg5+ 16.Ke4
-+ (-9.13) Depth: 22/31 00:00:00 131kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 h5 4.Kg3 h4+ 5.Kh3 Kf4 6.Kxh4 g5+ 7.Kh3 Kf3 8.Kh2 g4 9.Kg1 Kg3 10.Kf1 Kh2 11.Kf2 g3+ 12.Kf3 g2 13.Ke4 g1Q 14.Kd5 Qg5+ 15.Kd4 Kg3
-+ (-9.24) Depth: 23/33 00:00:00 168kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 h5 4.Kg3 h4+ 5.Kh3 Kf4 6.Kxh4 g5+ 7.Kh3 Kf3 8.Kh2 g4 9.Kg1 Kg3 10.Kf1 Kh2 11.Kf2 g3+ 12.Kf3 g2 13.Ke4 g1Q 14.Kd3 Qc5 15.Ke4 Kg3
-+ (-9.35) Depth: 24/35 00:00:00 226kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 h5 4.Kg3 h4+ 5.Kh3 Kf4 6.Kxh4 g5+ 7.Kh3 Kf3 8.Kh2 g4 9.Kg1 Kg3 10.Kf1 Kh2 11.Kf2 g3+ 12.Kf3 g2 13.Ke4 g1Q 14.Kd5 Qg5+ 15.Kd4 Kg3 16.Kc4 Qf4+ 17.Kd5 Qf5+ 18.Kd4
-+ (-9.44) Depth: 25/35 00:00:00 291kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 h5 4.Kg3 h4+ 5.Kh3 Kf4 6.Kxh4 g5+ 7.Kh3 Kf3 8.Kh2 g4 9.Kg1 Kg3 10.Kf1 Kh2 11.Kf2 g3+ 12.Kf3 g2 13.Ke4 g1Q 14.Kd5 Qg5+ 15.Ke4 Kg3 16.Kd4 Qf5 17.Kc4 Qe4+ 18.Kc5
-+ (-9.51) Depth: 26/37 00:00:00 381kN
1...Kh5
-+ (-9.90 ++) Depth: 27/43 00:00:00 1150kN
1...Kh5
-+ (-10.70 ++) Depth: 27/43 00:00:00 1387kN
1...Kh5
-+ (-12.97 ++) Depth: 27/43 00:00:00 2270kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 Kh4 4.Kg2 g5 5.Kh2 g4 6.Kg2 g3 7.Kg1 Kh3 8.Kh1 h5 9.Kg1 g2 10.Kf2 Kh2 11.Ke3 g1Q+ 12.Kf3 Qg4+ 13.Kf2 h4 14.Ke3 Qg1+ 15.Kf3 h3 16.Ke2 Qg4+ 17.Kd2 Kg2 18.Ke3 Qf3+ 19.Kd4 h2 20.Ke5 Qe3+ 21.Kf5 h1Q 22.Kg6 Qb1+ 23.Kf6
-+ (-16.86) Depth: 27/49 00:00:00 2934kN
1...Kh5
-+ (-17.56 ++) Depth: 28/49 00:00:01 3609kN
1...Kh5
-+ (-18.97 ++) Depth: 28/49 00:00:01 4358kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 Kh4 4.Kg2 g5 5.Kh2 g4 6.Kg2 g3 7.Kf3 Kh3 8.Ke3 Kh2 9.Kf3 g2 10.Kf4 g1Q 11.Kf3 h5 12.Ke4 h4 13.Ke5 h3 14.Kd5 Kg2 15.Kc6 h2 16.Kc7 h1Q 17.Kb7 Kf2+ 18.Ka6 Qg6+ 19.Ka5 Qg5+ 20.Kb6 Qe3+ 21.Ka5 Qa3+ 22.Kb6 Qb4+ 23.Kc7 Qc5+ 24.Kd7 Qhd5+ 25.Ke8 Qg8+ 26.Kd7 Qgc8#
-+ (-#25) Depth: 28/57 00:00:11 62923kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 Kh4 4.Kg2 g5 5.Kh2 g4 6.Kg2 g3 7.Kf3 Kh3 8.Ke3 Kh2 9.Kf3 g2 10.Kf4 g1Q 11.Ke4 h5 12.Kd5 h4 13.Ke5 h3 14.Kf6 Kg2 15.Kg7 h2 16.Kf6 h1Q 17.Kf7 Qf2+ 18.Kg8 Qd4 19.Kf8 Qf6+ 20.Ke8 Qh5+ 21.Kd7 Qb5+ 22.Kc8 Qfc6+ 23.Kd8 Qg5#
-+ (-#23) Depth: 29/57 00:00:12 66596kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 Kh4 4.Kg2 g5 5.Kh2 g4 6.Kg2 g3 7.Kf3 Kh3 8.Ke3 Kh2 9.Kf3 g2 10.Kf4 g1Q 11.Ke4 h5 12.Kd3 h4 13.Ke4 h3 14.Kf4 Kg2 15.Ke5 h2 16.Kd6 h1Q 17.Kd5 Kg3+ 18.Kd6 Kg2 19.Kd5
-+ (-#22) Depth: 30/57 00:00:13 69691kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 Kh4 4.Kg2 g5 5.Kh2 g4 6.Kg2 g3 7.Kf3 Kh3 8.Ke3 Kh2 9.Kf3 g2 10.Kf4 g1Q 11.Ke4 h5 12.Kd3 h4 13.Ke4 h3 14.Kf4 Kg2 15.Ke5 h2 16.Kd5 h1Q 17.Kc6 Kf1+ 18.Kd6 Qd4+ 19.Ke6 Qhd5+ 20.Ke7 Qg7+ 21.Ke8 Qgf7#
-+ (-#21) Depth: 31/57 00:00:13 71470kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 Kh4 4.Kg2 g5 5.Kh2 g4 6.Kg2 g3 7.Kf3 h5 8.Ke2 Kg4 9.Ke1 h4 10.Kf1 h3 11.Kg1 g2 12.Kf2 Kf4 13.Ke2 h2 14.Kd3 h1Q 15.Kc3 g1Q 16.Kb3 Qc5 17.Kb2 Qb7+ 18.Ka2 Qca7#
-+ (-#18) Depth: 32/57 00:00:15 80571kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 Kh4 4.Kg2 g5 5.Kh2 g4 6.Kg2 g3 7.Kf3 h5 8.Ke2 Kg4 9.Ke1 h4 10.Kf1 h3 11.Kg1 g2 12.Kf2 Kf4 13.Ke2 h2 14.Kd3 h1Q 15.Kc3 g1Q 16.Kb3 Qd5+ 17.Kb2 Qd3 18.Ka2 Qgb1#
-+ (-#18) Depth: 33/57 00:00:15 81746kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 Kh4 4.Kg2 g5 5.Kh2 g4 6.Kg2 g3 7.Kf3 h5 8.Ke2 Kg4 9.Kf1 h4 10.Kg1 h3 11.Kf1 g2+ 12.Kf2 Kf4 13.Ke2 h2 14.Kd3 h1Q 15.Kc3 g1Q 16.Kb3 Qd5+ 17.Kb2 Qd3 18.Ka2 Qgb1#
-+ (-#18) Depth: 34/57 00:00:15 83761kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 Kh4 4.Kg2 g5 5.Kh2 g4 6.Kg2 g3 7.Kf3 h5 8.Ke2 Kg4 9.Kf1 h4 10.Kg1 h3 11.Kf1 g2+ 12.Kf2 Kf4 13.Ke2 h2 14.Kd3 h1Q 15.Kc3 g1Q 16.Kb3 Qd5+ 17.Kb2 Qd3 18.Ka2 Qgb1#
-+ (-#18) Depth: 35/57 00:00:16 87646kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 Kh4 4.Kg2 g5 5.Kh2 g4 6.Kg2 g3 7.Kf3 h5 8.Ke2 Kg4 9.Kf1 h4 10.Kg1 h3 11.Kf1 g2+ 12.Kf2 Kf4 13.Ke2 h2 14.Kd3 h1Q 15.Kc3 g1Q 16.Kb3 Qe3+ 17.Ka4 Qd5 18.Kb4 Qeb3#
-+ (-#18) Depth: 36/57 00:00:17 93965kN
1...Kh5 2.Kg3 Kxg5 3.Kf3 Kh4 4.Kg2 g5 5.Kh2 g4 6.Kg2 g3 7.Kf3 h5 8.Ke2 Kg4 9.Kf1 h4 10.Kg2 h3+ 11.Kf1 g2+ 12.Kf2 Kf4 13.Ke2 h2 14.Kd3 h1Q 15.Kc3 g1Q 16.Kb3 Qd5+ 17.Kb2 Qd3 18.Ka2 Qgb1#
-+ (-#18) Depth: 37/57 00:00:19 105mN
1...Kh5 2.Kg3 Kxg5 3.Kf3 Kh4 4.Kg2 g5 5.Kh2 g4 6.Kg2 g3 7.Kf3 h5 8.Ke2 Kg4 9.Kf1 h4 10.Kg2 h3+ 11.Kf1 g2+ 12.Kf2 Kf4 13.Ke2 h2 14.Kd3 h1Q 15.Kc3 g1Q 16.Kb3 Qd5+ 17.Kb2 Qd3 18.Ka2 Qgb1#
-+ (-#18) Depth: 38/57 00:00:23 124mN

(25.03.2011)