DiscoCheck 5.2 released

[Werner]

Just noticed that DiscoCheck 5.2 has problems mating with the queen. Timecontrol was 60 moves in 60 seconds.

[D]8/4q3/8/8/5K1k/8/8/8 b - - 99 141

DiscoCheck had 78 seconds left on the clock at move 121, but the mate scores (99.xx) never got realized and the game ended in a draw. Could it be a bug with my Mac compile, or is this on Windows too?

-Max

...perhaps some of your tbs are corrupt?
Discocheck mates here without Problems.

[Max]

...perhaps some of your tbs are corrupt?
Discocheck mates here without Problems.

Does it support tablebases? There is no uci parameter for tbs settings.

-Max

[Werner]

...perhaps some of your tbs are corrupt?
Discocheck mates here without Problems.

Does it support tablebases? There is no uci parameter for tbs settings.
-Max

I saw this: +99,87/13 1,3 - but it means of course M6 - there is no tbs support. So the question is open, why DC solves it here and not with you?
I use Windows 7 64bit, the 64bit compile and Shredder GUI.

Sorry!

[Max]

So the question is open, why DC solves it here and not with you?
I use Windows 7 64bit, the 64bit compile and Shredder GUI.

Thanks Werner for testing on Windows.

In the mentioned game, Discocheck 5.2 calculated 1.3 seconds for move 92 and found Qb7+ with mate score +99,91 (mate in 4 or 5?). Maybe some gui would now have ended the game as OctoChess agrees (with a mate in 8).

Code: Select all

92. .. Qb7+ {+99,91/52 1,3} 

With the next move it played Qa7 with the same mate score. So far so good.

Code: Select all

93. Kd6 {-99,82/14 0,7} Qa7 {+99,91/25 1,3} 

And now the mate score dropped by one. It gets strange.

Code: Select all

94. Kc6 {-99,80/14 0,4} Kf2 {+99,89/29 1,3} 

Seems like Discocheck chooses only random moves from now on untill the 50-moves rule flags for a draw at move 141.

Code: Select all

95. Kd6 {-99,84/16 0,6} Qh7 {+99,91/44 0,7} 

The game was played with ponder on and xboard 4.7.2.

-Max

[Werner]

Hi,
here is my example:


[Event "40 Züge in 2 min"]
[Site "Engine Match"]
[Date "2014.01.13"]
[Round "1"]
[White "Equinox 2.01 x64 1CPU"]
[Black "DiscoCheck 5.2 x64"]
[Result "0-1"]
...
92. Ke7 {-99,78/17
0,5} Qb4+ {-M5/127 17s} 93. Ke6 {-M9/1 0s (Kf6)} Kf2
{-M4/127 0s} 94. Ke5 {-M8/1 0s (Kf6)} Qd2 {-M4/127 0s}
95. Ke6 {-M7/1 0s (Kf6)} Ke2 {-M4/127 0s} 96. Kf7 {-M7/1 0s
(Kf6)} Qd7+ {-M4/127 0s} 97. Kf6 {-M6/1 0s (Kg6)} Ke3
{-M3/127 0s} 98. Kg6 {-M5/1 0s} Kf4 {-M2/127 0s} 99. Kf6
{-M4/1 0s (Kh5)} Ke3 {-M3/127 0s} 100. Kg6 {-M5/1 0s} Kf3
{-M3/127 0s} 101. Kh6 {-M4/1 0s (Kh5)} Kf4 {-M2/127 0s}
102. Kg6 {-M3/1 0s (Kh5)} Qe7 {-M2/127 0s} 103. Kh6 {-M2/1
0s (Kh5)} Kf5 {-M2/127 0s} 104. Kh5 {-M1/1 0s} Qh7#
{-M1/127 0s} 0-1

does not look straight Forward - so if Opponent chooses other moves maybe DC Fails ?

[Frank Quisinsky]

Hi Werner,

I have the same Problem in SWCR2. Wrote about in in my News.

First Problem:
With three or four pieces on the board DiscoCheck crashed after a while. Examples: KR vs. KR, or Kp vs. K ... 4-5 moves before mate, or KQ vs. K.

Second Problem:
Engines crashed perhaps after 10 games (abnormal) after book moves before the first calculation started.

For this reason I must stopped my SWCR2 test. I am sure it have to do with ponder = on because in debug I can see illegal ponder move under Shredder GUI.

Hope the programmer can create a bugfix. Must have this fantastic Piece of work in my SWCR2.

Best
Frank

[Sylwy]

Hi Werner,

I have the same Problem in SWCR2. Wrote about in in my News.

First Problem:
With three or four pieces on the board DiscoCheck crashed after a while. Examples: KR vs. KR, or Kp vs. K ... 4-5 moves before mate, or KQ vs. K.

Second Problem:
Engines crashed perhaps after 10 games (abnormal) after book moves before the first calculation started.

For this reason I must stopped my SWCR2 test. I am sure it have to do with ponder = on because in debug I can see illegal ponder move under Shredder GUI.

Hope the programmer can create a bugfix. Must have this fantastic Piece of work in my SWCR2.

Best
Frank

Hello !

In my test under Windows XP (4'+2" TC, Shredder12.ctg book, Fritz 11 GUI, 32-bit Tony Mokonen sse3 binary) all, but absolutely all, works perfectly. DiscoCheck 5.2 version is a great chess engine. Tomorrow in the evening (here or in Chess2U forum) I'll post the result.

Regards,
SilvianR

[Max]

Hi Werner,

I have the same Problem in SWCR2. Wrote about in in my News.

First Problem:
With three or four pieces on the board DiscoCheck crashed after a while. Examples: KR vs. KR, or Kp vs. K ... 4-5 moves before mate, or KQ vs. K.

Second Problem:
Engines crashed perhaps after 10 games (abnormal) after book moves before the first calculation started.

For this reason I must stopped my SWCR2 test. I am sure it have to do with ponder = on because in debug I can see illegal ponder move under Shredder GUI.

Hope the programmer can create a bugfix. Must have this fantastic Piece of work in my SWCR2.

Best
Frank

Hi Frank,

with xboard there was no crash in my test runs. But there seems to be a bug in Discocheck 5.2 which is related to mating scores and the 127 ply depth in ponder on games.

Another example. Discocheck 5.2 got this position against Crafty 23.8 and played 58. .. Bc3 with {+99,87/17 2,1}

[D]8/8/p7/4b3/K7/4k3/8/3n4 b - - 0 58

After 59. Kb3 it should be easy to mate within some moves. At move 60 both players get 60 seconds more on the clock. But Discocheck played „random“ and we got this position at move 102:

[D]8/8/p7/8/3n4/K1b5/8/1k6 b - - 0 102

Now the pawn stepped forward and again only some kind of random play until Crafty resigned at move 126!

It's a great program and I hope too that Lucas can fix this.

-Max

Code: Select all

58. .. Bc3 {+99,87/17 2,1} 59. Kb3 {-327,51/19 1,2} Kd4
{+99,93/127 1,4} 60. Kc2 {-327,47/19 0,8} Nb2 {+99,91/127 1,7} 61. Kb1 Kd3
{+99,91/59 1,0} 62. Kc1 {-327,51/20 0,6} Nc4 {+99,93/127 0,2} 63. Kd1
{-327,53/18 1,0} Ba5 {+99,95/127 0,2} 64. Kc1 {-10,24/2 0,2} Bd2+
{+99,97/127 0,2} 65. Kb1 {-327,57/20 0,3} Bc3 {+99,93/127 0,5} 66. Ka2
{-327,51/18 0,7} Be1 {+99,93/127 0,6} 67. Kb3 {-327,49/17 0,7} Ne3
{+99,91/51 1,1} 68. Kb2 {-327,49/19 1,0} Nd5 {+99,95/127 0,2} 69. Kb3
{-327,49/19 1,2} Bd2 {+99,93/127 0,5} 70. Kb2 {-327,49/19 0,7} Bc3+
{+99,93/127 0,5} 71. Kb3 {-327,51/20 0,7} Nb6 {+99,91/40 1,1} 72. Ka3 Kc2
{+99,93/127 0,2} 73. Ka2 Nd5 {+99,95/127 0,1} 74. Ka3 {-9,89/2 0,2} Bb2+
{+99,97/127 0,2} 75. Ka4 {-18,74/19 1,1} Bc3 {+99,95/127 0,2} 76. Ka3
{+0,01/2 0,2} Bd2 {+99,95/127 0,2} 77. Ka4 {-327,55/19 0,7} Nf4
{+99,93/127 0,4} 78. Ka3 {-9,73/2 0,2} Nd3 {+99,95/127 0,1} 79. Ka4
{-327,55/19 0,8} Be1 {+99,93/127 0,4} 80. Ka3 {-9,53/2 0,2} Bb4+
{+99,95/127 0,1} 81. Ka4 {-327,55/20 0,5} Bd2 {+99,93/127 0,2} 82. Ka3 Be1
{+99,95/127 0,2} 83. Ka4 {-327,55/20 0,8} Nf4 {+99,93/127 0,3} 84. Ka3
{-9,64/2 0,2} Nd5 {+99,95/127 0,1} 85. Ka4 {-327,55/19 0,9} Bd2
{+99,89/20 1,5} 86. Ka3 {-9,52/2 0,2} Bb4+ {+99,91/70 1,3} 87. Ka4 Be1
{+99,89/20 1,5} 88. Ka3 Nb4 {+99,91/20 1,2} 89. Ka4 {-8,85/2 0,2} Nc6
{+99,93/127 0,2} 90. Ka3 {-9,55/2 0,2} Bd2 {+99,95/127 0,2} 91. Ka4
{-327,51/19 1,0} Be1 {+99,93/127 0,2} 92. Ka3 {+0,01/2 0,2} Bb4+
{+99,91/98 1,4} 93. Ka4 {-327,51/20 1,0} Bc3 {+99,89/32 1,6} 94. Ka3 Na5
{+99,91/100 1,4} 95. Ka4 {-327,55/22 1,0} Bd2 {+99,89/71 1,6} 96. Ka3
{-9,50/2 0,2} Nc4+ {+99,91/127 0,4} 97. Ka4 {-327,55/20 0,5} Nd6
{+99,91/110 1,7} 98. Ka3 {-9,64/2 0,2} Nb5+ {+99,93/127 0,1} 99. Ka4
{-327,51/19 1,2} Nd4 {+99,89/24 1,7} 100. Ka3 {-327,53/21 0,2} Bc3
{+99,91/21 1,5} 101. Ka4 Kb1 {-97,20/29 1,7} 102. Ka3 {-9,31/2 0,2} a5
{+99,85/16 1,7} 103. Ka4 Kc2 {+99,87/19 1,6} 104. Ka3 Bb4+ {+99,89/18 1,6}
105. Ka4 {-327,47/20 1,2} Bd2 {+99,91/65 0,5} 106. Ka3 Nc6 {+99,93/127 0,2}
107. Ka4 {-327,51/19 1,2} Nd4 {+99,91/90 2,0} 108. Ka3 {+0,01/2 0,2} Ne2
{+99,93/127 0,1} 109. Ka4 {-327,41/19 2,3} Kd3 {+99,83/16 3} 110. Kb5
{-327,45/18 2,5} Nd4+ {+99,85/16 0,2} 111. Ka4 {-327,47/20 1,2} Be1
{+99,89/22 2,2} 112. Ka3 {-10,03/2 0,2} Ne2 {+99,91/72 2,0} 113. Kb3
{-327,47/20 2,4} Bc3 {+99,91/80 2,2} 114. Ka3 Kd2 {+99,93/127 0,2} 115. Kb3
{-327,45/17 1,6} Kd3 {+99,87/17 2,5} 116. Ka3 {+0,01/125 0,2} Bb4+
{+99,93/127 0,4} 117. Ka4 {-327,47/19 1,5} Nc3+ {+99,89/29 3} 118. Kb3 Kd2
{+99,91/123 2,9} 119. Kc4 {-327,51/19 1,2} Ke3 {+99,93/127 0,2} 120. Kb3
{-9,03/2 0,2} Kd3 {+99,95/127 0,2} 121. Kb2 {-327,49/2 0,2} a4
{+99,97/127 0,2} 122. Ka1 {-327,55/18 0,5} Bc5 {+99,95/127 1,0} 123. Kb2
{-11,10/2 0,2} Bd6 {+99,97/127 0,2} 124. Ka1 {-327,55/16 0,9} Bb4
{+99,95/127 0,2} 125. Kb2 {-10,93/2 0,2} Bc5 {+99,97/127 0,2}
{White resigns} 0-1

[Mike S.]

Ponder = on fails also under the Fritz GUI. Sometimes when a new game starts, Disco does not move at all and loses on time. So I conclude it is really a problem of Discocheck, not of the GUIs, in case of ponder on.

Otherwise, as the results tell, a very promising engine!

[carldaman]

Ponder = on fails also under the Fritz GUI. Sometimes when a new game starts, Disco does not move at all and loses on time. So I conclude it is really a problem of Discocheck, not of the GUIs, in case of ponder on.

Otherwise, as the results tell, a very promising engine!

I believe DC did not support pondering until recently, so its implementation may be buggy.