Article View: gnu.chess
Article #5727depth anomalies
From: Paul Vaughan
Date: Sun, 14 Jun 1992 20:37
Date: Sun, 14 Jun 1992 20:37
155 lines
8196 bytes
8196 bytes
This is about gnuchess 4p45, downloaded from prep with no patches
applied.
I posted this game earlier noting that gnuchess let itself be mated
unnecessarily. I've been trying to reproduce that, but have not
succeeded. But I was doing some work with gnuan and I noticed a few
depth anomalies that I don't understand. Even though I've given it 3
minutes to think to a depth of only 3 plies, it occasionally only
prints out 2 plies in the best line column. It certainly did not
exhaust its 3 minutes on any move and I don't think either side has a
forced mate or draw at these points. This leads me to believe that
perhaps the search algorithm is prematurely shutting down. Because it
still prints 3 in the depth column, this suggests that it thinks it
has completed the requested search. Is there a better explanation?
Another curious thing is the last move. On this move, it did spend
more time than on most, but still less than three minutes. I don't
understand why it would take more than a few seconds to find this move
for white. In the listing from the actual game, I get:
f5f6 0 0 0 2500 b6e3 220 3 25026 1100
h5h8 -9999 1 42 0
but from gnuan, I get
b6e3* 264 3 g7f6 g6g7 f8g7 h5f7 g7h8 f7h5 h8g7 f1d1
28 h5h8 9998 1 h5h8 c8c4
Note the difference in score (1 point) and that gnuan gives a move for
black in the best line, even though black is in fact mated at this
point. Here is the position:
r . r . . k . Q
p p . . . p p .
. . . . . P P .
. . . . p . . .
. . B . . . . .
. P . . q . . .
. P P . . . . K
. . . n . R . .
black to play
[8.85]bock) gnuan
Book used 4611 of 7500 entries.
Playing without hashfile
Input the file with the algebraic moves of the game.
CLp45.Jun11-145947B
Input the search depth you want to use. (1 to 29)
3
Input the maximum number of minutes per move you want to use.
3
3 3
Move White Black Score Depth Best Line
------------------------------------------------------------------
1 e2e4* 29 3 g1f3 g8f6 b1c3
c7c5* 4 3 g8f6 b1c3 b8c6
2 g1f3* 42 3 b1c3 g8f6 g1f3
b8c6* -20 3 g8f6 e4e5 f6e4
3 d2d4* 39 3 b1c3 g8f6 f1c4
c5d4* -17 3 d7d5 d4c5 d8a5 c1d2 a5c5
4 f3d4 32 3 f3d4 g8f6 b1c3
g8f6* -11 3 c6d4 d1d4 g8f6
5 b1c3 33 3 b1c3 e7e5 d4c6 d7c6 d1d8 e8d8
d7d6* -27 3 d8b6 d4c6 d7c6
6 f1e2* 47 3 d4c6 b7c6 c1e3 e7e5
e7e5* -9 3 c6d4 d1d4 e7e5
7 d4b3* 31 3 d4c6 b7c6 c1e3 a8b8
f8e7* -8 3 c8e6 e2b5 e6b3 a2b3
8 e1g1 -5 3 e1g1 c8e6 e2b5
e8g8 4 3 e8g8 c1e3 c8e6
9 c1e3 -17 3 c1e3 c8e6 b3d2
c8e6 5 3 c8e6 e2b5 e6g4
10 f2f4* -17 3 f2f3 a8c8 e2b5
d8d7 12 3 d8d7 f4f5 e6b3 a2b3
11 f4f5 -8 3 f4f5 e6b3 a2b3
e6b3 1 3 e6b3 a2b3 f8c8 e2c4
12 a2b3 3 3 a2b3 f8c8
f8c8 -18 3 f8c8 e2c4 h7h6
13 e2d3* 8 3 e2c4 h7h6 f1f3
d6d5 14 3 d6d5 c3d5 f6d5 e4d5 d7d5
14 e4d5* -10 3 c3d5 f6d5 e4d5 d7d5
c6b4* -88 3 c6d4 a1a4 f6d5 e3d4 e5d4 a4d4
15 d3c4* 84 3 d3b5 d7d6 b5c4
e7c5 -61 3 e7c5 e3c5 c8c5
16 e3c5 106 3 e3c5 c8c5 d5d6 e5e4
c8c5 -102 3 c8c5 d5d6
17 d1e2* 90 3 d5d6 e5e4 a1a4
d7d6 -62 3 d7d6 c3b5 d6e7 b5a7 f6d5
18 g2g4* 85 3 f1e1 c5c8 e2e5 d6e5 e1e5 b4c2
h7h6* 34 3 b4d5 c3d5 f6d5
19 h2h4* 44 3 f1e1 c5c8 c3e4
c5c8* 44 3 b4d5 c3d5 f6d5
20 g4g5* 36 3 f1d1 a7a6 c3e4 f6e4 e2e4
h6g5* 31 3 f6d5 f1d1 d5c3 b2c3
21 h4g5 -20 3 h4g5 f6d5
f6d5 16 3 f6d5 f1d1 d5c3 b2c3
22 c3d5 -7 3 c3d5 b4d5 a1d1 d6b6 g1h2
b4d5 11 3 b4d5 a1d1
23 a1d1* -12 3 e2f2 a7a6 f1d1
d6b6 84 3 d6b6 f1f2
24 g1h2* -95 3 f1f2 d5e3 d1b1
d5e3 181 3 d5e3 f1e1 e3d1 e1d1
25 g5g6* -160 3 d1d7 e3f1 e2f1 c8f8
g8f8* 122 3 c8c4 g6f7 g8f7 b3c4 e3f1 d1f1 b6b2
26 e2h5* -145 3 g6f7 b6h6 h2g1 h6g5 g1f2 e3f1 d1f1 g5f5 f2e1
e3d1 218 3 e3d1 h5d1 c8d8 d1e2
27 f5f6* -214 3 h5d1 c8d8
b6e3* 264 3 g7f6 g6g7 f8g7 h5f7 g7h8 f7h5 h8g7 f1d1
28 h5h8 9998 1 h5h8 c8c4
White's percentage was 35.71%. Black's percentage was 51.85%.
Elapsed time was 218 seconds.
[8.86]bock)
gnuchess 4p45 game
score depth nodes time score depth nodes time
e2e4 0 0 0 0 c7c5 -19 0 0 0
g1f3 0 0 0 200 b8c6 -44 0 0 0
d2d4 0 0 0 200 c5d4 -56 0 0 0
f3d4 0 0 0 100 g8f6 -34 0 0 0
b1c3 0 0 0 200 d7d6 -30 0 0 0
f1e2 0 0 0 200 e7e5 -19 0 0 0
d4b3 0 0 0 100 f8e7 12 0 0 0
e1g1 0 0 0 200 e8g8 -3 0 0 0
c1e3 0 0 0 400 c8e6 -16 4 12015 1000
f2f4 0 0 0 1200 d8d7 12 4 6008 500
f4f5 0 0 0 2400 e6b3 -14 5 6009 500
a2b3 0 0 0 1400 f8c8 -1 5 6008 500
e2d3 0 0 0 2400 d6d5 1 4 9010 800
e4d5 0 0 0 1000 c6b4 -390 4 22024 1300
d3c4 0 0 0 2400 e7c5 -75 4 11012 800
e3c5 0 0 0 1100 c8c5 -96 5 23024 1500
d1e2 0 0 0 2200 d7d6 -58 4 19025 1400
g2g4 0 0 0 500 h7h6 31 4 17020 1200
h2h4 0 0 0 500 c5c8 41 4 17018 1200
g4g5 0 0 0 600 h6g5 24 4 18019 900
h4g5 0 0 0 300 f6d5 9 5 18019 1100
c3d5 0 0 0 1400 b4d5 17 5 21022 1300
a1d1 0 0 0 800 d6b6 90 5 28029 1300
g1h2 0 0 0 2500 d5e3 164 5 26027 1300
g5g6 0 0 0 2300 g8f8 149 5 28029 1200
e2h5 0 0 0 3200 e3d1 171 3 16017 800
f5f6 0 0 0 2500 b6e3 220 3 25026 1100
h5h8 -9999 1 42 0
Black
Technical Details
Message-ID:
<9206142037.AA00807@bock.mcc.com>Path:
arc1-ca.pugleaf.net!sparky!uunet!cis.ohio-state.edu!bock.mcc.com!vaughanShow Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 10d71d,34c6961be996aded,start X-Google-Attributes: gid10d71d,public X-Google-ArrivalTime: 1992-06-14 16:52:18 PST Newsgroups: gnu.chess Path: arc1-ca.pugleaf.net!sparky!uunet!cis.ohio-state.edu!bock.mcc.com!vaughan From: vaughan@bock.mcc.com (Paul Vaughan) Subject: depth anomalies Message-ID: <9206142037.AA00807@bock.mcc.com> Sender: gnulists@ai.mit.edu Organization: GNUs Not Usenet Distribution: gnu Date: Sun, 14 Jun 1992 20:37:23 GMT Lines: 156