Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Discrepancy between score and PV

Author: Dann Corbit

Date: 23:18:04 07/21/03

Go up one level in this thread


On July 21, 2003 at 23:53:48, Normand M. Blais wrote:

>The following position occured in a game between two chess programs (My program
>had white).
>
>[d]r3r1k1/1p3ppp/2n1qn2/p1b1pNB1/3pP3/P2B4/1PP2PPP/R2Q1RK1 w - -
>
>After the game, I decided to let the program analyse the position again just for
>fun. But then I notice that, after finding the correct move, the pv line at
>depth 6 was ending in a checkmate while the score at that depth was 0.02. I
>suspect that this is a bug. Any idea how this could happend?
>
>NMB
>
>Here is the analysis:
>
>  1   -0.09      0        58  a4
>  2   -0.19      0      1404  a4 Rac8
>  3    0.05      0      2401  Qf3 Nd7 Rfd1
>  4   -0.05      2     12896  Qf3 Nd7 Rfd1 Rac8
>  5   -0.04      7     50855  Qf3 Nd7 Rfd1 Rac8 a4
>  6    0.02     23    147187  Bc4 Qxc4 Bxf6 Bf8 Qg4 Qxc2 Nxg7 Bxg7 Qxg7#
>  7    0.15     81    515963  Bc4 Qxc4 Bxf6 Bf8 Qg4 h5 Nh6+ Kh8
>  8    0.02    296   1695515  Bc4 Qxc4 Bxf6 Bf8 Nxg7 Bxg7 Qg4 Qxc2 Qxg7#

There seems to be some difference of opinion as to what the best thing to do
here is:

Crafty:
   clearing hash tables
   time surplus   0.00  time limit 16:39 (16:39)
   depth   time  score   variation (1)
     6->   0.44   0.67   1. Qf3 Nd7 2. b3 Qg6 3. h4 f6
     7     0.71   0.64   1. Qf3 Nd7 2. Qg3 Qg6 3. Nh4 Qh5 4.
                         Nf5
     7->   1.64   0.64   1. Qf3 Nd7 2. Qg3 Qg6 3. Nh4 Qh5 4.
                         Nf5
     8     3.24   0.53   1. Qf3 Nd7 2. b3 Qg6 3. Bd2 Qf6 4.
                         Bc4 a4 5. bxa4 Rxa4
     8->   6.52   0.53   1. Qf3 Nd7 2. b3 Qg6 3. Bd2 Qf6 4.
                         Bc4 a4 5. bxa4 Rxa4
     9     8.93   0.53   1. Qf3 Nd7 2. b3 Qg6 3. Bd2 Qf6 4.
                         Bc4 a4 5. bxa4 Rxa4
     9    23.07   0.58   1. f4 exf4 2. Rxf4 Nd7 3. Qf1 f6 4.
                         Bh4 Nde5 5. Be2
     9    24.39   0.90   1. Bc4 Qxc4 2. Bxf6 g6 3. Nh6+ Kf8
                         4. b3 Qc3 5. h3 a4
     9->  24.73   0.90   1. Bc4 Qxc4 2. Bxf6 g6 3. Nh6+ Kf8
                         4. b3 Qc3 5. h3 a4
    10    26.01   0.53   1. Bc4 Qxc4 2. Bxf6 g6 3. Nh6+ Kf8
                         4. Rc1 Be7 5. Ng4 Qc5 6. Qf3 Bxf6 7.
                         Nxf6
    10    50.74   0.57   1. f4 Nd7 2. Qg4 Kh8 3. Qg3 g6 4. Nh4
                         f6 5. Bh6 Rac8 6. fxe5 Ndxe5
    10->   1:10   0.57   1. f4 Nd7 2. Qg4 Kh8 3. Qg3 g6 4. Nh4
                         f6 5. Bh6 Rac8 6. fxe5 Ndxe5
    11     2:00   0.52   1. f4 exf4 2. Rxf4 Nd7 3. Qg4 Nde5
                         4. Qg3 Nxd3 5. cxd3 Rac8 6. Nh6+ gxh6
                         7. Be7+ Qg6 8. Bxc5 Qxg3 9. hxg3
    11     2:43   0.56   1. Qf3 Nd7 2. b3 h6 3. Bc4 Qg6 4. Bd2
                         Ne7 5. Nxe7+ <HT>
    11     3:06   0.59   1. Bc4 Qxc4 2. Bxf6 g6 3. Nh6+ Kf8
                         4. Ng4 Kg8 5. Qf3 Be7 6. Nxe5 Qxc2
    11->   3:19   0.59   1. Bc4 Qxc4 2. Bxf6 g6 3. Nh6+ Kf8
                         4. Ng4 Kg8 5. Qf3 Be7 6. Nxe5 Qxc2
    12     3:28   0.58   1. Bc4 Qxc4 2. Bxf6 g6 3. Nh6+ Kf8
                         4. Qf3 Be7 5. Bxe7+ Kxe7 6. b3 Qe6
                         7. Rad1 Kf8 8. h3
    12->   5:16   0.58   1. Bc4 Qxc4 2. Bxf6 g6 3. Nh6+ Kf8
                         4. Qf3 Be7 5. Bxe7+ Kxe7 6. b3 Qe6
                         7. Rad1 Kf8 8. h3
    13     5:46   0.52   1. Bc4 Qxc4 2. Bxf6 g6 3. Nh6+ Kf8
                         4. Qf3 Be7 5. Bxe7+ Kxe7 6. Rad1 a4
                         7. b3 axb3 8. cxb3 Qe6 9. b4
   time=16:39  cpu=98%  mat=0  n=320857243  fh=92%  nps=321k
   ext-> chk=6264616 cap=1032852 pp=132168 1rep=380104 mate=26868
   predicted=0  nodes=320857243  evals=88678232
   endgame tablebase-> probes done=0  successful=0

Ruffian:
  r3r1k1/1p3ppp/2n1qn2/p1b1pNB1/3pP3/P2B4/1PP2PPP/R2Q1RK1 w - -
  Depth   Time  Score    Best line
   8/18   1.41   0.18    Qe2 Ne7 Bc4 Qb6 Qf3 Nxf5 Qxf5 Qc6 Rfe1
   8-->   1.78   0.18    Qe2 Ne7 Bc4 Qb6 Qf3 Nxf5 Qxf5 Qc6 Rfe1
   9/23   3.41   0.14    Qe2 Ne7 Bc4 Qb6 Ng3 Rac8 Rab1 Qc7 Bxf6 gxf6
   9/23   6.37   x.xx    Qf3
   9/23   8.78   0.19    Qf3 Rac8 Nh6+ Kf8 Bxf6 Qxf6 Qxf6 gxf6 Bc4 Rc7 Nf5
   9-->   9.29   0.19    Qf3 Rac8 Nh6+ Kf8 Bxf6 Qxf6 Qxf6 gxf6 Bc4 Rc7 Nf5
  10/23  14.63   0.19    Qf3 Rac8 Nh6+ Kf8 Bxf6 Qxf6 Qxf6 gxf6 f4 Kg7 Nf5+
                         Kh8 Bc4
  10-->  19.43   0.19    Qf3 Rac8 Nh6+ Kf8 Bxf6 Qxf6 Qxf6 gxf6 f4 Kg7 Nf5+
                         Kh8 Bc4
  11/27   1:08   0.23    Qf3 Kh8 Qe2 Nd7 Rfd1 Ne7 Nxe7 Bxe7 Bc4 Qd6 Be3 Qc7
  11-->   1:20   0.23    Qf3 Kh8 Qe2 Nd7 Rfd1 Ne7 Nxe7 Bxe7 Bc4 Qd6 Be3 Qc7
  12/31   3:40   0.26    Qf3 Rac8 c3 dxc3 bxc3 Kh8 Rfb1 Rc7 Bc2 Nd7 Rd1 Rec8
                         Qe2
  12-->   4:08   0.26    Qf3 Rac8 c3 dxc3 bxc3 Kh8 Rfb1 Rc7 Bc2 Nd7 Rd1 Rec8
                         Qe2
  13/33  11:04   0.27    Qf3 Nd7 Rfd1 Ne7 Bb5 Nc6 c3 a4 Rac1 Qg6 cxd4 exd4
                         Bf4 Nde5
  13-->  12:02   0.27    Qf3 Nd7 Rfd1 Ne7 Bb5 Nc6 c3 a4 Rac1 Qg6 cxd4 exd4
                         Bf4 Nde5
  14...  12:02   1/43    Qf3

Time: 16:39, cpu 95.57%, nodes 458063864, nps 458504
Hash: 128M, used 99%, hits 23%
EGTB: probs 0, hits 0

Chess Tiger 14 likes Rc1:
r3r1k1/1p3ppp/2n1qn2/p1b1pNB1/3pP3/P2B4/1PP2PPP/R2Q1RK1 w - - ce 56; acd 14; acs
4; pv Rc1 Ne7 Qe2 Nxf5 exf5 Qb6 Bb5 Rec8 c3 dxc3 Rxc3 Nd5 Rf3 f6 Bc4;|Ra1c1

ChessMaster:
Time	Depth	Score	Positions	Moves
8:55	3/12	0.68	58866439	1.Qf3 Nd7 2.Bb5 Rec8 3.Rfd1 a4
					4.Rac1 g6 5.Ng3 Be7 6.Bd2 Bh4

The Chessmaster recommends: Queen to f3.

Analysis: You move your queen to f3. Black replies by moving the knight to d7.
You move your bishop to b5. Black responds with rook at e8 to c8. You move your
rook at f1 to d1. Black responds with the pawn to a4. You move your rook at a1
to c1. Black counters by moving the pawn to g6, which attacks your knight. You
move your knight to g3, which moves it to safety. Black counters by moving the
bishop to e7, which threatens your bishop at g5. You move your bishop to d2,
which moves it to safety. Black responds by moving bishop to h4.

As a result of this sequence of moves, Black's mobility is slightly decreased.
Additionally, Black's pawn structure slightly weaker.


Phalanx:
r3r1k1/1p3ppp/2n1qn2/p1b1pNB1/3pP3/P2B4/1PP2PPP/R2Q1RK1 w - -
    -> increment adds 1440 s to soft time limit
    -> soft time limit 3154.28 s
    -> hard time limit 41599.9 s
  6     35    63    58602  Qd1-e2  Nf6-d7  Pf2-f3  Pa5-a4
  6 ->   0:01.16   124167   0 turns
  7     55   231   289628  Qd1-e2  Nf6-d7  Qe2-g4
  7 ->   0:03.21   452253   0 turns
  8     36   471   728667  Qd1-e2  Nf6-d7  Qe2-g4  Qe6-g6  Nf5-h4  Qg6-e6
                           Qg4xe6  Re8xe6  Bd3-c4  Re6-d6
  8     40  1225  2022612  Pf2-f4  Nf6-d7  Qd1-g4  Qe6-g6  Qg4-f3
  8     45  1370  2317380  Bd3-c4  !
  8     70  1420  2443319  Bd3-c4  Qe6xc4  Bg5xf6  Pg7-g6  Nf5-h6  Kg8-f8
                           Qd1-d2  Bc5-e7  Pb2-b3  Qc4-e6
  8 ->   0:14.45  2474828   2 turns
  9     75  1688  2921312  Bd3-c4  Qe6xc4  Bg5xf6  Pg7-g6  Qd1-d2  Pa5-a4
  9 ->   0:24.03  4177378   0 turns
 10     59  3091  5453579  Bd3-c4  Qe6xc4  Bg5xf6  Pg7-g6  Qd1-d2  Pa5-a4
                           Nf5-h6  Kg8-f8  Ra1-d1  Re8-e6  Nh6-g4  Ra8-c8
 10 ->   0:52.36  9412593   0 turns
 11     62  8140 13821490  Bd3-c4  Qe6xc4  Bg5xf6  Pg7-g6  Qd1-d2  Pa5-a4
                           Nf5-h6  Kg8-f8  Qd2-g5  Ra8-a6  Ra1-c1  Ra6-b6
                           Pb2-b4
 11 ->   4:43.81 45012677   0 turns
 12     64 36867 58734717  Bd3-c4  Qe6xc4  Bg5xf6  Pg7-g6  Qd1-d2  Pa5-a4
                           Nf5-h6  Kg8-f8  Rf1-e1  Re8-e6  Nh6-g4  Kf8-g8
                           Qd2-h6  Bc5-f8
 12 ->  10:19.54 98381024   0 turns
 13     70 89488 141218659  Bd3-c4  Qe6xc4  Bg5xf6  Pg7-g6  Qd1-d2  Re8-e6
                           Nf5-h6  Kg8-f8  Nh6-g4  Kf8-g8  Pb2-b3  Qc4-c3
                           Qd2xc3  Pd4xc3  Rf1-d1  Pa5-a4  Ng4-h6  Kg8-f8
 13 ->  28:51.54 271014655   0 turns




This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.