Stockfish needs desperately tablebase support

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

Moderators: hgm, Rebel, chrisw

Jouni
Posts: 3291
Joined: Wed Mar 08, 2006 8:15 pm

Stockfish needs desperately tablebase support

Post by Jouni »

I was analysing very difficult correspondence game ending with Stockfish (great endgame player). In one variation happened this:

[D]7R/8/1KP5/8/8/4k3/4p2p/8 w - -

Analysis by Stockfish 1.5.1 JA:

19.c7 e1Q 20.Re8+ Kd2 21.Rxe1 Kxe1 22.c8Q Kf2 23.Qh3 Kg1 24.Qg3+ Kh1 25.Qd6 Kg1 26.Qd1+ Kg2 27.Qd5+ Kg1 28.Qd4+ Kg2 29.Qd2+ Kg1 30.Qc1+ Kg2 31.Qc2+ Kg1 32.Qb1+ Kg2 33.Qg6+ Kf2 34.Qh6 Kg1
+- (8.00) Depth: 18 00:00:23 15195kN
19.c7 e1Q 20.Re8+ Kd2 21.Rxe1 Kxe1 22.c8Q Kf2 23.Qh3 Kg1 24.Qg3+ Kh1 25.Qd6 Kg1 26.Qd1+ Kg2 27.Qd5+ Kg1 28.Qd4+ Kg2 29.Qd2+ Kg1 30.Qc1+ Kg2 31.Qc2+ Kg1 32.Qb1+ Kg2 33.Qg6+ Kf2 34.Qf5+ Kg1
+- (8.04) Depth: 19 00:00:40 26714kN
19.c7 e1Q 20.Re8+ Kd2 21.Rxe1 Kxe1 22.c8Q Kf2 23.Qh3 Kg1 24.Qg3+ Kh1 25.Qd6 Kg1 26.Qd1+ Kg2 27.Qd5+ Kg1 28.Qd4+ Kg2 29.Qd2+ Kg1 30.Qc1+ Kg2 31.Qc2+ Kg1 32.Qb1+ Kg2 33.Qg6+ Kf2 34.Qf5+ Kg1
+- (8.04) Depth: 20 00:01:12 47734kN

Position is draw of course. So in analysis You simply can't trust engines without table/bitbase support! Yes I know in playing their importance is
very small...

Jouni
Top
gerold
Posts: 10121
Joined: Thu Mar 09, 2006 12:57 am
Location: van buren,missouri

Re: Stockfish needs desperately tablebase support

Post by gerold »

Jouni wrote:I was analysing very difficult correspondence game ending with Stockfish (great endgame player). In one variation happened this:

[D]7R/8/1KP5/8/8/4k3/4p2p/8 w - -

Analysis by Stockfish 1.5.1 JA:

19.c7 e1Q 20.Re8+ Kd2 21.Rxe1 Kxe1 22.c8Q Kf2 23.Qh3 Kg1 24.Qg3+ Kh1 25.Qd6 Kg1 26.Qd1+ Kg2 27.Qd5+ Kg1 28.Qd4+ Kg2 29.Qd2+ Kg1 30.Qc1+ Kg2 31.Qc2+ Kg1 32.Qb1+ Kg2 33.Qg6+ Kf2 34.Qh6 Kg1
+- (8.00) Depth: 18 00:00:23 15195kN
19.c7 e1Q 20.Re8+ Kd2 21.Rxe1 Kxe1 22.c8Q Kf2 23.Qh3 Kg1 24.Qg3+ Kh1 25.Qd6 Kg1 26.Qd1+ Kg2 27.Qd5+ Kg1 28.Qd4+ Kg2 29.Qd2+ Kg1 30.Qc1+ Kg2 31.Qc2+ Kg1 32.Qb1+ Kg2 33.Qg6+ Kf2 34.Qf5+ Kg1
+- (8.04) Depth: 19 00:00:40 26714kN
19.c7 e1Q 20.Re8+ Kd2 21.Rxe1 Kxe1 22.c8Q Kf2 23.Qh3 Kg1 24.Qg3+ Kh1 25.Qd6 Kg1 26.Qd1+ Kg2 27.Qd5+ Kg1 28.Qd4+ Kg2 29.Qd2+ Kg1 30.Qc1+ Kg2 31.Qc2+ Kg1 32.Qb1+ Kg2 33.Qg6+ Kf2 34.Qf5+ Kg1
+- (8.04) Depth: 20 00:01:12 47734kN

Position is draw of course. So in analysis You simply can't trust engines without table/bitbase support! Yes I know in playing their importance is
very small...

Jouni
Some engines just don't play well at all in the end game.
Stockfish is one of them.

Best.

Gerold.
Top
User avatar
michiguel
Posts: 6401
Joined: Thu Mar 09, 2006 8:30 pm
Location: Chicago, Illinois, USA

Re: Stockfish needs desperately tablebase support

Post by michiguel »

Jouni wrote:I was analysing very difficult correspondence game ending with Stockfish (great endgame player). In one variation happened this:

[D]7R/8/1KP5/8/8/4k3/4p2p/8 w - -

Analysis by Stockfish 1.5.1 JA:

19.c7 e1Q 20.Re8+ Kd2 21.Rxe1 Kxe1 22.c8Q Kf2 23.Qh3 Kg1 24.Qg3+ Kh1 25.Qd6 Kg1 26.Qd1+ Kg2 27.Qd5+ Kg1 28.Qd4+ Kg2 29.Qd2+ Kg1 30.Qc1+ Kg2 31.Qc2+ Kg1 32.Qb1+ Kg2 33.Qg6+ Kf2 34.Qh6 Kg1
+- (8.00) Depth: 18 00:00:23 15195kN
19.c7 e1Q 20.Re8+ Kd2 21.Rxe1 Kxe1 22.c8Q Kf2 23.Qh3 Kg1 24.Qg3+ Kh1 25.Qd6 Kg1 26.Qd1+ Kg2 27.Qd5+ Kg1 28.Qd4+ Kg2 29.Qd2+ Kg1 30.Qc1+ Kg2 31.Qc2+ Kg1 32.Qb1+ Kg2 33.Qg6+ Kf2 34.Qf5+ Kg1
+- (8.04) Depth: 19 00:00:40 26714kN
19.c7 e1Q 20.Re8+ Kd2 21.Rxe1 Kxe1 22.c8Q Kf2 23.Qh3 Kg1 24.Qg3+ Kh1 25.Qd6 Kg1 26.Qd1+ Kg2 27.Qd5+ Kg1 28.Qd4+ Kg2 29.Qd2+ Kg1 30.Qc1+ Kg2 31.Qc2+ Kg1 32.Qb1+ Kg2 33.Qg6+ Kf2 34.Qf5+ Kg1
+- (8.04) Depth: 20 00:01:12 47734kN

Position is draw of course. So in analysis You simply can't trust engines without table/bitbase support! Yes I know in playing their importance is
very small...

Jouni
Stockfish in this case lacks a specific knowledge of Q vs P endgame.TBs help but it could be done without them.

Gaviota without TBs get only a +0.27 score (Gaviota is not a good endgame player at all)

Code: Select all

setboard 7R/8/1KP5/8/8/4k3/4p2p/8 w - - 
d
+-----------------+
| . . . . . . . R |
| . . . . . . . . |
| . K P . . . . . |
| . . . . . . . . |    Castling: 
| . . . . . . . . |    ep: -
| . . . . k . . . |
| . . . . p . . p |
| . . . . . . . . | [White]
+-----------------+

analyze
********* Starts iterative deepening, thread = 0
set timer to infinite
         5   1       0.0    -1.06  1.Rxh2 e1=Q
       271   1       0.0    +0.07  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   h1=Q
       350   1       0.0    +2.56  1.Re8+ Kd2
       398   1       0.0    +2.67  1.Rh3+ Ke4
       714   2       0.0    +2.67  1.Rh3+ Ke4 2.Rh4+ Kd3 3.Rh3+ Ke4
      2961   3       0.0      :-(  1.Rh3+
      5280   3       0.0      :-(  1.Rh3+
      8191   3       0.0    -0.00  1.Rh3+ Kf4 2.c7 e1=Q 3.c8=Q h1=Q 4.Rxh1
                                   Qxh1
     10561   3       0.0    +0.07  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   h1=Q
     14527   4       0.0    +0.07  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   h1=Q
     27780   5       0.1    +0.07  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   h1=Q
     49458   6       0.1      :-)  1.c7
     58229   6       0.1      :-)  1.c7
     83043   6       0.1    +2.78  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   Kf2
    101281   6:      0.2    +2.78  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   Kf2
    146970   7       0.2      :-)  1.c7
    322660   7       0.5      :-)  1.c7
    430171   7       0.7    +4.29  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   Kf2 5.Qc6
    463191   7:      0.8    +4.29  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   Kf2 5.Qc6
    696604   8       1.1      :-(  1.c7
    852034   8       1.4      :-(  
    987772   8       1.6      :-(  1.c7
   1174870   8       1.9      :-(  
   1423901   8       2.3    +1.18  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=R
                                   Kf2 5.Rc1 Kg2 6.Rc2+ Kg1
   1604428   8:      2.6    +1.18  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=R
                                   Kf2 5.Rc1 Kg2 6.Rc2+ Kg1
   2372818   9       3.9    +0.96  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=R
                                   Kf1 5.Rh8 Kg2 6.Rg8+ Kf2
   2830543   9:      4.6    +0.96  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=R
                                   Kf1 5.Rh8 Kg2 6.Rg8+ Kf2
   6309616  10      10.4      :-(  1.c7
   7373578  10      12.1      :-(  
  11008538  10      18.0    +0.27  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   Kf2 5.Qh3 Kg1 6.Qg3+ Kh1 7.Qe5
  11996223  10:     19.5    +0.27  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   Kf2 5.Qh3 Kg1 6.Qg3+ Kh1 7.Qe5
  20977822  11      34.3    +0.27  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   Kf2 5.Qh3 Kg1 6.Qg3+ Kh1 7.Qf3+ Kg1
                                   8.Kc5 h1=Q
  24591724  11:     39.9    +0.27  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   Kf2 5.Qh3 Kg1 6.Qg3+ Kh1 7.Qf3+ Kg1
                                   8.Kc5 h1=Q
  83573923  12     135.4    +0.27  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   Kf2 5.Qh3 Kg1 6.Qg3+ Kh1 7.Qe1+ Kg2
                                   8.Kc5 h1=Q 9.Qe4+ Kh2
  98202365  12:    158.0    +0.27  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   Kf2 5.Qh3 Kg1 6.Qg3+ Kh1 7.Qe1+ Kg2
                                   8.Kc5 h1=Q 9.Qe4+ Kh2
with GTBs the score is 0.00

Code: Select all

analyze
********* Starts iterative deepening, thread = 0
set timer to infinite
         5   1       0.0    -1.06  1.Rxh2 e1=Q
       271   1       0.0    +0.07  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   h1=Q
       350   1       0.0    +2.56  1.Re8+ Kd2
       398   1       0.0    +2.67  1.Rh3+ Ke4
       714   2       0.0    +2.67  1.Rh3+ Ke4 2.Rh4+ Kd3 3.Rh3+ Ke4
      2877   3       0.0      :-(  1.Rh3+
      4942   3       0.0      :-(  1.Rh3+
      7370   3       0.0    -0.00  1.Rh3+ Kf4 2.c7 e1=Q 3.c8=Q h1=Q 4.Rxh1
                                   Qxh1
      9740   3       0.0    +0.07  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   h1=Q
     15074   4       0.0    +0.07  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   h1=Q
     28112   5       0.1    +0.00  1.c7 e1=Q 2.Re8+ Kd2 3.Rxe1 Kxe1 4.c8=Q
                                   <=TABLE
    125842   6       0.3    +0.01  1.c7 e1=Q 2.Rh3+ Kf4 3.c8=Q Qb1+ 4.Ka7
                                   Qa1+ 5.Kb8 h1=Q 6.Qc7+ Kf5 7.Rxh1 Qxh1
    139355   6&#58;      0.3    +0.01  1.c7 e1=Q 2.Rh3+ Kf4 3.c8=Q Qb1+ 4.Ka7
                                   Qa1+ 5.Kb8 h1=Q 6.Qc7+ Kf5 7.Rxh1 Qxh1
    390676   7       0.7    +0.02  1.c7 e1=Q 2.Rh3+ Kf4 3.c8=Q Qb1+ 4.Ka7
                                   h1=Q 5.Qc7+ Kg4 6.Qd7+ Kg5 7.Qd2+ Kf5
                                   8.Rxh1 Qxh1
    416959   7&#58;      0.8    +0.02  1.c7 e1=Q 2.Rh3+ Kf4 3.c8=Q Qb1+ 4.Ka7
                                   h1=Q 5.Qc7+ Kg4 6.Qd7+ Kg5 7.Qd2+ Kf5
                                   8.Rxh1 Qxh1
    889150   8       1.7    +0.00  1.c7 e1=Q 2.Rh3+ Kf2 3.Rxh2+ <=TABLE
    947188   8       1.8    +0.05  1.Rh3+ Kf4 2.c7 e1=Q 3.c8=Q Qb1+ 4.Ka7
                                   Qa1+ 5.Kb8 h1=Q 6.Qc4+ Ke5 7.Rxh1 Qxh1
    996023   8&#58;      1.9    +0.05  1.Rh3+ Kf4 2.c7 e1=Q 3.c8=Q Qb1+ 4.Ka7
                                   Qa1+ 5.Kb8 h1=Q 6.Qc4+ Ke5 7.Rxh1 Qxh1
   1597033   9       3.0    +0.00  1.Rh3+ Kf4 2.c7 e1=Q 3.c8=Q Qg1+ 4.Qc5
                                   h1=Q 5.Rxh1 <=TABLE
   2098332   9&#58;      3.9    +0.00  1.Rh3+ Kf4 2.c7 e1=Q 3.c8=Q Qg1+ 4.Qc5
                                   h1=Q 5.Rxh1 <=TABLE
   2681230  10       5.0    +0.00  1.Rh3+ Kf4 2.c7 e1=Q 3.c8=Q Qg1+ 4.Qc5
                                   h1=Q 5.Rxh1 <=TABLE
   4691222  10&#58;      8.6    +0.00  1.Rh3+ Kf4 2.c7 e1=Q 3.c8=Q Qg1+ 4.Qc5
                                   h1=Q 5.Rxh1 <=TABLE
  18493330  11      33.8    +0.00  1.Rh3+ Kf4 2.c7 e1=Q 3.c8=Q Qg1+ 4.Qc5
                                   h1=Q 5.Rxh1 <=TABLE
  29935631  11&#58;     56.1    +0.00  1.Rh3+ Kf4 2.c7 e1=Q 3.c8=Q Qg1+ 4.Qc5
                                   h1=Q 5.Rxh1 <=TABLE
Top
User avatar
AdminX
Posts: 6340
Joined: Mon Mar 13, 2006 2:34 pm
Location: Acworth, GA

Re: Stockfish needs desperately tablebase support

Post by AdminX »

Analysis by Yace 0.99.87:

1.Re8+ Kf3 2.Rxe2
= (0.00) Depth: 6/17 00:00:00 28kN

1.Rh3+ Kf4 2.c7 e1Q 3.c8Q Qg1+ 4.Kc7 h1Q 5.Qf8+ Kg4 6.Rxh1 Qxh1 7.Qg8+ Kf4 8.Qf8+ Kg4 9.Qg8+ Kf4 10.Qf8+ Kg4
= (0.01 !) Depth: 6/17 00:00:00 37kN
= (0.00) Depth: 13/33 00:00:53 30030kN
"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers
Top

Return to “Computer Chess Club: General Topics”