Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Testposition - Pawn Magic (Eval w/ Crafty - Attn: Bob)

Author: Uri Blass

Date: 13:50:57 08/19/01

Go up one level in this thread


On August 19, 2001 at 02:57:37, Slater Wold wrote:

>On August 18, 2001 at 15:18:32, Sune Larsson wrote:
>
>>
>>  [D]8/pR4pk/1b6/2p5/N1p5/8/PP1r2PP/6K1 b - - 0 1
>>
>>        Ortueta - Sanz, 1934
>>
>>
>>  This position might prove to be a good test. Deep Fritz had some problems
>>  here. Found the correct move after 13:14 with -0.28, then went green
>>  after 35:39 with -2.47. PIII 800 - 64 MB Hash.
>>
>>  Black wins after 1.-Rxb2! 2.Nxb2 c3 3.Rxb6 c4! 4.Rb4 a5!  True Pawn Magic!!
>>
>>
>>  Ortueta - Sanz
>>  8/pR4pk/1b6/2p5/N1p5/8/PP1r2PP/6K1 b - - 0 1
>>
>>  Analysis by Deep Fritz: PIII 800 Mhz
>>
>>
>>1...Bd8 2.Rxa7 Bf6 3.Rb7 Bd4+ 4.Kf1 Rf2+ 5.Ke1 Rxg2 6.h3
>>  =  (0.19)   Depth: 8/20   00:00:00  67kN
>>1...Bd8 2.Rxa7 Bf6 3.Kf1 Bxb2 4.Nxb2 Rxb2 5.a4 c3 6.a5
>>  =  (0.22)   Depth: 9/20   00:00:00  123kN
>>1...c3!
>>  =  (0.19)   Depth: 9/20   00:00:00  149kN
>>1...c3! 2.bxc3 Rxa2 3.Nxb6 axb6 4.Rxb6 Ra3 5.c4
>>  =  (0.06)   Depth: 9/24   00:00:00  161kN
>>1...c3 2.bxc3 Rxa2 3.Nxb6 axb6 4.Rxb6 Ra3 5.c4 Rc3 6.Kf2 Rxc4
>>  =  (0.06)   Depth: 10/26   00:00:00  242kN
>>1...Bd8!
>>  =  (0.03)   Depth: 10/27   00:00:01  382kN
>>1...Bd8 2.h3 Bf6 3.Kf1 Kg6 4.Rxa7 Bxb2 5.Nxb2 Rxb2 6.a4 c3 7.a5
>>  =  (-0.03)   Depth: 11/25   00:00:02  1076kN
>>1...Bd8 2.h3 Bf6 3.Kf1 Kg6 4.a3 Bd4 5.g4 Rh2 6.Rxa7 Bxb2
>>  =  (-0.06)   Depth: 12/27   00:00:03  2041kN
>>1...Bd8 2.h3 Bf6 3.Kf1 Kg6 4.g4 Rh2 5.Rxa7 Bxb2 6.Nxb2 Rxb2 7.a4
>>  =  (-0.03)   Depth: 13/28   00:00:07  4281kN
>>1...Bd8 2.h3 Bf6 3.Kf1 Kg6 4.a3 a6 5.Rb6 Kf5 6.Nxc5 Bxb2 7.Rxa6
>>  =  (-0.06)   Depth: 14/31   00:00:13  8712kN
>>1...Bd8 2.h3 Bf6 3.Kf1 Kg6 4.a3 Bd4 5.g4 Rh2 6.Rxa7 Rxh3 7.Re7
>>  =  (-0.16)   Depth: 15/32   00:00:36  23307kN
>>1...Bd8 2.h3 Bf6 3.Kf1 Kg6 4.g4 Rh2 5.Rxa7 Bxb2 6.Nxb2 Rxb2 7.Ra6+
>>  =  (-0.16)   Depth: 16/34   00:01:27  56494kN
>>1...c3!
>>  =  (-0.19)   Depth: 16/34   00:01:33  60739kN, tb=4
>>1...c3 2.bxc3 Rxa2 3.Nxb6 axb6 4.Rxb6 Ra3 5.c4 Rc3 6.Kf2 Rxc4 7.Re6
>>  =  (-0.19)   Depth: 17/37   00:02:08  84446kN, tb=19
>>1...c3 2.bxc3 Rxa2 3.Nxb6 axb6 4.Rxb6 Ra3 5.c4 Rc3 6.Kf2 Rxc4 7.Re6
>>  =  (-0.22)   Depth: 18/36   00:04:21  173788kN, tb=70
>>1...Bd8!
>>  =  (-0.25)   Depth: 18/37   00:11:10  443095kN, tb=84
>>1...Rxb2!
>>  ³  (-0.28)   Depth: 18/38   00:13:14  527020kN, tb=84
>>1...Rxb2! 2.Nxb2 c3 3.Rxb6 c4 4.Rb4 a5 5.Rb5 c2 6.Kf2 c1Q 7.Ke2
>>  -+  (-2.47)   Depth: 18/38   00:35:39  1415072kN, tb=108
>>
>>(Larsson, KIL 18.08.2001)
>
>
>hash table memory = 48M bytes.
>pawn hash table memory = 10M bytes.
>5 piece tablebase files found
>20386kb of RAM used for TB indices and decompression tables
>EGTB cache memory = 16M bytes.
>
>
>Crafty v18.10 (1 cpus)
>
>White(1): force
>White(1): setboard 8/pR4pk/1b6/2p5/N1p5/8/PP1r2PP/6K1 b - - 0 1
>
>       +---+---+---+---+---+---+---+---+
>    8  |   |   |   |   |   |   |   |   |
>       +---+---+---+---+---+---+---+---+
>    7  | *P| R |   |   |   |   | *P| *K|
>       +---+---+---+---+---+---+---+---+
>    6  |   | *B|   |   |   |   |   |   |
>       +---+---+---+---+---+---+---+---+
>    5  |   |   | *P|   |   |   |   |   |
>       +---+---+---+---+---+---+---+---+
>    4  | N |   | *P|   |   |   |   |   |
>       +---+---+---+---+---+---+---+---+
>    3  |   |   |   |   |   |   |   |   |
>       +---+---+---+---+---+---+---+---+
>    2  | P | P |   | *R|   |   | P | P |
>       +---+---+---+---+---+---+---+---+
>    1  |   |   |   |   |   |   | K |   |
>       +---+---+---+---+---+---+---+---+
>         a   b   c   d   e   f   g   h
>
>Black(1): name Administrator
>Crafty 18.10 vs Administrator
>Black(1): time 180000
>time remaining:  30:00 (crafty).
>Black(1): otim 180000
>time remaining:  30:00 (opponent).
>Black(1): go
>end-game phase
>              clearing hash tables
>              time surplus   0.00  time limit 30:00 (15:00)
>         nss  depth   time  score   variation (1)
>                7     0.19   0.50   1. ... Bd8 2. Rxa7 Bf6 3. Rf7 Bxb2
>                                    4. Nxb2 Rxb2 5. a4
>                7->   0.20   0.50   1. ... Bd8 2. Rxa7 Bf6 3. Rf7 Bxb2
>                                    4. Nxb2 Rxb2 5. a4
>                8     0.34   0.26   1. ... Bd8 2. Rxa7 Bf6 3. Rb7 Bd4+
>                                    4. Kf1 Rf2+ 5. Ke1 Rxg2 6. h4
>                8->   0.42   0.26   1. ... Bd8 2. Rxa7 Bf6 3. Rb7 Bd4+
>                                    4. Kf1 Rf2+ 5. Ke1 Rxg2 6. h4
>                9     0.67   0.01   1. ... Bd8 2. Rxa7 Bf6 3. Rb7 Bd4+
>                                    4. Kf1 Rf2+ 5. Ke1 Rxg2 6. h4 Rf2
>                9->   0.81   0.01   1. ... Bd8 2. Rxa7 Bf6 3. Rb7 Bd4+
>                                    4. Kf1 Rf2+ 5. Ke1 Rxg2 6. h4 Rf2
>               10     1.53   0.04   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g4 a6 5. Ke1 Rh2 6. Rb6
>               10->   1.97   0.04   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g4 a6 5. Ke1 Rh2 6. Rb6
>               11     3.50  -0.07   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    Rxa7 Rf2+ 5. Ke1 Rxg2 6. Rb7 Rh2 7.
>                                    Rb5
>               11->   4.08  -0.07   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    Rxa7 Rf2+ 5. Ke1 Rxg2 6. Rb7 Rh2 7.
>                                    Rb5
>               12     7.09  -0.09   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g3 Rf2+ 5. Ke1 Rh2 6. Rxa7 Bxb2 7.
>                                    Nxb2 Rxb2 8. Ra5 <HT>
>               12->   8.29  -0.09   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g3 Rf2+ 5. Ke1 Rh2 6. Rxa7 Bxb2 7.
>                                    Nxb2 Rxb2 8. Ra5 <HT>
>               13    18.58  -0.17   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g3 Kg6 5. a3 Rf2+ 6. Ke1 Rh2 7. g4
>                                    a6 8. Rb6+ Kg5
>               13->  21.90  -0.17   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g3 Kg6 5. a3 Rf2+ 6. Ke1 Rh2 7. g4
>                                    a6 8. Rb6+ Kg5
>               14    43.61  -0.17   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g4 Rh2 5. Rxa7 Bxb2 6. Nxb2 Rxb2 7.
>                                    Ra5 c3 8. Rxc5 c2 9. a4
>               14->  51.39  -0.17   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g4 Rh2 5. Rxa7 Bxb2 6. Nxb2 Rxb2 7.
>                                    Ra5 c3 8. Rxc5 c2 9. a4
>               15     2:41  -0.36   1. ... Bd8 2. Kf1 Bf6 3. h4 Bxh4 4.
>                                    Rxa7 c3 5. Nxc3 Rf2+ 6. Kg1 Rxb2 7.
>                                    Rc7 Bf6 8. Ne4 Bd4+ 9. Kf1 Rxa2 10.
>                                    Nxc5 Bxc5 11. Rxc5
>               15     3:33     ++   1. ... Rxb2!!
>         (2)   15->   5:02  -0.75   1. ... Rxb2
>               16     6:32     ++   1. ... Rxb2!!
>Black(0): ?
>              time=10:10  cpu=99%  mat=0  n=507729008  fh=16%  nps=831k
>              ext-> chk=47370576 cap=408675 pp=1112610 1rep=2420213 mate=67061
>              predicted=0  nodes=507729008  evals=94120254
>              endgame tablebase-> probes done=117  successful=117
>              hashing-> trans/ref=15%  pawn=7%  used=99%
>              SMP->  split=0  stop=0  data=0/64  cpu=10:10  elap=10:10
>Black(1): Rxb2
>
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
>max threads set to 2
>hash table memory = 48M bytes.
>pawn hash table memory = 10M bytes.
>5 piece tablebase files found
>20386kb of RAM used for TB indices and decompression tables
>EGTB cache memory = 16M bytes.
>
>
>Crafty v18.10 (2 cpus)
>
>White(1): force
>White(1): setboard 8/pR4pk/1b6/2p5/N1p5/8/PP1r2PP/6K1 b - - 0 1
>
>       +---+---+---+---+---+---+---+---+
>    8  |   |   |   |   |   |   |   |   |
>       +---+---+---+---+---+---+---+---+
>    7  | *P| R |   |   |   |   | *P| *K|
>       +---+---+---+---+---+---+---+---+
>    6  |   | *B|   |   |   |   |   |   |
>       +---+---+---+---+---+---+---+---+
>    5  |   |   | *P|   |   |   |   |   |
>       +---+---+---+---+---+---+---+---+
>    4  | N |   | *P|   |   |   |   |   |
>       +---+---+---+---+---+---+---+---+
>    3  |   |   |   |   |   |   |   |   |
>       +---+---+---+---+---+---+---+---+
>    2  | P | P |   | *R|   |   | P | P |
>       +---+---+---+---+---+---+---+---+
>    1  |   |   |   |   |   |   | K |   |
>       +---+---+---+---+---+---+---+---+
>         a   b   c   d   e   f   g   h
>
>Black(1): name Administrator
>Crafty 18.10 vs Administrator
>Black(1): time 180000
>time remaining:  30:00 (crafty).
>Black(1): otim 180000
>time remaining:  30:00 (opponent).
>Black(1): go
>end-game phase
>              clearing hash tables
>              time surplus   0.00  time limit 30:00 (15:00)
>         nss  depth   time  score   variation (1)
>starting thread 1
>                7->   0.13   0.50   1. ... Bd8 2. Rxa7 Bf6 3. Rf7 Bxb2
>                                    4. Nxb2 Rxb2 5. a4
>                8     0.19   0.26   1. ... Bd8 2. Rxa7 Bf6 3. Rb7 Bd4+
>                                    4. Kf1 Rf2+ 5. Ke1 Rxg2 6. h4
>                8     0.24  -0.12   1. ... Rxb2 2. Nxb2 c3 3. Nd3 c4+ 4.
>                                    Rxb6 axb6 5. Nf4 b5
>         (3)    8->   0.24  -0.12   1. ... Rxb2 2. Nxb2 c3 3. Nd3 c4+ 4.
>                                    Rxb6 axb6 5. Nf4 b5
>                9     0.24     --   1. ... Rxb2
>         (2)    9     0.27   0.67   1. ... Rxb2 2. Nxb2 c3 3. Rxb6 axb6
>                                    4. Nd3 Kg6 5. a4 Kf5
>                9     0.38     ++   1. ... Bd8!!
>                9     0.61   0.10   1. ... Bd8 2. Kf1 Bf6 3. Rxa7 Bxb2
>                                    4. Nxb2 Rxb2 5. a4 Ra2 6. a5
>                9->   0.69   0.10   1. ... Bd8 2. Kf1 Bf6 3. Rxa7 Bxb2
>                                    4. Nxb2 Rxb2 5. a4 Ra2 6. a5
>               10     1.07   0.04   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g4 Rh2 5. Ke1 a6 6. Rb6
>               10->   1.29   0.04   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g4 Rh2 5. Ke1 a6 6. Rb6
>               11     2.21  -0.07   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    Rxa7 Rf2+ 5. Ke1 Rxg2 6. Rb7 Rh2 7.
>                                    Rb5
>               11->   2.55  -0.07   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    Rxa7 Rf2+ 5. Ke1 Rxg2 6. Rb7 Rh2 7.
>                                    Rb5
>               12     5.16  -0.09   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g3 Rf2+ 5. Ke1 Rh2 6. Rxa7 Bxb2 7.
>                                    Nxb2 Rxb2 8. Ra5 <HT>
>               12->   5.88  -0.09   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g3 Rf2+ 5. Ke1 Rh2 6. Rxa7 Bxb2 7.
>                                    Nxb2 Rxb2 8. Ra5 <HT>
>               13    13.55  -0.17   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g3 Kg6 5. a3 Rf2+ 6. Ke1 Rh2 7. g4
>                                    a6 8. Rb6+ Kg5
>               13->  15.30  -0.17   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g3 Kg6 5. a3 Rf2+ 6. Ke1 Rh2 7. g4
>                                    a6 8. Rb6+ Kg5
>               14    25.68  -0.17   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g4 Rh2 5. Rxa7 Bxb2 6. Nxb2 Rxb2 7.
>                                    Ra5 c3 8. Rxc5 c2 9. a4
>               14->  29.54  -0.17   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g4 Rh2 5. Rxa7 Bxb2 6. Nxb2 Rxb2 7.
>                                    Ra5 c3 8. Rxc5 c2 9. a4
>               15     1:25  -0.25   1. ... Bd8 2. Kf1 Bf6 3. h3 Bd4 4.
>                                    g3 Rf2+ 5. Ke1 Rf3 6. Rxa7 Rxg3 7.
>                                    Rf7 Rg2 8. Kd1 Bxb2 9. Nxc5
>               15     1:52     ++   1. ... Rxb2!!
>               15     3:19  -1.93   1. ... Rxb2 2. Nxb2 c3 3. Rxb6 c4 4.
>                                    Rb4 a5 5. Rb5 c2 6. g4 c1=Q+ 7. Kg2
>                                    c3 8. Rh5+ Kg6 9. Nc4 Qc2+ 10. Kf3
>                                    Qxa2 11. Nxa5
>         (2)   15->   3:19  -1.93   1. ... Rxb2 2. Nxb2 c3 3. Rxb6 c4 4.
>                                    Rb4 a5 5. Rb5 c2 6. g4 c1=Q+ 7. Kg2
>                                    c3 8. Rh5+ Kg6 9. Nc4 Qc2+ 10. Kf3
>                                    Qxa2 11. Nxa5
>Black(0): ?
>              time=4:02  cpu=210%  mat=0  n=368346060  fh=89%  nps=1521k
>              ext-> chk=23467910 cap=366896 pp=962906 1rep=1135472 mate=77247
>              predicted=0  nodes=368346060  evals=95924812
>              endgame tablebase-> probes done=110  successful=110
>              hashing-> trans/ref=34%  pawn=7%  used=99%
>              SMP->  split=1510  stop=212  data=9/64  cpu=8:29  elap=4:02
>Black(1): Rxb2
>              time used:   4:02
>
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
>Bob, is there anyway to speed this up?  Granted, Deep Fritz would have to have
>10x this hardware to find it in the amount of time Crafty did on 1 of my 1.7Ghz
>CPU's, this _IS_ a pretty obvious move.
>
>GT 2.0 and CT 14.0 both find it in less than 10 seconds on one of my 1.7Ghz
>CPU's.  1:52 on 2x1.7Ghz is not good.  Deep Shredder found it on 2x1.7Ghz in
>less than 30 seconds.  And well, with Deep Fritz's AWEFUL endgame, it can't find
>it in any reasonable amount of time.
>
>Do you have any ideas how you might fix this?  Or is this a problem, where any
>attempt to find it faster, may result in a lowered performance elsewhere?

It is easy to find it faster by increasing the value of passed pawns or
evaluating 2 pawns in the 6th as better than a rook.

The problem may be in other positions when 2 pawns on the 6th do not win because
the opponent king is too close.

I remember that previous Crafty versions solved it faster.
It does not mean that they were better.
>
>Thanks!
>
>
>Slate
>
>PS I got an 1800 player to look at this position on ICC, and in less than 10
>seconds, he kibitzed, "Rxd2!"

I guess that you mean Rxb2
I believe that he simply knows it because Rxb2 is a famous move.
I read this move in a chess book when I was an unrated player.

Uri



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.