Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Does Your Program See the Brilliant Exchange Sac?

Author: Chuck

Date: 18:44:59 07/24/01

Go up one level in this thread


On July 23, 2001 at 18:42:57, John Merlino wrote:

>On July 23, 2001 at 18:22:49, Dann Corbit wrote:
>
>>On July 23, 2001 at 12:21:38, odell hall wrote:
>>
>>>[D] r2qr1k1/pp2ppbp/2p1nnp1/6B1/2PP4/5N1P/PPBQ1PP1/R3R1K1 w - - 0 15
>>>
>>>  Gambit2  Finds this immediately
>>>
>>>
>>>hall,o - Junior 7
>>>r2qr1k1/pp2ppbp/2p1nnp1/6B1/2PP4/5N1P/PPBQ1PP1/R3R1K1 w - - 0 1
>>>
>>>Analysis by Gambit Tiger 2.0:
>>>
>>>1.Bxf6 Bxf6 2.d5 Nc5 3.dxc6 Qxd2 4.Nxd2 bxc6
>>>  ²  (0.40)   Depth: 6   00:00:00  95kN
>>>1.Bxf6 Bxf6 2.d5 Nc5 3.Rad1 cxd5 4.b4 Na6 5.cxd5
>>>  ²  (0.42)   Depth: 7   00:00:00  112kN
>>>1.Bxf6 Bxf6 2.d5 Nc5 3.Rad1 cxd5 4.b4 Ne4 5.Qxd5 Qxd5 6.Rxd5
>>>  ²  (0.34)   Depth: 8   00:00:01  207kN
>>>1.Rxe6 fxe6 2.Re1 Qb6 3.b3 Rad8 4.Rxe6 Rd6 5.Qe3
>>>  ²  (0.48)   Depth: 8   00:00:01  260kN
>>>1.Rxe6 fxe6 2.Qe3 Qd6 3.Bf4 Qd7 4.Ne5 Qd8 5.Bh6 Bxh6 6.Qxh6 Qxd4
>>>  ²  (0.48)   Depth: 9   00:00:02  452kN
>>>1.Rxe6 fxe6 2.Qe3 Qd6 3.Bf4 Qd7 4.Ng5 Bh6 5.Nxe6 Bxf4 6.Nxf4
>>>  ²  (0.54)   Depth: 10   00:00:03  796kN
>>>1.Rxe6 fxe6 2.Qe3 Qd6 3.Bf4 Qd7 4.Ng5 Bh6 5.Nxe6 Bxf4 6.Nxf4 Red8 7.Ne6
>>>  ²  (0.48)   Depth: 11   00:00:07  1568kN
>>>hall,o - Junior 7
>>>r2qr1k1/pp2ppbp/2p1nnp1/6B1/2PP4/5N1P/PPBQ1PP1/R3R1K1 w - - 0 1
>>>
>>>Analysis by Gambit Tiger 2.0:
>>>
>>>1.Bxf6 Bxf6 2.d5 Nc5 3.dxc6 Qxd2 4.Nxd2 bxc6
>>>  ²  (0.40)   Depth: 6   00:00:00  95kN
>>>1.Bxf6 Bxf6 2.d5 Nc5 3.Rad1 cxd5 4.b4 Na6 5.cxd5
>>>  ²  (0.42)   Depth: 7   00:00:00  112kN
>>>1.Bxf6 Bxf6 2.d5 Nc5 3.Rad1 cxd5 4.b4 Ne4 5.Qxd5 Qxd5 6.Rxd5
>>>  ²  (0.34)   Depth: 8   00:00:01  207kN
>>>1.Rxe6 fxe6 2.Re1 Qb6 3.b3 Rad8 4.Rxe6 Rd6 5.Qe3
>>>  ²  (0.48)   Depth: 8   00:00:01  260kN
>>>1.Rxe6 fxe6 2.Qe3 Qd6 3.Bf4 Qd7 4.Ne5 Qd8 5.Bh6 Bxh6 6.Qxh6 Qxd4
>>>  ²  (0.48)   Depth: 9   00:00:02  452kN
>>>1.Rxe6 fxe6 2.Qe3 Qd6 3.Bf4 Qd7 4.Ng5 Bh6 5.Nxe6 Bxf4 6.Nxf4
>>>  ²  (0.54)   Depth: 10   00:00:03  796kN
>>>1.Rxe6 fxe6 2.Qe3 Qd6 3.Bf4 Qd7 4.Ng5 Bh6 5.Nxe6 Bxf4 6.Nxf4 Red8 7.Ne6
>>>  ²  (0.48)   Depth: 11   00:00:07  1568kN
>>
>>That is quite a beautiful move.
>>
>>Beowulf plays the cowardly (in comparison):
>>g5e3 g8f8 d2c1 d8d7 g1h2 a8d8 c4c5 h7h6 f3e5 d7d5
>>
>>But I am relieve to see that Phalanx (which has a propensity for brilliant moves
>>from time to time) plays the same dull move (with a much better plan):
>> 11    106 28454 53104772  Bg5-e3  Qd8-c7  Ra1-d1  Pb7-b5  Pb2-b3  Ra8-b8
>>                           Rd1-b1  Pb5xc4  Pb3xc4  Pc6-c5  Pd4-d5  Ne6-f8
>>
>>As does ExChess:
>> 13.   0.59   412 144845525   1. Be3 Nf8 2. Rad1 Qc8 3. Bf4 Ne6 4. Be5 Nd7 5.
>>Bh2 c5 6. dxc5 Rd8 7. Be5 Nexc5
>>
>>And Amy:
>>13    5:40  +0.384  1. Be3 Qc7 2. Qb4 b6 3. Qa3 Rad8 4. Bd3 Nd7 5. Re2 c5 6.
>>                    d5
>>        13     38   3400771924835 1. Be3 Qc7 2. Qb4 b6 3. Qa3 Rad8 4. Bd3 Nd7 5.
>>Re2 c5 6. d5
>>13    7:22  +0.384  1. Be3 Qc7 2. Qb4 b6 3. Qa3 Rad8 4. Bd3 Nd7 5. Re2 c5 6.
>>                    d5
>>        13     38   4425895899769 1. Be3 Qc7 2. Qb4 b6 3. Qa3 Rad8 4. Bd3 Nd7 5.
>>Re2 c5 6. d5
>>
>>And Bringer:
>>bm Rxe6;
>>============ Weiß / white  am Zug / to move ===
>> sT   .   .  sD  sT   .  sK   .
>> sB  sB   .   .  sB  sB  sL  sB
>>  .   .  sB   .  sS  sS  sB   .
>>  .   .   .   .   .   .  wL   .
>>  .   .  wB  wB   .   .   .   .
>>  .   .   .   .   .  wS   .  wB
>> wB  wB  wL  wD   .  wB  wB   .
>> wT   .   .   .  wT   .  wK   .
>>target response time (ms): -1
>>max. response time (ms): -1
>>Static score: mat = 5, pos = 54, total = 59
>>0:00:00.1  ( 6/12)      16093   0.52  g5-e3  d8-d6  f3-e5  f6-d7  e5xd7  d6xd7
>>0:00:00.3  ( 7/14)      58838   0.60  g5-e3  d8-d6  c4-c5  d6-d5  c2-b3  d5-e4
>>0:00:01.2  ( 8/14)     226263   0.45  g5-e3  d8-d6  d2-c3  a8-d8  c4-c5  d6-d5
>>0:00:03.2  ( 9/16)     619151   0.51  g5-e3  d8-d6  c4-c5  d6-c7
>>0:00:11.2  (10/17)    2018162   0.46  g5-e3  d8-c7
>>0:00:41.4  (11/24)    7827912   0.49  g5-e3  d8-c7  d2-d3  a8-d8
>>0:01:01.8  (12/24)   11640244   0.45  g5-e3  d8-c7  d2-d3  a8-d8  d3-a3  b7-b6
>>a1-d1  e8-f8  a3-b3  c6-c5  d4-d5  e6-f4
>>
>> 1.  g5-e3
>>
>>
>>Crafty [on the other hand] goes for the pedestrian:
>>12->   3:16   0.58   1. Bh4 Nh5 2. Rad1 Qc7 3. Qe3 c5 4.
>>                     d5 Nd4 5. Ba4 Nf5 6. Qg5 Nxh4 7. Nxh4
>>
>>As does Yace:
>>  37523348  1:56.2   0.45 10.  1.Bh4 Nh5 2.Rad1 Qc7
>> 102236489  5:26.9   0.50 11t  1.Bh4 a5 2.Rad1 Qb6 3.b3 Red8 4.Bg3 Nd7 5.Qe3H
>>                               Qb4H {HT} {10}
>>
>>
>>Gambit Tiger's move is definitely the most sexy of the lot.
>>However, all the programs seem to have a pretty low opinion of their selection.
>>Only Phalanx has an eval over one pawn in strength estimation.
>
>CM8000 also prefers "the pedestrian". But it DOES feel pretty good about it
>(comparatively):
>
>Time	Depth	Score	Positions	Moves
>0:00	2/5	1.09	12131		15. Be3 Qd6 16. Rad1 Nd7 17. Qc3 f5
>0:00	2/5	1.16	21778		15. Bh4 Qb6 16. b3 Nh5 17. Rad1 Qc7
>0:01	2/6	1.15	62789		15. Be3 Qd6 16. Rad1 a5 17. Ng5
>					Red8 18. Nxe6 Qxe6
>0:02	3/7	1.08	122767		15. Be3 Qd6 16. Rad1 Red8 17. Ng5
>					Nxg5 18. Bxg5 e6
>0:06	3/8	1.13	341852		15. Be3 Qd6 16. Rad1 Red8 17. Qa5
>					b6 18. Qc3 Rac8
>0:11	3/8	1.19	612722		15. Bh4 b5 16. b3 Nh5 17. Rad1
>					bxc4 18. bxc4 Qd6 19. g3 Rab8
>0:27	4/9	1.17	1540127		15. Bh4 Qd7 16. Rad1 Rad8 17. Qe3
>					b5 18. b3 Nh5 19. d5 cxd5 20. Rxd5
>1:01	4/10	1.08	3610156		15. Bh4 b5 16. b3 Nh5 17. Rad1
>					bxc4 18. bxc4 Rb8 19. d5 cxd5 20.
>					cxd5
>
>jm

It does seem pretty unclear which is the best move here, so here is what Hiarcs
7.32 thinks, followed by what it thinks after Rxe6 is played:

Hiarcs 7.32
r2qr1k1/pp2ppbp/2p1nnp1/6B1/2PP4/5N1P/PPBQ1PP1/R3R1K1 w - - 0 1

Analysis by Hiarcs 7.32:

1.Be3
  ±  (1.00)   Depth: 1   00:00:00
1.Be3 Qb6
  ±  (0.91)   Depth: 2/4   00:00:00
1.b4 Nxg5 2.Nxg5
  ±  (0.95)   Depth: 2/5   00:00:00
1.b4 Nxg5 2.Nxg5
  ±  (0.95)   Depth: 3/8   00:00:00
1.b4 Nxg5 2.Qxg5 Qd6 3.a3
  ±  (0.71)   Depth: 4/9   00:00:00
1.Be3 Qd6 2.Rad1 Rad8
  ±  (0.78)   Depth: 4/10   00:00:00
1.Be3 Qd6 2.Rad1 Rad8 3.Ne5
  ±  (0.74)   Depth: 5/13   00:00:00  8kN
1.Be3 Qd6 2.Rad1 Rad8 3.Qa5 a6
  ±  (0.73)   Depth: 6/17   00:00:00  47kN
1.Be3 Qd6
  ²  (0.60)   Depth: 7/21   00:00:02  151kN
1.Bh4 Qd7 2.Rad1 Rad8 3.b4 Nh5 4.d5
  ²  (0.67)   Depth: 7/23   00:00:06  518kN
1.Bh4 Qb6 2.Bb3 c5 3.d5 Nd4 4.Nxd4 cxd4
  ±  (0.73)   Depth: 8/25   00:00:14  1144kN
1.Bh4 b5 2.b3 Qd7 3.a4 bxc4 4.bxc4 Rad8
  ±  (0.77)   Depth: 9/27   00:00:47  3391kN
1.Bh4 Qd7 2.Rad1 Rad8 3.Qe3 Qd6 4.Bg3 Qb4 5.Bb3 c5 6.a3
  ±  (0.72)   Depth: 10/29   00:02:42  11568kN
1.Bh4 Nh5 2.d5 cxd5 3.Ba4 dxc4 4.Bxe8 Qxd2 5.Nxd2 Bxb2 6.Bd7 Bxa1 7.Rxa1 c3
8.Bxe6 fxe6
  ±  (0.73)   Depth: 11/30   00:10:56  46353kN

(Columbia 24.07.2001)

Now after Rxe6:

Hiarcs 7.32
r2qr1k1/pp2ppbp/2p1Rnp1/6B1/2PP4/5N1P/PPBQ1PP1/R5K1 b - - 0 1

Analysis by Hiarcs 7.32:

1...fxe6
  ³  (-0.48)   Depth: 1   00:00:00
1...fxe6 2.Ne5
  =  (-0.02)   Depth: 2/4   00:00:00
1...fxe6 2.Ne5
  =  (-0.02)   Depth: 2/5   00:00:00
1...fxe6 2.Ne5 Qd6
  =  (-0.12)   Depth: 3/7   00:00:00
1...fxe6 2.Re1 Qb6 3.Bb3 c5 4.dxc5 Qxc5 5.Rxe6
  =  (-0.08)   Depth: 4/7   00:00:00
1...fxe6 2.Qe3 Qd7 3.Re1 Rad8
  =  (0.02)   Depth: 5/13   00:00:00  8kN
1...fxe6 2.Re1 Qb6 3.b4 Rad8 4.Rxe6
  =  (0.13)   Depth: 6/18   00:00:00  36kN
1...fxe6 2.Qe3 Qd7 3.Re1 Rad8 4.b4 Qd6 5.Qxe6+ Qxe6 6.Rxe6
  =  (-0.02)   Depth: 7/18   00:00:01  131kN
1...fxe6 2.Qe3 Qd7 3.Re1 Rad8 4.b4 Qd6 5.Qxe6+ Qxe6 6.Rxe6
  =  (-0.02)   Depth: 8/25   00:00:09  698kN
1...fxe6 2.Re1 Qd7 3.Qe2 Rad8 4.b4 Nh5
  =  (-0.05)   Depth: 9/27   00:00:38  2777kN
1...fxe6 2.Re1 Qd7 3.Qe2 Rad8 4.Qe3 Nh5 5.g4 Nf6 6.Bf4 Kh8
  =  (-0.09)   Depth: 10/29   00:03:00  12833kN
1...fxe6 2.Qe3 Qd7 3.Ne5 Qc7 4.Nd3 e5 5.dxe5 Nd7 6.e6 Nf6 7.Bf4 Qa5
  =  (0.06)   Depth: 11/30   00:10:12  45264kN

(Columbia 24.07.2001)

As you can see, Hiarcs likes Bh4 better by +0.73 to +0.06 for Rxe6.

Chuck



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.