Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: a test position for chess programs

Author: Robert Hyatt

Date: 10:05:35 06/02/00

Go up one level in this thread


On June 02, 2000 at 12:51:00, blass uri wrote:

>On June 02, 2000 at 10:39:51, Robert Hyatt wrote:
>
>>On June 02, 2000 at 09:56:47, blass uri wrote:
>>
>>>On June 02, 2000 at 08:32:31, blass uri wrote:
>>>
>>>>This position is from the game tal2.03-Junior6a
>>>>
>>>>[D]r2qr3/pb3pk1/np3b1N/2pp3Q/3P4/1PR1P1P1/P4PBP/5RK1 w - - 0 1
>>>>
>>>>1)How much time does your program need to find 1.g4?
>>>>
>>>>2)How much time does your program need to avoid 1.g4?
>>>>
>>>>1.g4 is probably not the best move because of the line 1.g4 Rh8 2.Qxf7+ Kxh6
>>>>3.f4 Rg8 4.Qxb7 Nb4 5.g5+ Bxg5 6.fxg5+ Qxg5 7.Rf2 Raf8 8.Qd7 Rxf2
>>>>
>>>>Uri
>>>
>>>The best move may be e4 or Nxf7.
>>>
>>>If it g4 then white has not more than a draw and the question is how much time
>>>does your program need to see not more than a draw evaluation with the move g4
>>>or another move.
>>>
>>>Uri
>>
>>
>>Maybe "crafty" is simply dense here:
>>
>>         (2)    5->   0.35  -0.39   1. Nxf7 Qe7 2. Nh6 Qe6 3. Nf5+ Kg8
>>                6     0.42  -0.26   1. Nxf7 Qe7 2. Nh6 Qe6 3. Nf5+ Kg8
>>                                    4. h3
>>                6->   0.64  -0.26   1. Nxf7 Qe7 2. Nh6 Qe6 3. Nf5+ Kg8
>>                                    4. h3
>>                7     0.90  -0.07   1. Nxf7 Qd7 2. Nh6 Rf8 3. Ng4 Be7 4.
>>                                    Ne5
>>                7->   1.21  -0.07   1. Nxf7 Qd7 2. Nh6 Rf8 3. Ng4 Be7 4.
>>                                    Ne5
>>                8     1.54  -0.16   1. Nxf7 Qd7 2. Nh6 cxd4 3. exd4 Kf8
>>                                    4. Nf5 Rac8 5. Rfc1
>>         (4)    8->   2.56  -0.16   1. Nxf7 Qd7 2. Nh6 cxd4 3. exd4 Kf8
>>                                    4. Nf5 Rac8 5. Rfc1
>>         (3)    9     3.58  -0.15   1. Nxf7 Qd7 2. Nh6 cxd4 3. exd4 b5
>>                                    4. Nf5+ Kg8 5. Qg4+ Kh8 6. Qh5+ Qh7
>>                                    7. Qxh7+ Kxh7 8. h3
>>         (4)    9->   7.45  -0.15   1. Nxf7 Qd7 2. Nh6 cxd4 3. exd4 b5
>>                                    4. Nf5+ Kg8 5. Qg4+ Kh8 6. Qh5+ Qh7
>>                                    7. Qxh7+ Kxh7 8. h3
>>         (3)   10    11.57  -0.07   1. Nxf7 Qd7 2. Nh6 Rad8 3. Ng4 Rh8
>>                                    4. Nxf6 Rxh5 5. Nxh5+ Kf8 6. Nf6
>>         (4)   10->  29.00  -0.07   1. Nxf7 Qd7 2. Nh6 Rad8 3. Ng4 Rh8
>>                                    4. Nxf6 Rxh5 5. Nxh5+ Kf8 6. Nf6
>>         (3)   11    38.04  -0.02   1. Nxf7 Qd7 2. Nh6 Rad8 3. Bh3 Qc7
>>                                    4. Bf5 Kf8 5. Bg6 Re6 6. Qf5
>>         (4)   11->   2:44  -0.02   1. Nxf7 Qd7 2. Nh6 Rad8 3. Bh3 Qc7
>>                                    4. Bf5 Kf8 5. Bg6 Re6 6. Qf5
>>         (3)   12     3:10  -0.02   1. Nxf7 Qd7 2. Nh6 Rad8 3. h3 cxd4
>>                                    4. Nf5+ Kf8 5. Nxd4 Bxd4 6. exd4 Kg8
>>                                    7. Qg5+ Qg7 8. Qf4
>>               12     9:48  38/42   1. Qxd5
>>
>>
>>I will let it run longer, but at reasonable time controls, Nxf7 is all it would
>>play here.  It is neither happy nor worried about the position...
>
>Crafty17.10(engine for Hiarcs7.32) on pIII450 found g4 after almost 40 minutes
>at depth 12 with a small positive score(0.13)
>At depth 13 after more than anhour g4 failed low with evaluation of -0.48 so I
>guess that crafty is going to change its mind at this depth.
>
>Uri

It didn't change thru depth=14, where I stopped it.  No idea why 17.10 would,
but 17.11 would not, except that I am using an SMP search and this often
produces different results than a single-cpu program.




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.