Author: chandler yergin
Date: 08:22:48 01/30/06
Go up one level in this thread
On January 29, 2006 at 19:32:02, Uri Blass wrote:
>On January 29, 2006 at 19:06:13, stuart taylor wrote:
>
>>On January 29, 2006 at 16:33:51, Terry McCracken wrote:
>>
>>>On January 29, 2006 at 15:45:52, George Tsavdaris wrote:
>>>
>>>>
>>>> In the following position you must avoid 13...dxc4?? that loses by a rather
>>>>deep combination.
>>>> The reason i post this is because i don't see often Rybka fail to see something
>>>>and that makes me feel good and superior to it, since i played as white this
>>>>game and won....:-) Now back to reality: Rybka beta 13b considers 13...dxc4 as a
>>>>good move for black and gives -0.95 after 13 plies. It also can't find my
>>>>excellent move 14.d5!! in a reasonable time (actually because i'm too lazy i
>>>>waited only 2:30 minutes on AMD3500+).
>>>>
>>>> White (I) has set up a good trap in this position to black. Rybka can't see
>>>>this.... Any estimation from its author of when it will be able to avoid
>>>>13...dxc4 and see 14.d5!!.....?
>>>>
>>>>[D]rn1qk2r/pp3ppp/4b3/3p4/2PP1p2/B4N2/P3Q1PP/R4RK1 b kq - 0 13
>>>>
>>>>[Event "?????"]
>>>>[Site "????????"]
>>>>[Date "????.??.??"]
>>>>[Round "?"]
>>>>[White "????"]
>>>>[Black "????"]
>>>>[Result "*"]
>>>>[SetUp "1"]
>>>>[FEN "rn1qk2r/pp3ppp/4b3/3p4/2PP1p2/B4N2/P3Q1PP/R4RK1 b kq - 0 13"]
>>>>[PlyCount "34"]
>>>>[EventDate "????.??.??"]
>>>>
>>>>13... dxc4 $2 (13... Nc6) 14. d5 $3 Qxd5 15. Rad1 Qa5 (15... Qb5 16. Nd4 Qb6
>>>>17. Kh1 Nd7 18. Rxf4 O-O-O 19. Nxe6 Qxe6 20. Rxc4+ Kb8 21. Qxe6 fxe6 22. Rcd4
>>>>b5 23. Rxd7 Kc8 $18 {And now white is clearly winning.... There are 2 roads:
>>>>24.Rxd8 Kxd8 25.Rxd8 Kxd8 with a Bishop against 2 Pawns at the endgame where
>>>>white should be able to convert it to a win or 24.Rd2 for example with a
>>>>Bishop more for 2 Pawns....}) 16. Nd4 Qxa3 (16... Qb6 17. Kh1 Nd7 18. Rxf4
>>>>O-O-O 19. Nxe6 Qxe6 20. Rxc4+ Kb8 21. Qxe6 fxe6 22. Rcd4 b5 23. Rxd7 Kc8 $18 {
>>>>And now white is clearly winning.... There are 2 roads: 24.Rxd8 Kxd8 25.Rxd8
>>>>Kxd8 with a Bishop against 2 Pawns at the endgame where white should be able
>>>>to convert it to a win or 24.Rd2 for example with a Bishop more for 2 Pawns....
>>>>}) 17. Nxe6 Qe3+ 18. Qxe3 fxe3 19. Nc7+ Ke7 20. Rfe1 Rc8 21. Rxe3+ Kf8 22. Nxa8
>>>>Nc6 23. Rd7 Rxa8 24. Rxb7 f6 (24... Rc8 25. Re4 Rd8 26. Rc7 Nd4 27. Ree7 Ne6
>>>>28. Rxf7+ Kg8 29. Rfd7 Rxd7 30. Rxd7 $18) 25. Rc7 Ne5 26. Rg3 g6 27. Rxh7 Kg8
>>>>28. Rc7 a5 29. Rh3 Rd8 30. Rhh7 $18 *
>>>
>>>Excellent Game! I doubt any program could see 13..dxc4 as an error, the win is
>>>simply too far away!
Game playing and analysis are Apples & Oranges.
In a Game the Program has limited time to move, it makes the 'best/ move found
at the time. It does this every move.
Rybka has slow kN/s, it's not as fast as other Engines and only Scrolls the main
line, so you cannot see othe Candidate moves and how it evaluates them.
Example:
New game,
[D]rn1qk2r/pp3ppp/4b3/8/2pP1p2/B4N2/P3Q1PP/R4RK1 w kq - 0 1
Analysis by Rybka 1.0 Beta 32-bit:
2.Qe5 Nd7 3.Qxf4 Qa5 4.Bd6 0-0-0 5.Ng5 Nf6 6.Be7
ยต (-0.91) Depth: 13 00:00:45 2235kN
(, 30.01.2006)
>>>
>>>Congratulations!
>>>
>>>Terry
>>
>>If it can see a mate in 36, then can't we expect it to find a forced win in 17
>>or so?
>>S.Taylor
>
>The mate in 36 is based on a lot of check single reply that trigger extensions.
>
>win in 17 when you need to play moves that are not extended can be harder for it
>to find.
>
>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.