Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: crafty at the internet vs diep

Author: Peter Fendrich

Date: 07:23:26 05/14/98

Go up one level in this thread


On May 14, 1998 at 08:01:55, Vincent Diepeveen wrote:

>
>On May 13, 1998 at 13:54:13, Robert Hyatt wrote:
>
>>On May 13, 1998 at 11:55:37, Amir Ban wrote:
>>
>>>On May 13, 1998 at 08:00:40, Vincent Diepeveen wrote:
>>>
>>>
>>>>Big surprise was that Diep played so well in blitz in Paris,
>>>
>>>[snip]
>>>
>>>>Against Junior diep had a totally won
>>>>position, but then diep searched 5 seconds for a move, played it and
>>>>lost. at second 6 it would have gotten a fail low. 6 ply ain't enough to
>>>>see tactically everything. But in that position 6 ply would have been
>>>>enough
>>>>to win it.
>>>
>>>
>>>Vincent, I'm real busy now. Why do I have to waste time to correct your
>>>statements ?
>>>
>>>The Paris game Diep-Junior is below. Diep plays 32.Qc1 and loses
>>>immediately. It has better moves (say R1a5 or Rxb7), but this gives
>>>white equality at best. I'll take black here anytime.
>>>
>>>Pity Qc1 didn't work out. It was a real crushing move ! I'm a bit
>>>puzzled how you managed to play it. I looked what level I need to find
>>>32...Nd3 and I find it in depth 1. Perhaps I can find it in quiescence
>>>only. If you can't avoid such moves in 5 seconds, I suggest a major
>>>redesign of your program.
>>>
>>>
>>>[Event "?"]
>>>[Site "?"]
>>>[Date "?"]
>>>[Round "5"]
>>>[White "Diep"]
>>>[Black "Junior"]
>>>[Result "0-1"]
>>>[ECO "?"]
>>>
>>>1. d4 d5 2. a3 Bf5 3. Bf4 e6 4. e3 Bd6 5. Ne2 Ne7 6. Nd2 Nbc6
>>>7. c4 O-O 8. Bg3 a5 9. c5 Bxg3 10. hxg3 Qd7 11. Nf3 b6 12. Rc1 Rfb8
>>>13. Qd2 b5 14. Nf4 a4 15. Bd3 h6 16. Bxf5 exf5 17. Qe2 b4 18. axb4 Rxb4
>>>19. Ra1 Qc8 20. Nd3 Rbb8 21. Ra3 Qe8 22. Qc2 Nb4 23. Nxb4 Rxb4 24. O-O
>>>Qb5
>>>25. Rb1 Rb8 26. Qd1 Nc6 27. Nh4 g6 28. Rba1 Rxb2 29. Rxa4 Kh7 30. Ra6
>>>Nb4
>>>31. Ra7 Rb7 32. Qc1 Nd3 33. Rxb7 Qxb7 34. c6 Qb6 35. Qf1 Nxf2 36. Qe1
>>>Ng4
>>>37. Qc1 Qb3 38. Qe1 Nxe3 39. Kh1 Qd3 40. Nf3 Rxg2 41. Qxe3 Qxe3 42. Kxg2
>>>Qe2
>>>43. Kh3 Qxf3 44. Kh2 Qf2 45. Kh1 Qxg3 46. Rf1 f4 47. Ra1 Qf3 48. Kg1 Qe3
>>>49. Kh1 Qe4 50. Kh2 Qe2 51. Kh1 f3 0-1
>>>
>>>
>>>Amir
>>
>>
>>
>>I agree.  My "broken quiescence search" (Vincent's words, not mine)
>>finds
>>this as follows, after white's move 32:
>>              depth   time  score   variation (1)
>>                1     0.00   0.24   32. ... Nd3
>>                1->   0.00   0.24   32. ... Nd3
>>                2     0.00   0.44   32. ... Nd3 33. Rxb7 Qxb7
>>                2     0.00   0.47   32. ... Rxa7 33. Qxb2 Rxa1+ 34. Qxa1
>>                2->   0.01   0.47   32. ... Rxa7 33. Qxb2 Rxa1+ 34. Qxa1
>>                3     0.04   0.68   32. ... Rxa7 33. Rxa7 Qe2 34. Qf1
>>                3->   0.08   0.68   32. ... Rxa7 33. Rxa7 Qe2 34. Qf1
>>                4     0.09   0.44   32. ... Rxa7 33. Qxb2 Rxa1+ 34. Qxa1
>>                                    Kh8 35. Nf3
>>                4->   0.13   0.44   32. ... Rxa7 33. Qxb2 Rxa1+ 34. Qxa1
>>                                    Kh8 35. Nf3
>>                5     0.21     --   32. ... Rxa7
>>                5     0.25   0.14   32. ... Rxa7 33. Rxa7 Qe2 34. Qf1
>>Qxf1+
>>                                    35. Kxf1 Rb1+ 36. Ke2 Rb2+ 37. Kf3
>>                5     0.29     ++   32. ... Nd3!!
>>                5     0.30   1.64   32. ... Nd3 33. Rxb7 Qxb7 34. Qf1
>
>You need 5 ply to see Nd3 winning.
>Diep searched 5 ply starting with 6 in Paris.
>
>Nowadays it gets it within 2 seconds, but that's of no use.
>The most important blitz contest last year Junior won against Diep.
>
>It's interesting to see how Diep can get with 5 ply searches, sometimes
>6 ply searches such a position against Junior, getting way deeper.
>
>The depth of Junior always amazes me, just like genius depth (although
>genius does only print out its positional depth, it seems to see 2 ply
>more).
>
>>Nxf2
>>                                    35. Nf3
>>                5->   0.40   1.64   32. ... Nd3 33. Rxb7 Qxb7 34. Qf1
>>Nxf2
>>                                    35. Nf3
>>                6     0.47   1.77   32. ... Nd3 33. Qc3 Nxf2 34. Nf3
>>Rxa7
>>                                    35. Rxa7 Ne4
>>                6->   0.61   1.77   32. ... Nd3 33. Qc3 Nxf2 34. Nf3
>>Rxa7
>>                                    35. Rxa7 Ne4
>>                7     0.74   1.68   32. ... Nd3 33. Rxb7 Qxb7 34. Qf1
>>Nxf2
>>                                    35. Nf3 Ne4 36. Kh2
>>                7->   0.94   1.68   32. ... Nd3 33. Rxb7 Qxb7 34. Qf1
>>Nxf2
>>                                    35. Nf3 Ne4 36. Kh2
>>                8     1.15   1.82   32. ... Nd3 33. Rxb7 Qxb7 34. Qf1
>>Nxf2
>>                                    35. Nf3 Ne4 36. Kh2 Qc6
>>                8->   1.44   1.82   32. ... Nd3 33. Rxb7 Qxb7 34. Qf1
>>Nxf2
>>                                    35. Nf3 Ne4 36. Kh2 Qc6
>>                9     1.80   1.84   32. ... Nd3 33. Rxb7 Qxb7 34. Qf1
>>Nxf2
>>                                    35. Nf3 Qb3 36. Qc1 c6 37. Qe1
>>                9->   2.42   1.84   32. ... Nd3 33. Rxb7 Qxb7 34. Qf1
>>Nxf2
>>                                    35. Nf3 Qb3 36. Qc1 c6 37. Qe1
>>               10     3.20     ++   32. ... Nd3!!
>>               10->   4.89   2.13   32. ... Nd3 33. Rxb7 Qxb7 34. Qf1
>>Nxf2
>>                                    35. Nf3 Qb3 36. Qc1 c6 37. Qe1
>>               11     7.09   2.33   32. ... Nd3 33. Rxb7 Qxb7 34. c6 Qb3
>>                                    35. Qf1 Nxf2 36. Qe1 Qb5 37. Qc1 Qe2
>>                                    38. Nf3
>>               11->  10.08   2.33   32. ... Nd3 33. Rxb7 Qxb7 34. c6 Qb3
>>                                    35. Qf1 Nxf2 36. Qe1 Qb5 37. Qc1 Qe2
>>                                    38. Nf3
>>
>>
>>So it looks good at 1 ply, but turns *real* good after only .29 seconds.
>>It doesn't take much to find that.  If you don't in 5 seconds, something
>>is seriously wrong...

So what you're trying to say is that if your program was only a little
bit better it might as well won the game?
No problem with me...
Peter



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.