Author: Christophe Theron
Date: 05:36:35 11/10/01
Go up one level in this thread
On November 10, 2001 at 03:19:27, José Carlos wrote:
>On November 10, 2001 at 02:29:10, Christophe Theron wrote:
>
>>On November 10, 2001 at 02:01:07, Uri Blass wrote:
>>
>>>On November 09, 2001 at 23:26:30, Christophe Theron wrote:
>>>
>>>>On November 09, 2001 at 20:28:45, Peter Berger wrote:
>>>>
>>>>>On November 09, 2001 at 19:56:34, Thomas Mayer wrote:
>>>>>
>>>>>Yet again , a late night post :=)
>>>>>
>>>>>>Hi,
>>>>>>
>>>>>>in winboard forum one has posted some games of Palm Tiger against not so strong
>>>>>>winboard engines.
>>>>>
>>>>>"One" answers :)
>>>>>
>>>>>>Next is interesting:
>>>>>>
>>>>>>[Event "Palm Tiger vs. WinBoard"]
>>>>>>[Site "Athlon 1333 vs. Palm Vx"]
>>>>>>[Date "2001.11.08"]
>>>>>>[Round "-"]
>>>>>>[White "Chess Tiger 14.6"]
>>>>>>[Black "Grizzly 1.31"]
>>>>>>[Result "0-1"]
>>>>>>[TimeControl "900"]
>>>>>>
>>>>>>1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Bxc6 dxc6 5. d4 exd4 6. Qxd4 Bg4 7. Qxd8+
>>>>>>Rxd8 8. Nc3 Bb4 9. Bg5 Bxc3+ 10. bxc3 f6 11. Bf4 Bxf3 12. gxf3 Rd7 13. h4
>>>>>>Ne7 14. h5 b6 15. Rd1 Nc8 16. Rxd7 Kxd7 17. Rg1 Rg8 18. Ke2 Nd6 19. Bh6 g5
>>>>>>20. hxg6 hxg6 21. Bf4 g5 22. Kd3 Rd8 23. Bh2 Nb5 24. f4 Kc8+ 25. Ke3 gxf4+
>>>>>>26. Bxf4 Nxc3 27. Rg7 Rd7 28. Rg8+ Kb7 29. f3 Nxa2 30. Rf8 Nb4 31. c4 Nc2+
>>>>>>32. Ke2 f5 33. exf5 b5 34. cxb5 cxb5 35. f6 Nd4+ 36. Ke3 Ne6 37. Re8 Nxf4
>>>>>>38. Kxf4 b4 39. Re4 a5 40. Re5 Kb6 41. Re7 Rd8 42. f7 Rf8 43. Kf5 b3 44.
>>>>>>Kf6 b2 45. Re1 a4 46. Ke7 Rxf7+ 47. Kxf7 a3 48. Rb1 c5 49. f4 c4 50. f5 c3
>>>>>>51. Re1 c2 52. Re6+ Kc5 53. Re5+ Kd4 54. Ra5 c1=Q 55. Rxa3 b1=Q 56. Ra4+
>>>>>>Ke3 57. Ra7 Qxf5+ 58. Kg7 Qc3+ 59. Kg8 Qcc8+ 60. Kg7 Qcf8#
>>>>>>{Black mates} 0-1
>>>>>>
>>>>>>[Event "Palm Tiger vs. WinBoard"]
>>>>>>[Site "Athlon 1333 vs. Palm Vx"]
>>>>>>[Date "2001.11.08"]
>>>>>>[Round "-"]
>>>>>>[White "Grizzly 1.31"]
>>>>>>[Black "Chess Tiger 14.6"]
>>>>>>[Result "1-0"]
>>>>>>[TimeControl "900"]
>>>>>>
>>>>>>1. f4 d5 2. Nf3 Bg4 3. g3 Bxf3 4. exf3 e6 5. Ke2 d4 6. Bg2 d3+ 7. cxd3 Nc6
>>>>>>8. Re1 Qd7 9. Kf1 Nb4 10. d4 Nd3 11. Re3 Qxd4 12. Qc2 Nxc1 13. Qxc1 Nf6 14.
>>>>>>Nc3 O-O-O 15. Nb5 Qd7 16. a4 a6 17. Nxc7 Kb8 18. Nxa6+ bxa6 19. Qc4 Qc8 20.
>>>>>>Rc1 Qxc4+ 21. Rxc4 Nd5 22. Rb3+ Ka7 23. f5 Bd6 24. f4 Rc8 25. Rxc8 Rxc8 26.
>>>>>>Rd3 Rc1+ 27. Ke2 Be7 28. fxe6 fxe6 29. f5 Nc7 30. Rd7 Bb4 31. fxe6 Rc5 32.
>>>>>>d4 Rc2+ 33. Kd3 Rd2+ 34. Ke3 Kb6 35. e7 Rxg2 36. Rxc7 Bxe7 37. Rxe7 Rxb2
>>>>>>38. h4 Rb3+ 39. Kf2 Ra3 40. Rxg7 Rxa4 41. Ke3 h5 42. Rg5 Ra3+ 43. Kf4 Rd3
>>>>>>44. d5 a5 45. Ke4 Rb3 46. d6 a4 47. Rd5 Rxg3 48. d7 Rg4+ 49. Kf5 Rg8 50.
>>>>>>d8=Q+ Rxd8 51. Rxd8 Kc7 52. Ra8 Kc6 53. Rxa4 Kb5 54. Ra8 Kc6 55. Rh8 Kd5
>>>>>>56. Rxh5 Kd4 57. Rg5 Kd5 58. h5 Kc4 59. h6 Kb3 60. h7 Ka2 61. Rg3 Ka1 62.
>>>>>>h8=Q+ Kb1 63. Rg2 Kc1 64. Qa1#
>>>>>>{White mates} 1-0
>>>>>>
>>>>>>[Event "Palm Tiger vs. WinBoard"]
>>>>>>[Site "Athlon 1333 vs. Palm Vx"]
>>>>>>[Date "2001.11.08"]
>>>>>>[Round "-"]
>>>>>>[White "Chess Tiger 14.6"]
>>>>>>[Black "Grizzly 1.31"]
>>>>>>[Result "0-1"]
>>>>>>[TimeControl "900"]
>>>>>>
>>>>>>1. d4 d5 2. c4 e6 3. Nf3 dxc4 4. e3 b5 5. Nc3 Bb4 6. a3 Bxc3+ 7. bxc3 Nf6
>>>>>>8. Ne5 O-O 9. a4 c6 10. Qf3 Qd5 11. Ba3 Qxf3 12. gxf3 Re8 13. Bd6 a6 14.
>>>>>>Bxb8 Rxb8 15. Nxc6 Rb6 16. axb5 axb5 17. Nb4 Nd5 18. Nxd5 exd5 19. f4 b4
>>>>>>20. cxb4 Rxb4 21. Bg2 Bb7 22. O-O Ra8 23. Rab1 Rxb1 24. Rxb1 c3 25. Rxb7 c2
>>>>>>26. Rc7 Ra1+ 27. Bf1 c1=Q 28. Rxc1 Rxc1 29. Kg2 f5 30. Bb5 Kf7 31. Kf3 Ke6
>>>>>>32. Be8 Rh1 33. Kg2 Rb1 34. Kf3 g6 35. h3 Rb4 36. Bc6 Kd6 37. Be8 Rb8 38.
>>>>>>Ba4 Rb1 39. Be8 Rh1 40. Kg2 Rc1 41. Kg3 Ke7 42. Bb5 Rg1+ 43. Kf3 Ke6 44.
>>>>>>Bc6 Rh1 45. Kg3 Rc1 46. Bb7 Rg1+ 47. Kf3 Kd6 48. Ba6 Rh1 49. Kg3 Ra1 50.
>>>>>>Bb7 h6 51. Kf3 Ra5 52. Bc8 Ra7 53. Kg3 Kc6 54. Be6 Rc7 55. e4 dxe4 56. Kh4
>>>>>>Kd6 57. Bb3 Ke7 58. Bg8 Rc3 59. d5 Kf6 60. f3 Rxf3 61. Be6 g5+ 62. fxg5+
>>>>>>hxg5+ 63. Kh5 Rxh3#
>>>>>>{Black mates} 0-1
>>>>>>
>>>>>>In most rating lists, Grizzly is around 1800-1900 - this new version is a little
>>>>>>bit stronger, but I doubt that it is over 2000. Does this mean that Chess Tiger
>>>>>>on Palm is also clearly under 2000 ?
>>>>>
>>>>>Both, the second and the third game were quite interesting and very close.
>>>>>Grizzly is no weaky - I was very much impressed by the Tiger performance here !
>>>>>Still Grizzly won ; is Zephyr weak ? Or Blikskottel ? What was their score ? I
>>>>>wanted to have a little fun ; beating the Tiger on the 1000* hardware shows that
>>>>>there has been some "magic" barrier passed - only Grizzly succeeded.
>>>>>
>>>>>Or is this another point where
>>>>>>engine-engine matches are very different to engine - human matches ? Anyway, I
>>>>>>would like to see some more matches of Chess Tiger on Palm against the lower
>>>>>>edge of the winboard-engines... And besides that, frequently here are some
>>>>>>postings of games ChessTiger on Palm against some well known dedicated chess
>>>>>>computers which are listed on the SSDF-list. ChessTiger wins many games, does
>>>>>>that mean that those old irons are still overestimated on the SSDF-list. Or
>>>>>>should we finally stop to compare SSDF-ELO with human ELO ? (Which I say for a
>>>>>>long time now - there is no real comparrison... The 100 ELO-penalty for the hole
>>>>>>list was somehow irregular because the old iron often have showed there strength
>>>>>>against humans, so why decrease them 100 ELOs ?)
>>>>>>
>>>>>>Greets, Thomas
>>>>>>
>>>>>>P.S.: I think a solution for the SSDF would be a completely different rating.
>>>>>>Just stop to make them compareable with human rating and set a starting level
>>>>>>for the lowest listed engine of e.g. 0 or e.g. 10000 so just the numbers are
>>>>>>very different to FIDE-ELO and noone would compare anymore.
>>>>>
>>>>>Well - it's too early for any conclusion IMHO . My personal guess is that the
>>>>>Palm Tiger is about the same strength as the Novag Diamond/Sapphire. When it is
>>>>>about the WinBoard engines being able to beat the Tiger even if on some insane
>>>>>Athlon 1300++ against some miserable Palm Vx - seems to be some _real_ challenge
>>>>>still , huh ?
>>>>>
>>>>>Regards,
>>>>>pete
>>>>
>>>>
>>>>
>>>>I think another real challenge is to write a 1800 elo program for the Athlon
>>>>1333.
>>>>
>>>>I think that by just using plain vanilla brute force alpha beta without
>>>>transposition table nor null move and a static piece square table you are
>>>>already above 1800 elo. Easily.
>>>>
>>>>Actually the Sargon III program does just that, and on an Athlon it must be well
>>>>over 2100 elo (maybe 2200).
>>>
>>>
>>>I doubt it
>>>Do you say that SargonIII does not evaluate passed pawns?(you cannot evaluate
>>>passed pawns only by piece square table evaluation).
>>>
>>>
>>>I believe that if you use only alpha beta without evaluation that has only
>>>material and piece square table and use no qsearch then you can get program that
>>>is even weaker than 1800 on Athlon1333.
>>>
>>>Uri
>>
>>
>>
>>Maybe. But just following the basic published algorithms of alphabeta+QSearch
>>brings you well over 1800 in no time.
>
> Don't forget there's some point in the life of a programmer when (s)he begins.
>The first time you read about alfabeta and qsearch and you decide to write a
>small chess program, you write something that won't play well in any hardware,
>because of it's braching factor, simple eval, bugs...
> Nowadays, programmers share their engines in a very early stage of
>developement. This is great, because weak human players can find a variety of
>weak chess programs to play with.
> I don't know haw strong is Grizzly, but shouldn't be surpresed that there're a
>lot of programs that will perform below 1800 fide in a fast machine. And those
>programmers are not stupid at all, they're learning.
Naturally they are not stupid. I went through this process myself not so long
ago.
But even at the time I was using a 386, my program has quickly been able to beat
me.
> Also, most amateur programmers have very little time to work on their engines.
>In my case, about a couple of hours per week.
The fact that the engine is able to run flawlessly under Winboard for long
enough to get a rating already shows that a good deal of debugging has been done
on it.
>>I don't think Sargon III had any second order evaluation of passed pawns.
>>
>>And I know you can go easily above 1800 elo with a piece square table only
>>evaluation. I know it sounds surprising, but that tells something about the
>>"essence" of chess.
>
> You're thinking from your experienced point of view. But there was a time when
>you were a beginner too :)
And I remember this time very well. I was using 286s and 386s at that time.
On a hardware that is between 500 and 1000 times faster, any program is approx.
200 elo points stronger...
Christophe
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.