Comet B22 stronger !

[Thorsten Czub]

Geschrieben von: / Posted by: Thorsten Czub at 29 May 2000 12:13:12:
Mir deucht der neue Comet B22 ist dem Uli aber gelungen ))
Der hat den Wchess auseinandergenommen, unglaublich...
It looks the new comet is really a step forward.
well done uli !

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 29 May 2000 13:52:32:
Als Antwort auf: / As an answer to: Comet B22 stronger ! geschrieben von: / posted by: Thorsten Czub at 29 May 2000 12:13:12:

Mir deucht der neue Comet B22 ist dem Uli aber gelungen ))
Der hat den Wchess auseinandergenommen, unglaublich...
It looks the new comet is really a step forward.
well done uli !

It's to early to tell if Comet B22 is better than B21, especially if you base your conclusion on one game alone. Here's a couple of games from my Super WB tournament against Phalanx (40 moves in 100 minutes, ponder off, 16Mb hash, PII-266MHz/64Mb):
[Event "SuperWB Tournament."]
[Site "MOGENS"]
[Date "2000.05.27"]
[Round "3.1"]
[White "Comet_B22"]
[Black "Phalanx_22c"]
[Result "0-1"]
[ECO "A67"]
[PlyCount "130"]
1. d4 Nf6 2. c4 e6 3. Nc3 c5 4. d5 exd5 5. cxd5 d6 6. e4 g6 7. f4 Bg7 8. Bb5+
Nfd7 9. a4 O-O 10. Nf3 Na6 11. O-O Nc7 12. Bd3 a6 13. Qe1 f5 14. exf5 gxf5 15.
Qd1 b6 16. Re1 Bb7 17. Qc2 Bxc3 18. bxc3 Bxd5 19. Bb2 Bxf3 20. gxf3 d5 21. Bxf5
Kh8 22. Bxh7 Rxf4 23. Kh1 Qh4 24. c4+ d4 25. Be4 Rg8 26. Rg1 Rff8 27. a5 bxa5
28. Qd2 Ne6 29. Bd5 Rxg1+ 30. Rxg1 Nf4 31. Rg4 Qf6 32. Bc1 Nxd5 33. Qg2 Rf7 34.
Rg8+ Kh7 35. Rg3 Rg7 36. Qh3+ Kg8 37. Qxd7 Qxf3+ 38. Rxf3 Rxd7 39. cxd5 c4 40.
Kg2 Rxd5 41. Rf6 c3 42. Rc6 a4 43. Kf3 Re5 44. Rc4 Re1 45. Ba3 d3 46. Rc8+ Kh7
47. Rc7+ Kh6 48. Rc6+ Kh5 49. Rc5+ Kh4 50. Rc4+ Kh3 51. Rc6 Rf1+ 52. Ke4 d2 53.
Rxc3+ Kh4 54. Be7+ Kh5 55. Rc5+ Kg6 56. Rc6+ Kf7 57. Rd6 d1=Q 58. Rxd1 Rxd1 59.
Bc5 Ra1 60. Kd3 a3 61. Kc2 a2 62. Kb3 Rc1 63. Be3 a1=Q 64. Bg1 Qc3+ 65. Ka4
65... Ra1# {Black mates} 0-1
[Event "SuperWB Tournament."]
[Site "MOGENS"]
[Date "2000.05.28"]
[Round "3.2"]
[White "Phalanx_22c"]
[Black "Comet_B22"]
[Result "1-0"]
[ECO "B22"]
[PlyCount "112"]
1. e4 c5 2. c3 d5 3. exd5 Qxd5 4. d4 Nc6 5. Nf3 e6 6. Bd3 cxd4 7. cxd4 Nf6 8.
O-O Be7 9. Nc3 Qd6 10. a3 O-O 11. Be3 a6 12. Qc2 Bd7 13. Bg5 g6 14. Rad1 Rfc8
15. Rfe1 Be8 16. Be4 Nxe4 17. Qxe4 Bxg5 18. Nxg5 h6 19. Nf3 Ne7 20. Ne5 Qb6 21.
Rd2 h5 22. Qf4 Nf5 23. Ne4 Kg7 24. Nc5 Rd8 25. Nc4 Qb5 26. Rc1 Rac8 27. Rc3 Kg8
28. Ne3 Bc6 29. h3 b6 30. a4 Qb4 31. Nxa6 Qxa4 32. Ra3 Qb5 33. Nc7 Qb4 34. Nxf5
exf5 35. Rc3 Qa4 36. Qe5 Rf8 37. Kh2 Bb7 38. Qf6 Qd7 39. Qxb6 Rb8 40. d5 Rfc8
41. f4 Kh7 42. Qc5 Kh6 43. b4 Ba8 44. b5 Bb7 45. b6 Rd8 46. Re3 Qa4 47. Rd4 Qa1
48. Qb4 Qf1 49. Qe7 Qf2 50. Qg5+ Kh7 51. Re7 Rf8 52. Ne6 h4 53. Rd3 Bxd5 54.
Rxd5 Qg3+ 55. Qxg3 hxg3+ 56. Kxg3 56... Kg8 {Black resigns} 1-0

Best wishes...
Mogens

[Thorsten Czub]

Geschrieben von: / Posted by: Thorsten Czub at 29 May 2000 23:43:43:
Als Antwort auf: / As an answer to: Re: Comet B22 stronger ! geschrieben von: / posted by: Mogens Larsen at 29 May 2000 13:52:32:

It's to early to tell if Comet B22 is better than B21, especially if you base your conclusion on one game alone. Here's a couple of games from my Super WB tournament against Phalanx (40 moves in 100 minutes, ponder off, 16Mb hash, PII-266MHz/64Mb):

I am evaluating what i have seen. and i have seen a good game with accurate
evaluation, activ play and thats enough for me.
maybe it is because your machines are slower and the computation-time you
used is faster. you also had no pb and less hash-tables.
together this could be a whole class difference.

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 30 May 2000 00:46:54:
Als Antwort auf: / As an answer to: Comet B22 stronger ! geschrieben von: / posted by: Thorsten Czub at 29 May 2000 23:43:43:

I am evaluating what i have seen. and i have seen a good game with accurate
evaluation, activ play and thats enough for me.
maybe it is because your machines are slower and the computation-time you
used is faster. you also had no pb and less hash-tables.
together this could be a whole class difference.

Well, I'm not going to question your observational skills. I'm sure your keen eye has detected the significant improvement that Comet has undergone by watching this single game. I wish I had that kind of perception. I guess some are born with it and some are not.
Probably so, but that goes for the other programs as well I would assume. Somehow I feel that making no judgement on the basis of two games, is better than making one on the basis of one game. This applies regardless of various parameters involved.
Best wishes...
Mogens

[Gábor Szõts]

Geschrieben von: / Posted by: Gábor Szõts at 30 May 2000 10:09:49:
Als Antwort auf: / As an answer to: Re: Comet B22 stronger ! geschrieben von: / posted by: Mogens Larsen at 29 May 2000 13:52:32:

Mir deucht der neue Comet B22 ist dem Uli aber gelungen ))
Der hat den Wchess auseinandergenommen, unglaublich...
It looks the new comet is really a step forward.
well done uli !

It's to early to tell if Comet B22 is better than B21, especially if you base your conclusion on one game alone. Here's a couple of games from my Super WB tournament against Phalanx (40 moves in 100 minutes, ponder off, 16Mb hash, PII-266MHz/64Mb):

Best wishes...
Mogens

I let Comet B.21 and B.22 play against each other a 20-game match under the following conditions:
Pentium 166, 32 MB RAM
G/5, 4 MB for hash
Result: 10-10 (8-8-4)
Nevertheless, Comet has been making quite a progress recently, and B.22 is clearly better than B.18, let alone B.12.

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 30 May 2000 11:20:33:
Als Antwort auf: / As an answer to: Re: Comet B22 stronger ! geschrieben von: / posted by: Gábor Szõts at 30 May 2000 10:09:49:

I let Comet B.21 and B.22 play against each other a 20-game match under the following conditions:
Pentium 166, 32 MB RAM
G/5, 4 MB for hash
Result: 10-10 (8-8-4)
Nevertheless, Comet has been making quite a progress recently, and B.22 is clearly better than B.18, let alone B.12.

Yes, I agree. B21 and B22 seem to be a lot better than the previous versions, which is excellent because I like the playing style of Comet.
Best wishes...
Mogens

[Thorsten Czub]

Geschrieben von: / Posted by: Thorsten Czub at 30 May 2000 12:59:47:
Als Antwort auf: / As an answer to: Re: Comet B22 stronger ! geschrieben von: / posted by: Mogens Larsen at 30 May 2000 00:46:54:

Probably so, but that goes for the other programs as well I would assume. Somehow I feel that making no judgement on the basis of two games, is better than making one on the basis of one game. This applies regardless of various parameters involved.
Best wishes...
Mogens

Making NO judgement about the quality of a coke
of the basis of drinking 2 glasses,
thats what YOU don't do.
or making a judgement about the quality of a wine
of the basis of drinking on glass when you drink
at least one glass each day since 1978...
thats what i do.
When i see something. i speak it out.
better than NOT saying anything.
i am used to finding out about playing strength.
let the program play against the version before is NO
way to find out about the playing-strengh. letting it
play against another program you don't know the playing strength either
is also no good way to find out the playing strength.

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 30 May 2000 13:30:43:
Als Antwort auf: / As an answer to: Comet B22 stronger ! geschrieben von: / posted by: Thorsten Czub at 30 May 2000 12:59:47:

Making NO judgement about the quality of a coke
of the basis of drinking 2 glasses,
thats what YOU don't do.
or making a judgement about the quality of a wine
of the basis of drinking on glass when you drink
at least one glass each day since 1978...
thats what i do.
When i see something. i speak it out.
better than NOT saying anything.
i am used to finding out about playing strength.
let the program play against the version before is NO
way to find out about the playing-strengh. letting it
play against another program you don't know the playing strength either
is also no good way to find out the playing strength.

If you're going to use analogies, at least try to use some that are appropriate.
I'm sure you are, but that's not the point.
Playing one game against a program of known strength still doesn't tell you anything fundamental, no matter how much wine you've been drinking. You didn't even post the game, so that others might attempt to validate your opinion. The headline of your thread is undocumented and that's basically it.
Best wishes...
Mogens

[Gábor Szõts]

Geschrieben von: / Posted by: Gábor Szõts at 30 May 2000 13:32:20:
Als Antwort auf: / As an answer to: Comet B22 stronger ! geschrieben von: / posted by: Thorsten Czub at 30 May 2000 12:59:47:

Probably so, but that goes for the other programs as well I would assume. Somehow I feel that making no judgement on the basis of two games, is better than making one on the basis of one game. This applies regardless of various parameters involved.
Best wishes...
Mogens

Making NO judgement about the quality of a coke
of the basis of drinking 2 glasses,
thats what YOU don't do.
or making a judgement about the quality of a wine
of the basis of drinking on glass when you drink
at least one glass each day since 1978...
thats what i do.
When i see something. i speak it out.
better than NOT saying anything.
i am used to finding out about playing strength.
let the program play against the version before is NO
way to find out about the playing-strengh. letting it
play against another program you don't know the playing strength either
is also no good way to find out the playing strength.

How interesting! I feel both of you are right.
For me that means that to assess the quality of a wine is equally hard as to assess the quality of a chess program...

[U.Türke]

Geschrieben von: / Posted by: U.Türke at 30 May 2000 15:24:47:
Als Antwort auf: / As an answer to: Comet B22 stronger ! geschrieben von: / posted by: Thorsten Czub at 30 May 2000 12:59:47:
First of all, thanks for your interest.
I think that you guys use two different methods in testing. IMO, both methods are reasonable and interesting.
1. You can proceed quantitatively, i.e. proceed a large probe of games and evaluate statistically. You will finally reliably deduce a kind of ELO number, representing a very reasonable estimate for the playing strength. In order to do this you even do not have to look at the games, only the results count. Advantage: you get a reliable estimate; Drawback: you don't know why the program is bad or good, i.e. we have no hint what to do better. This is the method done by SSDF, and it is unavoidable to get an ELO number.
2. You may carefully observe a game and find weak (or strong) points regarding the program's playing style. The advantage is that this method can lead to detailled hints. The drawback is that you cannot learn too much about the playing strength in terms of an ELO number. A program playing very impressive one game, may look like a fool in the next one.
So, I really appreciate what both of you are doing. I myself am using both approaches. I have time to observe a game I use approach 1. Sometimes I play matches at night and look next day what happened.
Thanks again and have a nice day,
Uli

[Thorsten Czub]

Geschrieben von: / Posted by: Thorsten Czub at 30 May 2000 19:38:34:
Als Antwort auf: / As an answer to: Re: Comet B22 stronger ! geschrieben von: / posted by: Mogens Larsen at 30 May 2000 13:30:43:

If you're going to use analogies, at least try to use some that are appropriate.
Playing one game against a program of known strength still doesn't tell you anything fundamental, no matter how much wine you've been drinking.
You didn't even post the game,
so that others might attempt to validate your opinion. The headline of your thread is undocumented and that's basically it.
Best wishes...
Mogens

there is no appropriate analogie. there is nobody who decides it
is appropriate other than me.
analogies cannot be measured.
wrong.
it says much. it gives you at least 2 x 40 test-positions
to observe the programs. and it does not only give you the
output of the 2 programs, also it shows you how the program
is capable to fight through the positions.
the games of my millennium-tournament are posted several places.
rgcc, often gambit soft, sometimes here too.

)) consuming is too easy isn't it. you want that the birds
fly directly into your mouth... ))

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 30 May 2000 20:01:25:
Als Antwort auf: / As an answer to: Comet B22 stronger ! geschrieben von: / posted by: Thorsten Czub at 30 May 2000 19:38:34:

there is no appropriate analogie. there is nobody who decides it
is appropriate other than me.
analogies cannot be measured.
wrong.
it says much. it gives you at least 2 x 40 test-positions
to observe the programs. and it does not only give you the
output of the 2 programs, also it shows you how the program
is capable to fight through the positions.
)) consuming is too easy isn't it. you want that the birds
fly directly into your mouth... ))

Well, analogies can be measured as either good or bad. Your analogy was bad.
Sounds reasonable, except the fact that the empiciral knowledge you claim to gain in one game isn't reproduceable in another, unless the observations are very elemental. Thereby you end up analysing more than one game anyway to make a sound judgement. Unless you are a GM player, I wouldn't trust your estimation of program strength based on one game.
Not entirely, I kinda like birds ). I just wanted to see your analysis or comments of the game. They had to be thorough to form a foundation for such a statement.
Best wishes...
Mogens

[Thorsten Czub]

Geschrieben von: / Posted by: Thorsten Czub at 30 May 2000 21:20:42:
Als Antwort auf: / As an answer to: Re: Comet B22 stronger ! geschrieben von: / posted by: Gábor Szõts at 30 May 2000 10:09:49:

I let Comet B.21 and B.22 play against each other a 20-game match under the following conditions:
Pentium 166, 32 MB RAM
G/5, 4 MB for hash
Result: 10-10 (8-8-4)
Nevertheless, Comet has been making quite a progress recently, and B.22 is clearly better than B.18, let alone B.12.

you get 10-10 and what is your conclusion: that there is NO
progress at all ?
yes. for sure.
but i see not much sense it letting comet play against comet.

[Thorsten Czub]

Geschrieben von: / Posted by: Thorsten Czub at 30 May 2000 21:25:19:
Als Antwort auf: / As an answer to: Maybe, maybe not. geschrieben von: / posted by: Mogens Larsen at 30 May 2000 20:01:25:

Well, analogies can be measured as either good or bad. Your analogy was bad.

wrong.

Sounds reasonable, except the fact that the empiciral knowledge you claim to gain in one game isn't reproduceable in another, unless the observations are very elemental.
Thereby you end up analysing more than one game anyway to make a sound judgement. Unless you are a GM player, I wouldn't trust your estimation of program strength based on one game.
Not entirely, I kinda like birds ). I just wanted to see your analysis or comments of the game. They had to be thorough to form a foundation for such a statement.
Best wishes...
Mogens

:-))) i think thats to shallow.
when you see 40 positions ? you get a lot of information.
also you do not only get information from WHAT you see but also from WHEN
(related to the other program that is also computing)
you don't have to.
mark, chris, christophe and john have done. so : maybe it works sometimes )
why do you need comments when you have the game ?
[Event "millennium-tournament 1999-2000"]
[Site "k6-400 40/120"]
[Date "2000.05.29"]
[Round "12"]
[White "Wchess2000"]
[Black "Comet B22"]
[Result "0-1"]
[TimeControl "40/7200"]
1. d4 Nf6 2. Nf3 e6 3. e3 c5 4. c4 cxd4 5. exd4 d5 6. Nc3 Bb4 7. Bd3 dxc4
8. Bxc4 O-O 9. O-O b6 10. Re1 Bb7 11. Bg5 h6 12. Bh4 Nc6 13. Rc1 Rc8 14.
Bb5 a6 15. Bd3 Re8 16. Bb1 Be7 17. Bg3 b5 18. a3 Qb6 19. Qd3 Red8 20. Ne4
Nxd4 21. Nxf6+ Bxf6 22. Nxd4 Rxc1 23. Qh7+ Kf8 24. Rxc1 Qxd4 25. Rf1 Qxb2
26. Bc7 Rd7 27. Qh8+ Ke7 28. Qb8 Qxa3 29. h3 Qc5 30. Qxb7 Rxc7 31. Qf3 a5
32. Rd1 a4 33. Qe4 a3 34. Qd3 Bd4 35. Kh1 Rd7 36. Ba2 b4 37. f4 g6 38. Qe4
Bf2 39. Rxd7+ Kxd7 40. Bc4 Bd4 41. Qc2 Ke7 42. g3 Qc6+ 43. Kh2 Qf3 44. Qg2
Qc3 45. Qa2 g5 46. Kg2 gxf4 47. gxf4 Qe3 48. Qc2 Qxf4 49. Qa2 Qe4+ 0-1

[Thorsten Czub]

Geschrieben von: / Posted by: Thorsten Czub at 30 May 2000 21:31:33:
Als Antwort auf: / As an answer to: Re: Comet B22 stronger ! geschrieben von: / posted by: U.Türke at 30 May 2000 15:24:47:

First of all, thanks for your interest.
I think that you guys use two different methods in testing. IMO, both methods are reasonable and interesting.
1. You can proceed quantitatively, i.e. proceed a large probe of games and evaluate statistically. You will finally reliably deduce a kind of ELO number, representing a very reasonable estimate for the playing strength. In order to do this you even do not have to look at the games, only the results count. Advantage: you get a reliable estimate; Drawback: you don't know why the program is bad or good, i.e. we have no hint what to do better. This is the method done by SSDF, and it is unavoidable to get an ELO number.
2. You may carefully observe a game and find weak (or strong) points regarding the program's playing style. The advantage is that this method can lead to detailled hints. The drawback is that you cannot learn too much about the playing strength in terms of an ELO number. A program playing very impressive one game, may look like a fool in the next one.
So, I really appreciate what both of you are doing. I myself am using both approaches. I have time to observe a game I use approach 1. Sometimes I play matches at night and look next day what happened.
Thanks again and have a nice day,
Uli

hi uli. thank you for your wise words.
do you remember when we watched the game comet - chess-tiger12 in paderborn.
i have forgotten WHEN it was. but i remember that the clou of the game
was, that tiger didn't care about the king-safety and instead concentrated
on pushing the pawn. we both watched the game and the main-lines of
the programs during the game. this was what i mean when i talk about
"observations" during the game. i remember another thing, that i knew
at this time (i did not told anybody at that time) tiger had NO
(in words: NO ~ NADA ~ NOTHING) king-safety terms at all.
so while we watched the game, i knew WHY tiger fought the game THIS
particular way. I could not tell you. because it was secret at this time
of the working process of christophe.
of course from time to time you need autoplayer games to prove that observations have been true and not random. i am having 2 machines to autoplay.
i always had. it 's not that i do not believe in matches. only: i learn much
more about the thing when watching the games live.
this makes those events so nice, because you can watch the programs live.
and discuss during the game.

[Dann Corbit]

Geschrieben von: / Posted by: Dann Corbit at 30 May 2000 22:00:22:
Als Antwort auf: / As an answer to: Maybe, maybe not. geschrieben von: / posted by: Thorsten Czub at 30 May 2000 21:25:19:
I think that a few games is definitely not enough to judge, *especially* by looking at the games. Sometimes, we see a computer make an absolutely _brilliant_ positional move or sacrifice. Unfortunately, it was not actually intended! It only fell into it.
I have seen the very best computer programs make prodigious blunders -- unthinkable C-level moves and outright gaffes. 99.9% of the time (however) they don't play stupidly. If I looked at two or three games that happened to have a horrible blunder (trapping your own piece, 3-fold repetition in the face of a clear win etc...) I might conclude that the program was terrible. On the other hand, if I saw two games with no mistakes and both of which contained a stunning and brilliant move, I might conclude that the program was a sheer genius.
Surely you have seen both sides from any computer program that has played at least 100 games against an opponent of approximately equal strength.
Do you know any program that never makes stupid mistakes -- ever?
Do you know any program that always makes brilliant moves, even positional?
Yes, a mathematical analysis lacks art and beauty. But it is a lot more stable than a touch-feely impression, even by a super GM.
IMO-YMMV.


My FTP site

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 30 May 2000 22:08:56:
Als Antwort auf: / As an answer to: Maybe, maybe not. geschrieben von: / posted by: Thorsten Czub at 30 May 2000 21:25:19:

:-))) i think thats to shallow.
when you see 40 positions ? you get a lot of information.
also you do not only get information from WHAT you see but also from WHEN
(related to the other program that is also computing)
you don't have to.
mark, chris, christophe and john have done. so : maybe it works sometimes )
why do you need comments when you have the game ?

As shallow as the analogy perhaps )).
Yes, but it's still only one game. The information you are able to gather is very simple and almost trivial, unless they are either reproducable against the same opponent or recognizeable from games against other programs. That's how you determine whether it's a inherent error or a coincidence caused by various parameters determined by your setup.
Nope.
That's their choice to make and no business of mine. However, I doubt that they change anything due to complications from a single game, unless it's a very obvious fault. This leads us back to the trivial errors again ).
Just wanted to see you incisive analysis on print. That's all.
Best wishes...
Mogens

[pete]

Geschrieben von: / Posted by: pete at 30 May 2000 22:14:27:
Als Antwort auf: / As an answer to: Maybe, maybe not. geschrieben von: / posted by: Thorsten Czub at 30 May 2000 21:25:19:

mark, chris, christophe and john have done. so : maybe it works sometimes )

What is your point I ask myself ?
If you have superior insight : great .
If you know many chessprogrammers and are respected by them : great .
I could continue like this and let it either finish in a nice or a nasty way .
Why is it so important for you to be special is the obvious question ?
If you post an analysis of the game I think you will find many friendly people willing to discuss your impression , but all this know this know that is simply a little boring
Either : you know everything best for yourself ; then it is absolutely meaningless to discuss it in a public forum unless sometimes to remind people how great you are or you want to share and learn ; then at least you should try .
regards.
pete

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 30 May 2000 22:23:20:
Als Antwort auf: / As an answer to: Re: Maybe, maybe not. geschrieben von: / posted by: Dann Corbit at 30 May 2000 22:00:22:

I think that a few games is definitely not enough to judge, *especially* by looking at the games. Sometimes, we see a computer make an absolutely _brilliant_ positional move or sacrifice. Unfortunately, it was not actually intended! It only fell into it.
I have seen the very best computer programs make prodigious blunders -- unthinkable C-level moves and outright gaffes. 99.9% of the time (however) they don't play stupidly. If I looked at two or three games that happened to have a horrible blunder (trapping your own piece, 3-fold repetition in the face of a clear win etc...) I might conclude that the program was terrible. On the other hand, if I saw two games with no mistakes and both of which contained a stunning and brilliant move, I might conclude that the program was a sheer genius.
Surely you have seen both sides from any computer program that has played at least 100 games against an opponent of approximately equal strength.
Do you know any program that never makes stupid mistakes -- ever?
Do you know any program that always makes brilliant moves, even positional?
Yes, a mathematical analysis lacks art and beauty. But it is a lot more stable than a touch-feely impression, even by a super GM.
IMO-YMMV.

I agree with what you're saying. My point of view is that strength is determined through a sufficient (as many as possible) amount of games against a program of "known" strength. If you want to improve the program and maybe make it stronger, then you can analyse the individual games one by one. Hopefully, you're able to recognize mistakes or miscalculations that are inherent to certain positions and correct them.
If I were to write a program that would be my approach. Then if someone told me that my program was getting stronger based on one game, I would probably ask the person in question to try another x number of games and get back to me.
Best wishes...
Mogens

MCL

[Thorsten Czub]

Geschrieben von: / Posted by: Thorsten Czub at 31 May 2000 00:29:54:
Als Antwort auf: / As an answer to: Re: Maybe, maybe not. geschrieben von: / posted by: pete at 30 May 2000 22:14:27:

What is your point I ask myself ?
I could continue like this and let it either finish in a nice or a nasty way .
Why is it so important for you to be special is the obvious question ?
If you post an analysis of the game I think you will find many friendly people willing to discuss your impression , but all this know this know that is simply a little boring
Either : you know everything best for yourself ; then it is absolutely meaningless to discuss it in a public forum unless sometimes to remind people how great you are or you want to share and learn ; then at least you should try .
regards.
pete

my point is, that if it would NOT work, i mean , to
find out about the strength and weakness of a program
by watching single games and not waiting for long match series...
if it would NOT work, how would i be able to find out about programs
strength before the strength of the programs has been proven in
long test series ???
i mean, practise refutes your ideas how it works.
of course. if you like such a kind of reflection.
for me only one thing counts: is it working.
if something is not working , i let it.
i am not doing this here for narcistic reasons.
the only goal is that it makes fun.
but it should be allowed, in a discussion where somebody
mentions that it is NOT working, to claim that it has worked (from
my point of view) so far. i remember that without doing lots of swedish-
autoplayer series it was easy to register the strength of
phalanx or zarkov or hiarcs or chess-tiger or or or or.
if you find the autoplayer more useful, ok - use it.
but don't tell me that it is not working my way.
so my point is that i have generated enough evidence for myself to make
sure THAT it works.
and it may be allowed to critisize the fact that somebody claims
that "my way" is not possible when in fact i have done it lots of time
and it works.

who says this.
i am working in secret normally. it takes 1/2 year or years until
the thing i am working on is coming on the market. normally
nobody knows about the stuff. i am not interested in letting you know.
i am in no way special. that exists in your mind.
anybody could do it. it's primitive observation. i guess i do have much
time, others maybe don't have.
we talked about programs and testing methods. you want to make
it a dialogue about narcistic behaviour. good for you. but for me
this is off-topic.

any chessplayer can replay games himself.
for analysis use fritz6.
it is capable of analysing a game. should be enough for most of the
chess players.

look. i don't need to remind you on anything.
i wanted to show or put the finger on the fact that - yes - that i know
or have recognized that comet b22 is ok.
i thought maybe somebody would be interested in finding
out himself. maybe interpret it as a kind of hint.
but stop your foolish discussion about what to post and how to post.
an open forum is open. this means anybody can post anything.
and neither you, nor me, have the right to value other peoples post.
doing this shows only lack of tolerance and nettiquette, as you do.

if you want to behave intolerant, please do it. but please
not in a chess forum. if you are different opinion about comet b22
you can tell us here. maybe with evidence. but discussions like you
do are off-topic.
and neither your bad style nor your anonymous name may hide you
from outing this.