Discussions (not only) about Xboard/Winboard, Computer Chess and Chess Programming
by pete » 07 Apr 2000, 23:31
by Pete Galati » 08 Apr 2000, 01:20
Crafty is pretty strong, we all know that, but I don't take SSDF very seriously.I am about to go to bed but just read about the latest SSDF list at ICC where posting is not possible at the moment .
Crafty is at the same level as Hiarcs7.32 and Nimzo99
This is though suspected for awhile by a few Internet guys a rather big sensation I think .
Too tired to post more ; just thought some people here would need no more than that and felt like sharing ...
pete
by Frank Quisinsky » 08 Apr 2000, 07:35
Hi Pete,Crafty is pretty strong, we all know that, but I don't take SSDF very seriously.I am about to go to bed but just read about the latest SSDF list at ICC where posting is not possible at the moment .
Crafty is at the same level as Hiarcs7.32 and Nimzo99
This is though suspected for awhile by a few Internet guys a rather big sensation I think .
Too tired to post more ; just thought some people here would need no more than that and felt like sharing ...
pete
They put out a constant effort, and that's good, but there's allways questions about the equipment they use, very inconsistant. I'm glad SSDF is doing their work, but I don't know if it's a very valid indication of Crafty's strength. And I should probably mention that this is not the Winboard version of Crafty, this is the Chessbase verion because of auto232 problems with running the Winboard version, it's also a Fritz opening book, that must speak well for the quality of the Fritz book.
Pete
) I think people say this because Crafty is a free program )) Shredder is the team for the last 10 minutes and Chess Tiger has a good strategy for the second half time ! ColChess, Averno played at the moment with 8 players (not 11) and The Crazy Bishop with 10 players by pete » 08 Apr 2000, 11:20
I can't follow here . Inconsistant equipment ? Why ?Crafty is pretty strong, we all know that, but I don't take SSDF very seriously.
They put out a constant effort, and that's good, but there's allways questions about the equipment they use, very inconsistant. I'm glad SSDF is doing their work, but I don't know if it's a very valid indication of Crafty's strength. And I should probably mention that this is not the Winboard version of Crafty, this is the Chessbase verion because of auto232 problems with running the Winboard version, it's also a Fritz opening book, that must speak well for the quality of the Fritz book.
Pete
) was lower than what we see now .) .by Colin Frayn » 08 Apr 2000, 16:48
Nice analogyChess Genius is the German football team (safe and boring)
AnMon is the Argentina football team (active and interesting) !
Fritz and Junior is the Brazil football team and Crafty is the team from England (hard and good)
ColChess, Averno played at the moment with 8 players (not 11)
)))by Frank Quisinsky » 08 Apr 2000, 18:40
Maybe, 9 playersNice analogyChess Genius is the German football team (safe and boring)
AnMon is the Argentina football team (active and interesting) !
Fritz and Junior is the Brazil football team and Crafty is the team from England (hard and good)
ColChess, Averno played at the moment with 8 players (not 11)
LOL
Oh well... I'll have to work hard on ColChess in the near future. I haven't coded much on the engine since version 6 came out a while ago.
Cheers,
Col
by Pete Galati » 08 Apr 2000, 20:05
A group of programs running on 450mhz, another group on 200mhz, and another group running on 90mhz computers. Well, we know that the programs running on the 450mhz computers (with the most mem btw) are going to come out on top.I can't follow here . Inconsistant equipment ? Why ?Crafty is pretty strong, we all know that, but I don't take SSDF very seriously.
They put out a constant effort, and that's good, but there's allways questions about the equipment they use, very inconsistant. I'm glad SSDF is doing their work, but I don't know if it's a very valid indication of Crafty's strength. And I should probably mention that this is not the Winboard version of Crafty, this is the Chessbase verion because of auto232 problems with running the Winboard version, it's also a Fritz opening book, that must speak well for the quality of the Fritz book.
Pete
by pete » 08 Apr 2000, 23:58
and ??A group of programs running on 450mhz, another group on 200mhz, and another group running on 90mhz computers. Well, we know that the programs running on the 450mhz computers (with the most mem btw) are going to come out on top.I can't follow here . Inconsistant equipment ? Why ?Crafty is pretty strong, we all know that, but I don't take SSDF very seriously.
They put out a constant effort, and that's good, but there's allways questions about the equipment they use, very inconsistant. I'm glad SSDF is doing their work, but I don't know if it's a very valid indication of Crafty's strength. And I should probably mention that this is not the Winboard version of Crafty, this is the Chessbase verion because of auto232 problems with running the Winboard version, it's also a Fritz opening book, that must speak well for the quality of the Fritz book.
Pete
Pete
by Pete Galati » 09 Apr 2000, 00:39
Almost without exception, the list had 3 groups of program, the first group was the programs that had the highest rating & percentage of wins, they were on the fastest (450mhz) computers. Then the group with the next highest ratings & percentage of wins who were on the next fastest (200mhz) computers, Then the last group of programs who had the lowest group of ratings & percent of win who were on the slowest (90mhz) computers.and ??A group of programs running on 450mhz, another group on 200mhz, and another group running on 90mhz computers. Well, we know that the programs running on the 450mhz computers (with the most mem btw) are going to come out on top.I can't follow here . Inconsistant equipment ? Why ?Crafty is pretty strong, we all know that, but I don't take SSDF very seriously.
They put out a constant effort, and that's good, but there's allways questions about the equipment they use, very inconsistant. I'm glad SSDF is doing their work, but I don't know if it's a very valid indication of Crafty's strength. And I should probably mention that this is not the Winboard version of Crafty, this is the Chessbase verion because of auto232 problems with running the Winboard version, it's also a Fritz opening book, that must speak well for the quality of the Fritz book.
Pete
Pete
I still can't understand . The chess programs at 90mhz are supposed to lose ; right ; but the ELO system is good at handling this : statistics about programs on different hardware .
the problem you see is probably that a computer program running on slower hardware will score less than what can be expected by ELO standards ; this is well possible , but I have not seen much data to back up this argument yet.
by pete » 09 Apr 2000, 18:54
I think the results of SSDF have to be interpreted : sure you can't compare apples with pears ( don't know if this translates well )Almost without exception, the list had 3 groups of program, the first group was the programs that had the highest rating & percentage of wins, they were on the fastest (450mhz) computers. Then the group with the next highest ratings & percentage of wins who were on the next fastest (200mhz) computers, Then the last group of programs who had the lowest group of ratings & percent of win who were on the slowest (90mhz) computers.
What are you asking about, I don't have any data on the ssdf, I'm just looking at what was posted at CCC.
Pete
) . .by Pete Galati » 09 Apr 2000, 21:37
I'm one of those people still using a 586-90 computer, so when I compare the search depths that can be reached on my computer with Crafty analysis, with for example the search depths that Dann can reach analyzing the same position on his newer/faster computer, then I see a very drastic problem with the difference in computer speeds, because while I can only reach a few search depths and take a long time doing it, Dann can easilly go sometimes go 8 or 10 deeper than me, and do it in a very _small_ amount of time. So, I understand that's it's a very real situation that allot of people are using slow computers like me, I don't think it's _really_ fair to the programs to have to run on them.I think the results of SSDF have to be interpreted : sure you can't compare apples with pears ( don't know if this translates well )Almost without exception, the list had 3 groups of program, the first group was the programs that had the highest rating & percentage of wins, they were on the fastest (450mhz) computers. Then the group with the next highest ratings & percentage of wins who were on the next fastest (200mhz) computers, Then the last group of programs who had the lowest group of ratings & percent of win who were on the slowest (90mhz) computers.
What are you asking about, I don't have any data on the ssdf, I'm just looking at what was posted at CCC.
Pete
Sure you can't compare Fritz on AMD450 with Chessmaster6000 on P200 or Tiger on P90 if you want to come to conclusions about program strength .
But I think you can compair the results of the programs in the _same_ group very well even if the opponents have been out of the other group .
if Program A ( on 450 ) beats program C ( on 200 ) 13:7 ; and Program B ( on 450 ) bests Program C ( on 200 ) 19:1 I think it is possible to compare the results of A and B .
The results are still consistent ; there are two variables : the hardware and the chosen program ; a crafty on a AMD450 _is_ stronger than a Tiger on a P90 and even the margin seems to be rather predictibable .
Two reasons are posted over and over against the SSDF list :
1.) the autoplayer is buggy
I won't engage in that but as I see it at least for crafty this can't be a big problem as it uses the chessbase autolayer in its games ; and in fact if you leave out Rebel there have never been many facts to back up this argument ( and the Rebel versions that had problems don't participate )
2.) The system itself is wrong .
I more or less agree to this partly ; I think though the results are still consistent when you take into consideration what they in fact are ; programs playing long series of tournament game matches ( with or without autoplayer ) against each other . And it seems the manually played matches conducted in a similar way correlate very well with the SSDF list ( for example the Erizaquoi tournaments ) .
This won't necessarily mean it correlates well with say a 100 round swiss tournament between the 50 best SSDF programs ( double-round ) ( for example learning is then no major issue anymore ) .
What _I_ don't like about the list is that I expect that a program on a 450 will win too many games against the program on the 90 ( more than expected ) ; so the list is inflationary in itself and then it _would_ make a huge difference if you give it opponents on a 450 a 200 or a 90 .
This is IMHO the interesting question and the reason why I would like to see some facts ( but it seems not enough as else I could simply calculate a little
The results of the current SSDF seem to suggest that Crafty on a 450 in longer matches against other programs is playing at the same level as Hiarcs or Nimzo ; and _yes_ : this still surprises me a little and within the posted ( still quiet high ) error margin it is more or less statistically proven so you can't even discuss about it
pete
by Dann Corbit » 10 Apr 2000, 18:08
This stems from not understanding how the SSDF works. In fact, doing as you suggest would render the SSDF results nearly useless.But I do agree with the SSDF results about Crafty being that strong, I just wish SSDF could run all on the same equipment.
Return to Archive (Old Parsimony Forum)
Users browsing this forum: No registered users and 27 guests