Author: Albert Silver
Date: 08:11:40 07/22/04
Go up one level in this thread
>>>I agree but commercial programmers prefer to sell their program and not to get
>>>correct rating for it.
>>>
>>>Uri
>>
>>How big a difference would it necessarily make? Today's top programs are in
>>their Nth generation, and many times the engine isn't really so different from
>>the previous. FOr example, despite not doubting for a minute that Shredder 8 is
>>stronger than Shredder 7.04, are its replies so different that preparation would
>>necessarily be so hard?
>>
>> Albert
>
>SMK had to think about it earlier.
>
>He did not have to sell previous versions but he prefered to make money.
>Now it is too late to get reliable rating for Shredder and the best candidate
>that I can think about to get a rating that we can trust is falcon(unfortunately
>I do not expect it to play in human tournaments).
>
>Uri
I think that's losing perspective.
First of all, the tournament proves very little. We saw that Bahrain qualifier
where Junior started out with 5-0 against Fritz, and Fritz is no lightweight.
Things happen.
Note that you mention he didn't have to sell previous versions but preferred to
make money. That's as odd a statement as I can think of.
First because Shredder 7.04 was a big advance over Shredder 6, and it scored a
2752 TPR at the same tournament last year
(http://www.chessbase.com/newsdetail.asp?newsid=1111). One can hardly claim in
any circumstance possible that this version is suddenly 150 points weaker.
Furthermore, that report only mentions the GM tournament result, which according
to Chessbase 8 actually had a TPR of 2745. Yet, it also played in the IM
tournament that same year, just as it did here. Do you know how it did? It
scored 7.5/10 (as opposed to the 8.5/10 in the GM tournament) and scored a TPR
of 2627.
If one accepts that Shredder 7 was a significant improvement over Shredder 6,
why should its release have been delayed? To wait for Shredder 9 and then see
how that does, 4 years later, in a couple of GM tournaments? The people who lose
in this deal are both SMK and the consumers. Me, I'd much rather have Shredder
7.04 (which I do) and not worry so much about its genuine FIDE rating.
Since I don't believe for a second that Shredder 8 is weaker than Shredder 7,
and yet it scored a TPR of close to 100 points less overall, it just shows two
thing IMO:
- One tournament really means very little when judging a program's ability.
- SSDF is absolutely NOT a measuring stick for judging a program's ability
against human players. Aside from the obvious arguments about different rating
pools, remember that engines play differently than humans. They are far more
tactical, and far less positional. So beating up on engines just shows how well
it can beat up on engines, but since humans play differently, the only way you
can measure how well it will play against humans is by playing humans.
Albert
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.