Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Chessbase will win!

Author: Robert Hyatt

Date: 19:43:12 08/17/00

Go up one level in this thread


On August 17, 2000 at 05:19:32, Vincent Diepeveen wrote:

>On August 17, 2000 at 01:38:32, Uri Blass wrote:
>
>>On August 16, 2000 at 18:55:43, Marcus Kaestner wrote:
>>
>>>>>be sure that the improvement of the chessbase engines lay far behind rebel or
>>>>>shredder for example.
>>>>
>>>>Did you test all the chessbase engines to be sure about it?
>>>>Did chessbase send you their latest versions?
>>>>
>>>
>>>the answer is very easy:
>>>if they want to have a realistic chance, fritz has to gain a minimum improvement
>>>of 50 elo, junior has to gain a minimum of 100 elo.
>>>
>>>i know for sure that this is not the case.
>>
>>How do you know?
>>Did the programmers tell you that it is not the case?
>>>
>>>so chessbase will hardly be able to win the title.
>>>
>>>marcus
>>
>>Based on previous post of Amir Ban Frans Morsch is the most unrespected
>>programmer and amir also claimed that he is most likely going to release an
>>atomic bomb in his next version.
>
>I think progress on DIEP is most (but had to win a lot of terrain also),
>progress has been made clearly by Stefan Meyer Kahlen, Amir Ban and Frans
>Morsch. SOS heavily improved, Zchess heavily improved, and most likely
>Rebel also improved if the comments of Jeroen Noomen are taken into
>account for Rebel. I don't doubt that Tiger will be improved and
>most likely Insomniac last time i played it (1999 at dutch champs) also
>will be heavily improved.
>
>So with all respect. All programs are obviously going to be improved,
>where the stronger engine makers make more sound when making a small improvement
>it must be said that some stronger engines you need to work weeks to get
>a 1% speedup or to get 10 rating points. History has proven so, advancing
>at the top is real hard. No 100 points will be won here easily.
>
>This where especially the rest that was clealry below all this is
>getting up real quick. The real questions now are of course: do they
>have a real weak point yet?
>
>Engines with a real weak point are gonna have a problem, Also if their
>book is real weak you have a real problem.
>
>No matter how an engine improved, take for example crafty, if book is
>weak then you can usually already shake it, as in some programs level
>really has gone up there.
>
>Personally i think openingsbook as made by Jeroen Noomen are giving
>the best tournament preparement, though the strategical lines in shredder
>also give clear positions where shredder is having a positional finer
>position (but hard to call that killerlines).
>
>This year there will be some real strong programs with books based
>upon either old lines or automatically generated lines. Here the Kure
>book is giving a deadly advantage, as it's the best to litterary kill
>those old/automatically generated lines.
>
>>Based on this information I will not be surprised if Fritz's version is 100 elo
>>better than Fritz6.
>
>That's the biggest nonsense i ever heart. Fritz will never get up 100
>points easily. Frans would have worked years for nothing otherwise and
>then suddenly within a few months he gets up 100?
>
>Only when the version before that first got down 100!
>
>>I will be surprised if fritz's version is less than 50 elo better than Fritz6
>>because of Amir's post.
>
>Amir is commercial obviously. He needs to drum to the world. I doubt it's
>50 rating points better as the previous version, unless the previous version
>was not his best version from the past times.
>
>>My guess for the first 3 places in order to win chessbase8 was:
>>1)Fritz
>>2)Shredder
>>3)Crafty
>
>Why do you place crafty that high, considering it has a real weak point?
>
>>I guess that all these programs are more than 50 elo better than fritz6 when in
>>the case of crafty the main reason is better hardware.
>
>Crafty played at even more superb hardware at the WMCC 1997.

No it didn't.  we played on a plain 21164 at 600mhz.  Remember that we didn't
have one of the Kryo machines at all, just a plain (and not fast) alpha.



>
>>I guess that Rebel,Tiger and Nimzo8 and maybe Junior are also better than
>>fritz6.
>
>I don't give a penny for the nimzo engine. I'd die to get my hands on
>a legal usage of the nimzo book however. It isn't tuned to play the
>opening itself other than a few single lines it plays. Nimzo is an aggressive
>engine that's real good tuned, but i'm everything but impressed.
>
>Especially knowing the comments of its maker upon how he sees
>computerchess : "nimzo improved only in playing strength when i REMOVED
>knowledge", we can assume that unless he has come back to that statement,
>that the engine will not have more knowledge.
>
>>Uri



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.