Author: Rolf Tueschen
Date: 10:11:14 06/04/03
On June 04, 2003 at 12:22:12, Cliff Sears wrote: >On June 04, 2003 at 12:05:49, Thorsten Czub wrote: > >>computerchess is dead. >> >>nothing new under the sun. everything boring. >>no new ideas. all ONE big company (how boring). >> >>It's like 40 years SED in the GDR, boring, boring , boring. >> >>I will switch the hobby ... >> >>fritz7, 8, shredder7, 8, junior 7,8 >> >>boring, boring boring. >> >>all the same user interfaces. all the same bugs in the chessbase >>gui. >> >>it's really boring. >> >>why buying ?? >> >>chessbase monopol was the end of computerchess. >>infinite power. infinite boredom. >> >>good bye. > >It does seem stupid to pay almost $50 and all you are getting is a new engine >(and not a new updated GUI that you already have) Why stupid? In computerchess business that has a long tradition after CC had left science. Normal would be to test something until the results are valid but SSDF tests until the next dateline is arriving and then they present their data. Saying that validity is unneccessary and that in the next publication the rest of the test results is included. That spooky tradition has a simple reason. The business is eager to get the listing in time when the next selling date comes nearer. You get the idea: with invalid data you can prove almost everything. And that again gives advantage to the company that arrived in time to be "tested". It's a real Kuddelmuddel as we say in German. But if you say a word people become very angry because they only see the old and traditional five or seven testers in Sweden who allegedly should be discriminated. Of course it's the other way round. If someone tells them how to test on a scientific base and what they simply cannot do in practice, then that is _real_ worshipping. Openess in critics is friendship, hiding the scientific truth is bullshit. Perhaps we should learn this phrase by heart! Normal would be to present a new update when something spectacular has been reached. Something in chess! But reality is that too many people believe in bugs. That is because they let autoplay instead of playing themselves against a program. Normal would be to be honest about the real strength of computerchess programs. But in CC it is tradition to always call the new update the best program ever, mostly on the base of a recent WIN against a human super-GM who had been bought=engaged in a show event. In reality everybody who plays decent chess does know that chessprograms cannot play real chess yet. They have their strengths but also their weaknesses. Now - by definition - in show events the human chess Grandmasters are NOT engaged to play their normal chess level which included "nasty" play. Nasty against computerprograms is by definition playing the weaknesses of the machine. In tradition of CC however the GM is engaged to "work-around" the weaknesses of the machine so that it appears as if the machine could really play chess. Let me add a few ideas to this appearing of playing chess. To weak players (=clients who should buy a program) the programs seem to be very strong because with their exact play for a certain strictly defined limitation of depth the programs win every game against players who lose pieces and pawns by mere oversights. Weaker players can't imagine that the genius of human GM is mainly their memory and their exactness of calculating. But the real strength is their genius in finding a way to solve any position no matter how deep it is. So the two first talents guarantee that they keep up the pace and the third one is for the winning execution. On the other side of the board it is the main weakness of the machines that they are rather determined in their behavior. Ok, they might vary in 27 aspects but what is the number 27 when a human GM has thousands of such variations. And the moment a human GM has the chance to adapt to the specific weaknesses of a machine, the main part of the execution is done. All this is so basic, so trivial, that one wonders why computerchess freaks still believe in magic. Even the best programmers believe the de facto results out of show events. So Amir does believe that Deep Junior is a real GM. The only solution for computerchess is science. It's a hard and frustrating distance to go. But it doesn't help. Fakes cannot replace science. The GM won't tell what's going on as long as they profit from the traditions in CC. ============= P.S. I took this message into CCC because the topic belongs into CCC. The thread actually exists in CTF, the twin group of CCC, where someone (Ed Schröder, a former CC World Champion) opened the question How could we make computerchess interesting again in CCC... Rolf Tueschen
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.