Top

Topcomputers gaan ook de strijd aan

20 september 2007 4:25 | Laatste bewerking: 4:30

Net als bij het vorige WK zal er tijdens dit WK in Mexico City een computermatch gespeeld worden, die morgen begint. De sterkste engines van het moment treffen elkaar: Rybka (die deze zomer in Amsterdam wereldkampioen werd) en Zappa (wereldkampioen in 2005 en vicekampioen dit jaar). De match bestaat uit 10 partijen, met een speeltempo van 60 minuten plus 20 seconden per zet. We hebben de (vrij interessante) persconferentie al op film.

Het schema van de match:
20 september – partij 1
21 september – partij 2
22 september – partijen 3 and 4
23 september – partij 5
24 september – partij 6
25 september – partij 7
26 september – partijen 8 and 9
27 september – partij 10

Aangezien we al aardig wat artikelen over Rybka hebben gepubliceerd hier op ChessVibes, gaan we er op dit moment vanuit dat dit programma even geen introductie nodig heeft. Zappa staat vooral bekend om zijn goede parallel speedup, wat betekent dat het dramatisch in sterkte toeneemt op computers met dual core, 4- of 8-core (dat laatste wordt gebruikt tijdens deze match). Het beste resultaat was natuurlijk het winnen van het WK in Reykjavik in 2005, met een score van 10,5/11. De programmeur is Anthony Cozzie, een Amerikaanse aio informatica die overigens ook voor Google werkt.

Hier de persconferentie, waarin verschillende onderwerpen werden besproken, van de horizon van computers en het risico dat het schaakspel wordt ‘opgelost’ tot de Freestyle-toernooien en zelfs valsspelen bij computerschaak.

Comments

9 Reacties to “Topcomputers gaan ook de strijd aan”

  1. Bert de Bruut on 20 september 2007 9:00

    Will the machines this time use the same hardware (unlike the silly Deep Junior vs Deep Fritz match, in which the latter had been equipped with a severe handicap)? Will the opening books be the exact same?

    If not both these requirements are met, not much value can be added to the outcome of the match.

  2. Bert de Bruut on 20 september 2007 9:00

    (yes you’re right, I am to lazy to watch the video or read the tourney regulations)

  3. Dennis on 20 september 2007 12:33

    Bert, you mention “If not both these requirements are met, not much value can be added to the outcome of the match.” Why? Do you think the same about humans? Must Kramnik and Anand have the same opening books?

    To me, the value of this match is which program is the strongest (in this match). And that includes all aspects (opening books, speed of computer, knowledge inside program, and even luck).

  4. Felix on 20 september 2007 13:24

    Indeed, this match is not to figure out which engine is stronger (we have ratinglists like CEGT, CCRL etc. for that), it’s clear that Rybka is stronger than Zap!Mexico. The question is which team, or we could say engine-book combination, is stronger in this tournament.
    I guess Jeroen and Dagh will prove that Rybka’s book is superior.
    Anyway, this match is something very nice for the spectators, since we can look forward to even more top level chess in Mexico, so just enjoy the games :)

  5. Andre on 20 september 2007 21:07

    Rybka vs Hydra either in a free style tournament or pure computer match is the next step

  6. Nelson Hernandez on 21 september 2007 6:49

    Heh, I think Rybka vs. the current Freestyle Champion is the next step.

  7. Jeroen on 21 september 2007 12:06

    Hi Nelson,

    Ha, the current Freestyle champion!? Now who could that be….

    Best wishes, Jeroen

  8. Anson Williams on 22 september 2007 14:45

    The current Freestyle Champion? Isn’t that 99% Rybka anyway? :-)

  9. Christian on 24 september 2007 0:26

    Hi all,

    I was very pleased to hear about the conditions in Rybka/Ehlvest the sequel where Rybka’s book was cut down to 3 moves.

    My question is what would happen if Rybka (and all computers) were disallowed the use of a book entirely. 0 moves if you will. Would they still cruise through the WCC undefeated as Zappa’s programmer suggested.

    Some may say that no book is unfair to the computer’s but the way I look at it is that opening knowledge is really a human thing that we have spent many generations developing and giving it away to the computers for free is really unfair to us.

    Consider the evolution of our opening knowledge from classical to modern and hypermodern. Nimzowitch to Botvinnik. From Fischer to Kasparov. Things that were considered most dangerous are now considered a mistake. Due to the work of humans. I mean opening theory is really the property and birthright of humans. Is it not??

    I would really like to see a top level computer play a tourney against the top ten with no book. If it won that would really impress me and I think it would be very interesting the way computers and and top humans interacted. Perhaps we would all learn something new about chess itself. And if the humans exposed the computer in some fundamental way that would provide the computer’s with a new set of challenges and also a new way forward for them to continue their development

    P.S. what do you think Jeroen?

Iets te zeggen?





Bottom