Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Tablebase sizes: 6 man? 7? 8? ...

Author: Robert Hyatt

Date: 06:48:08 11/18/98

Go up one level in this thread


On November 18, 1998 at 01:28:36, Hristo wrote:

>
>Bob you made your point! I made mine!
>The technical details you are going into, "seem" to be stretching you a bit!
>:)))   Lets not confuse ANSI C IO and actual platform IO. SCSI(IO) supports
>large hardrives regardless of the WORD width of the processor(motorola, intel,
>...) or the bus IO ... you are absolutely correct when talking about prtability.
>I was simply trying to separate those things ANSI IO and PC IO ... we can go at
>this forever :)) but what is the point?!
>
>


I agree.  Just remember that *everything* I talk about when I talk about "C"
means ANSI/POSIX compliant C.  Everyone here knows enough about "Crafty" to
understand that it works on every possible platform it has been tried on, due
to this standards compliance.  And that it even includes code to work around
issues (endian for example) where the standards are "weak".

My post really was nothing more than a query to Eugene on how to access large
files, although I was really thinking of how to do this inside the ansi
framework..





>DeepBlue is an interesting topic! (much more than some silly IO stuff)
>Would you, please, start a new thread to set the record straight!(I wouldn't
>know what would be the topic)
>I watched the match. Played(analyzed) all games. None of them proves(or even
>shows) beyond the shadow of a doubt that DB was the supperior player.( I guess,
>I expected more from a machine that evaluates 200,000,000 pos per second)
>Despite that DB won. Kasparov lost. At this given instant(match) DB was the
>better player.(I hope you get the difference between "the supperior" and "the
>better") I wouldn't even consider calling the DB team "cheaters"! However I
>would like to see more games by DB, even if they were DB vs DB.
>
>hristo.




I don't think 6 games are enough to conlcude "DB is superior".  They  are
certainly enough to conclude "Deep Blue was superior in that 6 game match"
which is a different thing.

I would like to see more games.  DB Junior exists, still.  I'd love to see
Hsu and group put it up on ICC.  It would answer a *bunch* of questions in
a very short period of time.  They could give every programmer a 20-game
match shot at it... then we'd *know* what they could and couldn't do.  They
could also extend their absolutely awesome win/lose record on ICC (finger
scratchy for an example).

My main conclusion from the DB match was that the thing is an awesome calculator
and it is *not* a positional dummy either.  IE it drew games none of the rest of
us could have drawn.  It won a game easily (round 6) that none of the rest of us
could/would have won.  It just played very strongly.  Even if Kasparov made the
occasional slip, he still wins his matches.  And tournaments.




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.