Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Tournament update(about crafty)

Author: Robert Hyatt

Date: 19:55:15 01/14/99

Go up one level in this thread


On January 14, 1999 at 09:05:10, Enrique Irazoqui wrote:

>On January 14, 1999 at 08:35:17, Robert Hyatt wrote:
>
>>On January 13, 1999 at 18:00:48, Enrique Irazoqui wrote:
>>
>>>So we have another bug in Crafty 16.1 for Fritz. Before the first match, I went
>>>to ?Engine parameters?, unchecked ?No tricks?, checked ?Draw normal? and then
>>>clicked on Save. I looked at it just now and parameters are not saved at all,
>>>which means ?Tricks? are on and ?Draw normal? off unless this is modified every
>>>time Crafty is loaded... Same for 16.3, by the way.
>>>
>>>Enrique
>>
>>While it won't help you here, I am going to 'default' these things to what makes
>>sense for your kind of tournament, and let my 'icc' interface code turn them
>>back on when appropriate.  Until xboard 4.0, I didn't have enough information
>>to do this, so I had to assume I was playing a 'human' on a chess server and
>>deal with that as the default.  But now I can pretty well tell that I am playing
>>on a server or not.  And I'll fix this so it won't be a problem.  I can't do
>>much about the book/tablebase issues however, since they are beyond my control
>>apparently
>
>If tablebases are not used at the root, they better be on disk, no? or otherwise
>the access would be too slow. How much space do they take? I read somewhere that
>Crafty tablebases use some 20GB... Using them the Fritz way has the advantage of
>leaving them on CD.

yes...  they need to be on disk.  I am using all of them, and they need 22.5
gigs.  Eugene has done the on-the-fly decompression code, but I haven't had any
time to test it as of yet.  That will reduce this to maybe 5-6 gigs, and do the
decompression on the fly at a not very significant penalty.

The fritz approach means a CD is ok, but it really doesn't help the engine
play any better, because you have to _know_ you are trading into a won KRP vs
KR ending, not do it and then find out you are drawn.  Lonnie played three
games vs Crafty this morning and _every one_ screwed up.  One was likely
drawn but he let crafty find a way to trade into a won KPP vs KP ending.
In another he had a won game but traded into a drawn KRP vs KR, when he
could have traded into a mate in 39 KRP vs KR if he had been probing in the
tree.  This happens _regularly_.




>
>What about releasing a book for Crafty that can work in Fritz?


I'm willing to do most anything, but I have no idea what is needed, since
I don't know how they read the book, how their format looks, etc.  IE it
seems odd to have a book algorithm that works perfectly well already built
into crafty and it can't use it...




>
>> (as is the hash table issue and the continuity from move to move that
>>is broken by stuffing moves into crafty every time the opponent makes a move).
>
>The clearing of hashtables is supposed to be solved in the engine Crafty 16.3
>for Fritz, but there are other wonderful bugs, and I already got 3 different
>versions (bug fixes) of 16.3... The saga continues. :(
>
>Enrique


where in the devil are these bug fixes coming from?  Even I don't have that
many in one version... usually... (knock on wood)...

:)



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.