Author: Christophe Theron
Date: 23:43:52 01/04/00
Go up one level in this thread
On January 04, 2000 at 23:44:14, Robert Hyatt wrote:
>On January 04, 2000 at 17:10:49, Christophe Theron wrote:
>
>>On January 04, 2000 at 15:32:13, Robert Hyatt wrote:
>>
>>>On January 04, 2000 at 14:02:33, Christophe Theron wrote:
>>>
>>>>On January 04, 2000 at 12:45:55, James T. Walker wrote:
>>>>
>>>>>On January 03, 2000 at 20:34:52, Robert Hyatt wrote:
>>>>>
>>>>>>On January 03, 2000 at 19:34:43, James T. Walker wrote:
>>>>>>
>>>>>>>On January 03, 2000 at 02:20:43, Thorsten Czub wrote:
>>>>>>>
>>>>>>>>Want to remind you on the fact that i will do a tournament between
>>>>>>>>a fast-searcher .-))) rebel-tiger12 without book,
>>>>>>>>and vincents slow-searcher diep in a few days...
>>>>>>>>
>>>>>>>>it will be 40 games 40/120. on 400 Mhz machines.
>>>>>>>>maybe we get so many data concerning this topic,
>>>>>>>>that we can have nice discussions here.
>>>>>>>>
>>>>>>>>I am mainly vincents opinion.
>>>>>>>
>>>>>>>Hello Thorsten,
>>>>>>>I'm sure you will post the games here for all to see. Frankly, I will be
>>>>>>>surprised if Diep is within 150 points of ChessTiger. Have you tested Junior 6
>>>>>>>yet? I'm very impressed with this program.
>>>>>>>Jim Walker
>>>>>>
>>>>>>
>>>>>>I will be surprised if they are very far apart. Diep plays a lot on ICC and it
>>>>>>is _not_ a pushover. At longer time controls it is very dangerous.
>>>>>
>>>>>Hello Bob,
>>>>>I hope you are right! I have only seen Diep at G/30 or less. It was not very
>>>>>impressive but I also believe it has faster hardware now. I hope it is an
>>>>>interesting match.
>>>>>Jim Walker
>>>>
>>>>
>>>>I wonder why Diep is supposed to be stronger with a faster computer and Tiger is
>>>>not...
>>>>
>>>>Maybe some people are already brainwashed by Vincent's words?
>>>>
>>>>
>>>> Christophe
>>>
>>>
>>>No... but some programs react better. If you run crafty on a machine, and
>>>cut the clock speed by 1/2 every now and then, the playing skill will drop
>>>steadily until some threshold value, then it will plummet... because you just
>>>dropped the search depth into the range where my null-move search assumptions
>>>begin to fail badly. A non-null program won't have that threshold to cross and
>>>at some point it will suddenly start to kill mine when the hardware gets too
>>>slow.
>>>
>>>Not uncommon at all...
>>>
>>>Whether this is Vincent's issue or not is unknown.. I only know that his program
>>>plays _very_ strongly on ICC... Against Tiger, against mine, against anyone...
>>>
>>>It never pays to take someone too lightly...
>>
>>
>>That's not how I see it.
>>
>>I'm taking the risk in public to challenge somebody who has a big mouth.
>>
>>I know many peoples are irritated by Vincent blahblah, but I see nobody (almost
>>nobody) standing up to tell him.
>>
>>What do I have to win from this match? Can you tell me, Bob?
>>
>>I have nothing to win, and much to lose.
>
>That is probably true, of course. I'm only warning you that his program
>is not a sargon or some such thing from the 80's. It is very strong and
>knows a lot of things about the game that a lot of programs don't. How it
>will do is anybody's guess. It is pretty slow in terms of NPS, so it can
>get into tactical trouble, given a chance. But if it doesn't, its knowledge
>can spell trouble for the opponent if the opponent is not a very 'smart'
>program.
>
>Will be interesting to see the games...
Of course. I don't assume it is over before it has started.
However having a lot of knowledge is not all. Lots of knowledge can fail against
less knowledge more efficiently packed.
I mean I prefer to have the right knowledge rather than tons of knowledge.
That's an important part of my work (and yours too I assume) to carefully
implement as little knowledge as needed, and not too much. Then to identify what
other bit of knowledge is missing and to implement it, and so on.
All the time being very careful not to break what's already there, and to avoid
drastic NPS drop.
It requires more patience than rhetoric.
Christophe
>>However I take this match exactly like the matches that have been played by the
>>SSDF, Tiger against X, X being a strong commercial program. I have never been
>>afraid of such matches, I'm not going to be afraid today just because Vincent
>>claims he has such a wonderful program and such a wonderful understanding of
>>chess programming.
>>
>>Actually when I read such claims I feel more confident. I don't know why...
>>
>>
>>
>> Christophe
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.