Computer Chess Club Archives


Search

Terms

Messages

Subject: De Gorters review of Crafty 16.6 (integral text, English translation)

Author: Jeroen van Dorp

Date: 05:25:46 04/25/00

Go up one level in this thread


As I don’t know if we’re talking the same, I dug it up and here it is:

This is an excerpt from “Computerschaak” (CSVN) , 1/2000 (February),
“Topprogramma’s bestrijden elkaar in vijf toernooien” (Top programs battle each
other in five tournaments”) by De Gorter.

*The article reviews Crafty, Rebel C, Nimzo 7.32, Hiarcs 7.32, Fritz 6, Junior
5, Shredder 4 and Chess Tiger 12.*

I know about copywright, but I think De Gorter’s critic has become distorted
because of the indeed very bad text at KasparovChess.

Feel about the review as you like, but in “Computerschaak” it’s not the same
conclusion.
It’s a critical but fair review of *all* the above mentioned top programs. De
Gorter mainly looks at playing style and the input the programmers / opening
book creators / GUI developers have made. His original statement is that all
companies tell us their product is super strong, so he wants to test those
claims.
A few cites now are drifting around, somewhat in the direction of “De Gorter
says Crafty is worthless.” That is definitely not true.

Basically I also think Crafty doesn’t belong in this article as a freeware
engine, but the reason he discusses it is because it is such a well known engine
under the ChessBase GUI
It’s Crafty 16.6 and I have no information if he used tablebases, but from a few
remarks (else) in the article I suppose not, and certainly no 5-man.

This is a translation (under my sole responsibility) of the paragraph (original
in Dutch) about Crafty, BTW the very first program to be reviewed in the
article.
Between brackets two remarks for clarification by me, they don’t belong to the
article itself.

“Crafty
On the Internet Chess Club al kinds of versions of this programs play , almost
always on superiour hardware. In this case I think of two or four 500MHz
processors and a large amount of RAM. The results are in accordance with
expectations: SingaCrafty, CraftyCrafty and Crafty are a couple of names of
players who have a rating of over 3000, gathered in blitz games. Crafty is a
free engine with ChessBase if you purchase Fritz. That is the reason I was
curious how it  performed against other programs if playing on the same
hardware.

For this tournament and all following I used two Mendecino (Celeron) PC’s on
433MHz with 64Mb RAM. Time controls were 30 minutes a player a game. The results
of the A tournament leave no doubt. Under the same conditions Crafty is no match
for the best commercial programs. (*remark: it ended last, Fritz 5.32 was first,
then junior 5 and Chessmaster 6000 – JIMvD) Even under computer programs it’s
tactically very strong, but regarding mobility of pieces and some endgames it is
not really up to the rest.
I have to add that this is only the result of one tournament, and that is much
too less. In tournament C (*remark: also 30,0 where Crafty ended second together
with Fritz 6, Junior 4.6, only after Junior 5 at one– JIMvD) you will see that
Crafty indeed can be very dangerous.

But my opinion  about the weak points didn’t change because of that result. The
program is not into full balance and will experience peaks to one side or the
other  because it sometimes stumbles upon a tactical  windfall”

So far. So it’s not: “a bad program, don’t use this fluffy stuff”, but “it is a
very strong freeware program, of which I think that at moments is playing not
consistent enough.” It is not that well *balanced* over the game as a
whole,that’s the bottom line..

Fair if you consider the fact that Crafty is the typical development project.

Hard if you’re the programmer, as we are all mere humans wanting only
compliments.  The other programs get their share of critic  as well. His
conclusion BTW is that Nimzo 7.32 is the most well balanced program in all
stages of the game, although that doesn’t show from the tournament results.

Basically he is looking at a program that plays balanced throughout the game,
keep that in mind while interpreting the results. It is one way of testing and
evaluating, not only looking at points and tournament results.


--
Jeroen ;-}
-------------------------------
That ChessBrat @:
jimvandorp@wxs.nl
http://zip.to/jeroen
ICQ#45740870
-------------------------------




This page took 0.01 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.