Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Fritz5

Author: Don Dailey

Date: 18:43:07 05/23/98

Go up one level in this thread


On May 23, 1998 at 16:11:04, Wayne Lowrance wrote:

>I am seeking opinion information on the playing strength of Fritz 5.
>I have ordered Fritz5, not yet received it. I have been reading internet
>stuff that alarms me about Fritz5. Such as.
>
>Its current number 1 ranking ahead of nimzo98 and others is biased in
>favor of fritz because of a testing platform that may be friendly too
>Fritz, 100 meg of ram, not the version as the commercial version etc and
>on it goes.

I would ignore all this information.  The only hard evidence available
is the actual tests themselves and this favors Fritz.   The postings
are a natural response to people who are upset by a sudden change of
guard.  Most people take some time to adjust to change.  If you look
at a lot of the posts, they are laced with comments that make this
pretty obvious.

If it turns out that indeed Fritz had some unfair advantage, you will
likely find this is not a big deal, Fritz will prove to be a very
strong program and in the top 2 or 3 no matter how you slice it.

Unless you are Grandmaster, you probably want to pay a lot of
attention to features and other issues anyway.  In this area Fritz
is awesome compared to most of the others.
In short, I don't think you have anything to worry about.


>
>A test of all the top programs is reported where all programs start at a
>pre designated open position, all opening books are turned off. Fritz
>finished 9th in this test which suggests to me that just maybe it does
>not play as strong in endgame as other programs, like 8 others.

I don't know anything about this result.   It sounds like a way to
isolate a program from it's opening book.  Would you feel comfortable
if we tested programs without playing the endgames?  It might be
useful in some ways but it wouldn't be a very valid way to measure a
program's overall playing strength.  I know a lot of human players
who would not do nearly as well if their openings were chosen for
them.  And this has nothing to do with book fixing, it has a lot
more to do with a program's playing style.

Be really careful about conclusions people draw from any test or
experiment.  The test may be valid and useful but people will always
pick and choose the results that confirm their suspicions and they
will always believe everything else is somehow invalid.  This
of course applies to the rating list we are talking about but this
has always been pretty much trusted (until Fritz that is) and has
the merit of playing real games, not subsets of them.

But every program has it's strengths and weaknesses.  I don't have
any evidence that Fritz plays a WEAK endgame although I beleive
some other programs are somewhat better in this area.  But it would
be no simple matter to get close to the top (or on top) with a really
weak ending so this is certainly more than adequate.  It is known to
be a very strong tactician and is incredibly fast.  As far as the
opening book stuff is concerned, most if not all the top programs
work hard to "fix" their books and this is a hotly debated issue on
this newsgroup.  Since this is a technique grandmasters used, it's
pretty hard to fault them for it.   But my opinion is that Fritz has
always been a really good program and until now has been handicapped
because they were one of the last ones to NOT fix their book.  Now
that have payed attention to this, they are playing in a level field
and this has inspired tremendous jealousy.

>
>The program, i learned does not come with a CD titled "powerbook" which
>is apparently very important to its rating, at least I understand that
>powerbook is used in all its games.
>
>I am pleased to say that I have been told that I may return the program
>and purchase another program suchs as perhaps nimzo98  etc.
>
>By way of background I have been playing my oldest Son chess, program
>and against program. We started out with him using Sargon 5 on MAC
>g3-233 and I using CM5K on a laptop 100 meg pentium. cm5k usually won.
>He purchased Hiarcs4 or 5 not sure as I write and I loose with cm5k all
>the time ! I want to be the rascal, so i am looking for the best of the
>best, iE;
>Fritz5
>Nimzo98
>Hiuarcs 6.0
>Rebel 9.0
>Mchess Pro 7
>etc.
>
>I am as I said using a Toshiba  laptop pentium 100, with 40 meg of ram
>and lots of hard drive.
>
>Shower me with opinions please--I do not have but a couple of days to
>decide wheather to send Fritz5 back when I get it or keep it.
>
>P.S. Are there plans that Fritz5 is going to fadeaway as implied in one
>of the open letters this site titled "Fritz finds great move" or
>something likek that.
>
>Thanks folks  Wayne Lowrance



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.