Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Proposal: New testing methods for SSDF (1)

Author: Dirk Frickenschmidt

Date: 10:44:34 04/14/98

Go up one level in this thread


Hi Thorsten,

On April 13, 1998 at 15:22:38, Thorsten Czub wrote:

...[snipped most]

>The question is not if they manipulate.
>The point is that the auoplayer is not an open device.
Correct: their auto232 autoplayer is not a device available in public.

Perhaps the best really would be if a neutral programmer named by the
SSDF checked the integrity of the autoplayer. I accept what Ed said: a
programmer would be best suited to find out if this autoplayer shows any
irregular behaviour compared with the public auto232 or not. Of course
this programmer should be neutral and no competitor.

I already stated that I myself saw no strange effects when I used it.

So you may understand I hardly believe that a programmer would find big
sensations... So even if this point is debatable and perhaps needs
clarification (after I read Amir Ban's substantial post on the matter I
can at least imagine what kind of irregularities are imaginable: nobody
else discussing the matter named this precisely as he did), after all
which I saw there's still no reason for hysteria and right away calling
for disqualifying the Fritz5 results .

>You can talk about if and what and all kind of stuff. But the main
>problem is: here an exception was made to give ONE participant
>a special device,

If you like to call something which seems to be doing principally the
same as a standard auto232 device (namely playing autoplayer tournament
games in a row) "a special device" you can of course do so.
Though it sounds a bit funny to me, like they were using a time tunnel
or Scotty's beaming device or something like this within a simple
auto232-player, so it makes me smile a bit...
I for one hardly would call the Fritz auto232-player "a special device"
after all I could see :-)
But I find it a good idea to have a programmer check it to end this
discussion once for all.

Now soon as you show me some real reason for coming to other conclusions
- and not just by bad feeling or raising suspicion without *any* proof,
without *any* observation about published SSDF-games which anybody can
easily replay - we can talk about it. And I will say: "you were right".
If you were right.
No problem.

By the way, don't you think that someone like Ossi Weiner has been
searching - or has others search - day and night within the published
SSDF games to find possible inconsitencies? This is one *more* reason
why I hardly believe in all the funny rumors spread everywhere: *if*
there was anything substantial about them we could read the evidence in
"open letters" and advertising from this very person as *fast* as
possible, possibly yesterday :-)))

You know me well. I regard arguing without substantial observations as
poor arguing. And I know you well enough to be shure you can do better
and argue based on concrete observations, too.

>a special hardware platform,

Wrong.
The SSDF began upgrading to 64MB. They not only use these machines for
Fritz5. So why should this be a "special hardware platform"?.

I am in favour of uniform platforms anywhere: in the SSDF as well as in
world championships. As a user I like to see what these programs do on a
P200MMX (or perhaps a Pentuium II-300 next year) with hardware I as a
user have myself (32Mb recently, now 64Mb - and not for the sake of
Fritz5! :-) )

Using exactly the same (P200MMX and 64Mb) as the SSDF did is *one*
reason why I don't believe in all these weak conspiration theories - and
frankly, am laughing aloud everytime when I have a look at Ossi Weiner's
advertising, claiming he has a "clean" SSDF-list - done by himself and
having his programs as leading programs on top.
Best computer chess joke I have seen since a while ... :-)))
And nearly any computer chess expert I know is laughing about this as
well.

>a NON open version,

Wrong again: if Chessbase did not lie. On their website they say:

"The Fritz engine that played in Sweden and in the Selective Search
tests is exactly the same as the one you get in the Fritz5 box. There is
no "secret formula-1 Fritz" which we are holding back from our customers
(why on earth would we ship an inferior version?). All games played in
Sweden can easily be reproduced using a standard, off-the-shelf Fritz5."

The *only* acceptable reason I can see for claiming the opposite of
their official statement is in my eyes *proving* the opposite or at
least showing *some* sort of evidence - or shutting up...

>non open book,
I seem to be using the same book (Fritz5 power-book) as far as I can see
from the published games.

So what do you regard as "non-open" with their book?

>etc. etc.

What do you mean by "etc. etc."?

Do they use a secret stealth device to make Fritz5 unseen to the enemy
and thus confuse opposite programs?
Or perhaps Fritz5 is reading the opposite computer's e-mail with his
"special device" because he would be bored by just playing chess
otherwise? :-)))

... [more snipped]

>If they refuse to be fair or refuse to compete with publically available
>stuff, they refuse to compete.

If someone calls a program "not publically available" of which
*anything* ssems to be available in public (engine-version, playing
conditions as I can easiliy reproduce them, RAM size like on more and
more computers, powerbook like I have it myself) - with the one exeption
of an auto232-device which they don't like to be misused for modern
methods of book-cooking,

sigh, then this rather reminds me of the Ossi Weiner kind of advertising
and propaganda than of serious reasoning which I could follow.

Last word: I am really open to any kind of evidence and will immidiately
consider it as soon as I see it, but I am no longer even a bit open to
constant rumors witout recognizable substance...

And believe me, I will refuse to confuse you with Ossi as long as I can!
;-)

(Hart aber herzlich gemeint...)
Yours Dirk





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.