Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Latest millenium news?

Author: Mogens Larsen

Date: 03:00:30 04/17/01

Go up one level in this thread


On April 16, 2001 at 22:37:24, Chessfun wrote:

>I'm not sure either Amir or Franz would see it like that.
>Shredder had a choice and it made it.

Again you're relying on information you don't have available. Without knowing if
the breaking point was money, conditions or something else, you can't know if
they had a "choice".

>After months of negotiation it went knowhere.

You don't know if the negotiations have been taking months, the exact progress
or the problems, if there were any.

>Being done, who else is there.
>As for the rest IMO.
>Ferret...Not commercial (for me I'm therefore glad
>Any other non commercial see above.
>Patzer...Comeon.
>Diep.....LOL
>Crafty...IMO would do ok against the humans.

Adding PConners (with a GM norm), Zugzwang and SOS (amateur WC) makes it even
less ludicrous to talk about this kind of arrangement. And others who may have
an interest in developing a SMP program before the event. Probably not many, but
nonetheless.

I might add that they invited Deep Blue, which is neither commercial or in one
piece. But maybe Han Solo (or whatever his name is) is working all night in his
lab to get ready. I kind of doubt that. Just a commecial ploy as this event.

>So the choices were really rather slim.

Noone claimed otherwise, which is another good reason for making the effort. Not
a lot of engines equals more rounds and a reasonable justification of the
winner.

>Will be done with a number of matches not just a 10 round swiss.

Again, you don't know how many matches will be played. I doubt they'll continue
until there's statistical certainty, which makes "the strongest program" issue
just as unclear. And who's talking about a swiss?

>Again why you insist on all these other programs I find hard to understand.
>I have seen you suggest Diep and Patzer? do you really honestly believe it
>is worth the trouble or that those programs deserve to be considered.?

I've explained in a previous post that there are two ways to do this fairly. One
is selecting the champion and the other is to open the event completely, that
means challenge by right (champion) or qualification (all eligible). An
invitational infers limitation by choice (discrimination if you wish), which
contradicts fairness in this circumstance. Either it's a champion vs. champion
match or a man vs. strongest comp match. Neither option is solved by
invitations.

This proposed match is pure exhibition, ie. Kramnik vs. ChessBase, despite the
forseeable headlines. Nothing else at all. And this is backed up by the
conditions relayed here by the organizers. Apparently, ChessBase agreed to just
about anything presented before them to exploit the (maybe) troublesome
negotiations between Millennium and BGN.

Arranging an invitational with conditions that they knew would scare Millennium
off (otherwise there would have been no negotiation problems) is a blatant scam,
giving the current "surprising" result. A pure attempt at making the challenger
event legitimate. There's no need to infer secret money transactions as TC does,
it stinks enough already.

>It is, the original tournament you mention may be of what 10 games, very easy
>for the best program not to make the playoff where it would win.

The format is easily adjustable given the number of engines. A pool division
with seedning according to tournament results in ICCA approved tournaments with
many engines. Give it a week and you'll reach a reasonable amount of games in
each pool with autoplay. With author participation, fewer matches but smaller
margin for random error. The playoff could be twenty matches. With very few
engines then a knockout format with seedning. Still not a real problem.

Another thing worthy of consideration is the current arrangement. How many days
do you really think they will go on with supervision, accomodations and media
interest (and Bertil's vacation)? I doubt they'll go beyond a week, probably
less. How many games will they play in that time? Maybe four or five a day,
which results in approx. twenty to thirty games. Not enough to demonstrate
strongest. And since it's autoplayeer games, they should be checked for errors
as well.

The timeframe would actually be quite similar for both types of arrangement and
the number of games for the eventual winner wouldn't differ much either. Not to
mention a more trustworthy winner given different opponent types.

>Yes I have seen the names and I'm still smiling LOL.

Well, I'm just sad that you've eaten the scam with hook, line and sinker.

>I have an equal idea at least to yours....no?

Apparently not. If you did, then you would be knowledgable about previous
official tournaments where they participated and that they're no pushovers
judging by performance.

>That being the case I am entitled to give my opinion, if yours
>is to include Diep and Patzer so be it.

Certainly, but that isn't the case.

>I agree. Actually my opinion is on equal hardware Tiger 14 is stronger than Deep
>Fritz. However to say that it's impossible to determine the strongest
>programs is IMO a fantasy. It seems to me very similar to seeding the favorites
>something done everyday. The favs in this case determined by two notable
>computer experts.

No, it's comparing the possible number of games with statistics. Nothing else,
whic eliminates the "strongest" claim.

As for the experts. Well, I don't think most people here are very impressed.
Enrique manages to taint his reputation by arranging a invitational with a given
result, while also being a beta tester for that company. And Bertil offers the
SSDF stamp of approval to this arrangement knowingly or not. All in all, not
something that gives bragging rights.

>No, because the terms of the actual match with Kramnik also have to be taken
>into account prior to any qualifying tournament. Again it would IMO takes
>months.

The BGN acts on behalf of Kramnik, which is displayed by the conditions made
behind the invitational, so that wouldn't be a real problem. The terms of the
match are not different than they are know. The identity of the challenger
doesn't matter as long as there's a three month notice, which there would be.

Regards,
Mogens



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.