Strongest Fruit - what's the verdict?

Discussion of anything and everything relating to chess playing software and machines.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
User avatar
mclane
Posts: 16181
Joined: Thu Mar 09, 2006 5:40 pm
Location: US of Europe, germany
Full name: Thorsten Czub
Contact:

Re: Strongest Fruit - what's the verdict?

Post by mclane » Fri Sep 21, 2007 10:01 am

i am discussing with you. part of a discussion is that i don't have to be your opinion but can stay to mine.
if that is to tough for you to handle...

nczempin

Re: Strongest Fruit - what's the verdict?

Post by nczempin » Fri Sep 21, 2007 10:29 pm

mclane wrote:i am discussing with you. part of a discussion is that i don't have to be your opinion but can stay to mine.
if that is to tough for you to handle...
There are plenty of discussions that we could have that would be far more interesting, and where any opinions were actually involved.

But Mathematics (==bean-counting? Did your maths teacher not like you?) does not involve opinions. You can be of the opinion that you know which Fruit version is strongest, and, based on the data given so far, you'd be wrong. And you'd be misleading all the others.

You can be of the opinion that God exists, all Christians are evil, or that the Bush administration is actually behind the 9/11 bombings, and that the moon landing was staged, and weigh the pros and cons in a discussion. But with Mathematics this option does not exist, no matter how much you wish for it.

Uri Blass
Posts: 8590
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: Strongest Fruit - what's the verdict?

Post by Uri Blass » Fri Sep 21, 2007 11:00 pm

nczempin wrote:
mclane wrote:i am discussing with you. part of a discussion is that i don't have to be your opinion but can stay to mine.
if that is to tough for you to handle...
There are plenty of discussions that we could have that would be far more interesting, and where any opinions were actually involved.

But Mathematics (==bean-counting? Did your maths teacher not like you?) does not involve opinions. You can be of the opinion that you know which Fruit version is strongest, and, based on the data given so far, you'd be wrong. And you'd be misleading all the others.

You can be of the opinion that God exists, all Christians are evil, or that the Bush administration is actually behind the 9/11 bombings, and that the moon landing was staged, and weigh the pros and cons in a discussion. But with Mathematics this option does not exist, no matter how much you wish for it.
Note that data of games+results is more information than results only.

Mathematics does not prove that you cannot get correct conclusions based on all the information.

Inspite of it I believe that people who claim that they know which version is stronger only based on result and looking at games overestimate their ability.

Note that it is possible to make an experiement to prove it or the opposite
You can give somebody 2 engines with small difference and 100 games of both of them together with all the data of the main line and ask the person to evaluate which engine is stronger when in half of the cases the engine that scored worse is stronger.

You can see in how many cases they are going to guess correctly the stronger engine.

Uri

nczempin

Re: Strongest Fruit - what's the verdict?

Post by nczempin » Fri Sep 21, 2007 11:36 pm

Uri Blass wrote: Note that data of games+results is more information than results only.
Sure, but so far no-one has made a statement about the games. For example, Torsten, who is a much stronger player than me, could have said "I have looked at these games Fruit A and B played, and, even though the actual tournament doesn't show it all that obviously, I think that version A is stronger. Many of B's wins were due to atypical opening choices, and A had a number of draws that occurred because of a bug that is fixed now. ..." Or something like that.

Well, it turns out that he didn't do any such thing. Instead, he pointed us to a Web Page with a bunch of beans to be counted. Did he expect us _not_ to count the beans?

Mathematics does not prove that you cannot get correct conclusions based on all the information.
It can tell you that, using the results only, no conclusion can be made at any reasonable confidence level. It's a good way to complement a qualitative analysis I would expect from a strong player experienced with computer chess engines.

Inspite of it I believe that people who claim that they know which version is stronger only based on result and looking at games overestimate their ability.

Note that it is possible to make an experiement to prove it or the opposite
You can give somebody 2 engines with small difference and 100 games of both of them together with all the data of the main line and ask the person to evaluate which engine is stronger when in half of the cases the engine that scored worse is stronger.

You can see in how many cases they are going to guess correctly the stronger engine.

Uri
Well, as Bob mentioned in one of the posts about Gambling, sometimes it is the best to simply take a one-off 50 % chance.

Kohflote
Posts: 200
Joined: Wed Sep 19, 2007 9:07 am
Location: Singapore

Re: Strongest Fruit - what's the verdict?

Post by Kohflote » Mon Sep 24, 2007 2:50 am

Dear Ryan,

You mentioned "A new release that is clearly better than others is not far off though". Could you provide more information, please?
For me, I am interested in longer time control such as 40 moves in 150 minutes.

Thank you,
Kohflote

Post Reply