Author: Christophe Theron
Date: 09:08:45 11/09/00
Go up one level in this thread
On November 09, 2000 at 02:23:44, Uri Blass wrote:
>On November 08, 2000 at 21:24:48, Christophe Theron wrote:
>
>>On November 08, 2000 at 12:36:59, José de Jesús García Ruvalcaba wrote:
>>
>>>On November 07, 2000 at 14:53:51, Christophe Theron wrote:
>>>
>>>>On November 07, 2000 at 12:25:35, José de Jesús García Ruvalcaba wrote:
>>>>
>>>>>
>>>>>Based mainly on tournament results.
>>>>
>>>>
>>>>
>>>>
>>>>So you base your judgement that Shredder is the best program on 9 games played
>>>>back in August???
>>>>
>>>>
>>>>
>>>
>>>Hi Christophe!
>>>You know very well it has been more than nine games, Shredder has won several
>>>tournaments in the last two years (it has just won CCT2). Of course it has been
>>>previous versions of Shredder, but I have no reason to believe Shredder 5 will
>>>be weaker.
>>
>>
>>
>>If those previous versions were that superior, why don't we see them on top of
>>the SSDF list for example? Shredder 4 is currently number 4.
>>
>>
>>
>>
>>
>>>>> I do not really pay attention to private
>>>>>testing.
>>>>
>>>>
>>>>
>>>>You don't need to, of course, if you believe that 9 games are enough to tell
>>>>which program is better.
>>>>
>>>>
>>>>
>>>
>>>One tournament is not enough, of course. Perhaps I should emphazise the plural
>>>in "tournament results".
>>>Even several tournaments are not enough to determine objectively with certainty
>>>which program is stronger, but they are enough to influence my own personal
>>>subjective opinion.
>>>Thousands of privately played games have little impact in my opinion. I hope my
>>>position is now clear.
>>
>>
>>
>>
>>What is the difference between tournaments and privately played games? Do you
>>think the SSDF games are privately played?
>>
>>
>>
>>
>>
>>>>> I follow tournaments in which the authors are directly involved, like
>>>>>the ICCA tournaments, the IPCCC, the French championship (congratulations!), the
>>>>>Dutch open (congratulations again!), the Spanish championship (I have not heard
>>>>>from it in a long time, but I think you won the last one... congratulations!),
>>>>>and the CCT.
>>>>
>>>>
>>>>
>>>>So why don't you say that Gambit Tiger is the best program at this time?
>>>>
>>>
>>>It has won very convincingly two strong tournaments, and it ranks number two in
>>>my personal subjective list at this time. If it keeps winning important
>>>tournaments it will definitely replace Shredder in "my" number one spot. It is
>>>simply not yet there. But do not worry, I do not think my opinion influence a
>>>lot of people here.
>>>
>>>>BTW I wouldn't dare to say this myself based only on the few games of the French
>>>>and Dutch championship.
>>>>
>>>>
>>>
>>>But as the Gambit-Tiger author you should be very careful with what you say
>>>about its strenght, I do not have that problem as I am not involved in the
>>>development of any chess program.
>>
>>
>>
>>I am very careful with what I say about my program. For example, I would never
>>let somebody tell it is 100 elo points above Fritz6 without reacting...
>>
>>
>>
>>
>> Christophe
>
>I hope that the people who claim that tiger is 100 elo better than fritz6a are
>wrong and it is more than 100 elo better.
>
>I understood that Marcus believes that shredder5 and tiger13 are close to be 100
>elo better than Fritz6a in the ssdf list.
>
>I asked him by email about shredder5 and his prediction for shredder5 is 70-120
>elo better
>
>He claims that tiger and shredder earn from time more than Fritz6a and that it
>is the reason for the information that suggest that tiger is clearly less than
>100 elo better than fritz6a(this information is usually based on faster time
>control than the ssdf games).
>
>Uri
Neither Shredder 5 nor Chess Tiger 13 nor Gambit Tiger 1 will be 100 elo points
above Fritz6a. Let's get back to reality, please.
Now I'm expecting a "how do you know you don't have Shredder 5" question, to
which I'll not even answer, because I just have to wait a little bit and
everybody will see that it was silly to believe claims of a +100 improvement
over the former best.
It was extremely unlikely that a new program tops the former best by 100 elo
points 10 years ago, and it is now just more and more unlikely.
I wouldn't be honest if I let people say that my program is 100 elo above
Fritz6...
Christophe
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.