Author: Christophe Theron
Date: 17:00:55 04/14/02
Go up one level in this thread
On April 14, 2002 at 17:34:42, Roy Eassa wrote:
>On April 14, 2002 at 14:38:06, Christophe Theron wrote:
>
>>On April 14, 2002 at 13:57:07, Roy Eassa wrote:
>>
>>>On April 14, 2002 at 12:50:25, Christophe Theron wrote:
>>>
>>>>Yes, Fritz7 is great.
>>>>
>>>>
>>>>
>>>> Christophe
>>>
>>>
>>>The big question that's in a lot of people's minds (I'll bet) is how Chess Tiger
>>>15 will compare in strength to Fritz 7b.
>>>
>>>One thought is that you are not likely to release it until you are sure it's
>>>stronger. If that's true, it will be very exciting.
>>>
>>>On the other hand, you are probably in the later stages of the development of CT
>>>15 (I won't ask WHEN you will release it -- enough people ask that already!) and
>>>you already know it's clearly stronger than CT 14. Maybe you will decide that's
>>>enough and NOT spend (waste?) months testing it against Fritz 7b, but rather
>>>decide to release it based on its own merits and let the people test it against
>>>Fritz (and you could always release v15.01 later as an update).
>>>
>>>The big mystery -- and only YOU know the answer! :-)
>>
>>
>>
>>The answer is clear: I do not have Fritz7 and I will NOT test against Fritz7.
>>
>>As I said many times before, I don't try to improve Tiger by letting it play
>>against its probable future opponents.
>>
>>Anyway it would be a complete waste of time.
>>
>>I do not have Junior, I do not have Fritz, I do not have Shredder, and so on...
>>
>>I usually play against Genius5 on 386sx20MHz, when I have the time.
>>
>>
>
>Christophe, I hope you know that I am not challenging you and am not
>antagonistic to you at all! :-)
>
>I am curious sometimes about specific things and I ask questions, but I do not
>wish to appear to be challenging you. I ask questions and I learn from the
>answers. (I've annoyed people that way since I was very young.)
No Roy, no problem at all!
I know you are not asking question to annoy me.
It's just that I know that other people are going to read them.
And different people have different ways of understanding things. That's why I
repeat what I say in several different ways. Sorry to sound dense, but it's a
trick that I have learned here over the years: what you write is going to be
misunderstood or used against you. So you have to be extremely clear in your
message.
>You said, "I don't try to improve Tiger by letting it play against its probable
>future opponents." I actually did not remember heaving heard this before. I'm
>just curious why you think it would be a waste of time --
It's a waste of time because:
1) I do not know how to "tune" my program to make it better against a given
opponent. I do not know how to "tune" for a given time control either. "Tuning"
is something like an urban legend in computer chess. People have no idea of the
difficulty to do this, so when they need to explain something strange they will
say that some "tuning" has been done. People cannot admit that they do not know,
so they'll use whatever obscure explanation they have at their disposal.
"tuning" and "will be better at longer time controls" are great classics in this
area.
2) Even if I knew how to "tune" my program against Fritz for example, it is
likely that it would not be of any use against the other opponents (or else it
would not be "tuning" but just "improving", isn't it?). So it would be a lot of
work to gain almost nothing in strength, because the strength of a program is
measured against many opponents.
> if Fritz 7 is the
>current champ, wouldn't you be curious how your new and improved program would
>fare against it,
Yes I'm curious about the result, but this result is of mostly no use for me.
> or to see the moves of any games in which Tiger might lose?
That's the only thing that is useful. If I can see a pattern in the way Tiger
loses, then it's probably a weakness I can fix. But I need to see the same
weakness exposed over and over again to be able to fix it. So generally I'll
wait until I have a collection of position where my program fails, and
preferably against several different opponents, and then I'll try to address the
problem.
>(Would it be different if you had a couple "extra" equal-speed computers on
>which to run these contests?)
If I had more computers at my disposal I would not use them to play against my
top competitors.
>As far as Genius 5 on 386sx20, do you mean you run both programs on that
>platform and look at what percentage Tiger wins, using improvement there as one
>gauge of its strength? Or do you mean you play against Genius personally for
>fun? (Or are you referring to using it as an opponent for the Palm version of
>Chess Tiger?)
I have a real admiration for the work of Richard Lang. What is amazing is that
10 years ago or so, Richard already had found many subtle improvements that are
still missing in most of todays programs.
I let Chess Tiger for Palm play against Genius5 on 386dx20. Tiger is
disadvantaged by the hardware, but not much. And sometimes Genius plays so well
that Tiger looks like an amateur.
I believe it comes from the superior evaluation of Genius. There is not much I
can do because Tiger has to be faster and so I cannot add much into my
evaluation. But still it points me to what I should fix in priority.
>Your great success in your field speaks for itself (not to mention the fact that
>you read and post here at all!). I hope it's OK for me to ask these questions
>without seeming antagonistic.
You are absolutely right. I hope that now you know that I have nothing against
you. As I said, I have to make my message extremely clear and I have to keep in
mind that several people, with different ways of thinking, are going to read it.
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.