Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Which is Better Tactical or Strategical Knowledge for Chess Programs ?

Author: Jorge Pichard

Date: 07:45:53 03/07/00

Go up one level in this thread


On March 07, 2000 at 10:35:10, Jorge Pichard wrote:

>On March 07, 2000 at 01:37:02, blass uri wrote:
>
>>On March 06, 2000 at 20:56:45, Jorge Pichard wrote:
>>
>>>On March 06, 2000 at 13:50:49, blass uri wrote:
>>>
>>>>On March 06, 2000 at 12:24:55, Jorge Pichard wrote:
>>>>
>>>>>On March 06, 2000 at 01:18:05, blass uri wrote:
>>>>>
>>>>>>On March 05, 2000 at 18:09:29, Jorge Pichard wrote:
>>>>>>
>>>>>>>Most Programs perform excellent whenever there is a tactical position, but some
>>>>>>>programs perform better tactically than strategically. A good comparison is
>>>>>>>between Rebel 10 or Tiger Rebel, which are very good programs strategically
>>>>>>>compare to Deep Junior or Junior 6.0.which is very good tactically. But when you
>>>>>>>match them Deep Junior 6.0 is a little bit better than either of the previous
>>>>>>>mentioned programs.
>>>>>>
>>>>>>Why do you think that Junior6 is better in tactical positions relative to
>>>>>>TigerRebel?
>>>>>>
>>>>>>Junior6 searches more nodes per second but you cannot learn from it that it is
>>>>>>better in tactics.
>>>>>>
>>>>>>I know that Rebeltiger did better results in enrique's test suite relative to
>>>>>>Junior6.
>>>>>>I know that most of the improvement from Junior5 to Junior6 was about the
>>>>>>evaluation function.
>>>>>>
>>>>>>Uri
>>>>>I am basing my oppinion on a small test between Junior 6.0 Vs Rebel Tiger 12e of
>>>>>50 games in 60 minutes per side, using two identical AMD 800 Mhz and Junior 6.0
>>>>>won 28 games; probably the time control and speed of the computers used has a
>>>>>lot to do with the result. It could be that Junior 6.0  calculate deeper than
>>>>>Rebel Tiger 12e, providing faster processor and allowing a time control greater
>>>>>than 60 minutes per side per game.
>>>>>
>>>>>Pichard
>>>>
>>>>It is also possible that Junior6 had a better evaluation.
>>>>You cannot learn from the fact that Junior6 won that it calculated deeper.
>>>>
>>>>Uri
>>>My friend who owned Rebel Tiger 12e knew before the mini match how to set his
>>>program to play with the strongest setting I remember him setting his program to
>>>this setting:
>>>
>>>Rebel Tiger 12e
>>>Permanent Rrain ON
>>>Analysis Brain OFF
>>>
>>>He bought his AMD Athlon 1 week before I did and since my old AMD computer was
>>>too slow a K6 300 Mhz , he recommemded me to buy an AMD ATHLON 800 Mhz similar
>>>to his; I was undicided wether to buy an Intel Pentium III 800 Mhz but just
>>>because of the price and Performance difference and also due to the fact that he
>>>decided to Bet a small fortune of $ 250.00 I agreed to buy it since we are both
>>>programmers interested in the progress of chess softwares. We are facinated
>>>to see how much progress chess softwares have accomplished lately. We would like
>>>sometime to dedicate time to make our own  chess Program, but our dedication to
>>>Microsoft has enable us to program our own chess program, we understand that
>>>chess programs have advance so much, that there is little that could be improve
>>>nowaday except the capability of using Multiprosessors which Deep Junior 6.0 is
>>>capable of using.
>>
>>I do not believe that there is a little that could be improved.
>>
>>I believe that if people develop the right program then even a program on 386
>>can beat kasparov
>>
>>Uri
>Sorry but I disagreed with your logical reasoning. If you compare (Rebel 10,
>Rebel tiger 12e or Deep Junior 6.0) to Deep Blue, any top programmer will assure
>you that these programs have more Chess Knowledge than Deep Blue, But
>when you compare Deep Blue hardware capacity to the most advance PC, even the
>latest AMD Athlon 1000 Mhz just released, you probably need 50 processors like
>the latest AMD Athlon put into parallel just to match Deep Blue caculation
>capacity.



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.