Author: Tomas Casanovas Martinez
Date: 08:37:11 06/24/99
Go up one level in this thread
On June 23, 1999 at 23:06:50, Robert Hyatt wrote: >On June 23, 1999 at 14:19:45, Tomas Casanovas Martinez wrote: > >>On June 22, 1999 at 11:19:33, Robert Hyatt wrote: >> >>>On June 21, 1999 at 14:21:50, paul bedrey wrote: >>> >>>>Hi Tomas and thank you for posting your results. My question is what platform >>>>did you perform your test in? I assume that you used Hiarcs 7.32 and loaded >>>>the Fritz and Junior engines into it. Is this correct? The reason I ask is that >>>>in some prior posts it was questioned if the interface influenced the results. >>>>My second question is have you tried loading Hiarcs 7.32 into the Fritz 5.32 and >>>>Junior 5 modules? If your results are the same then you would have proved the >>>>interface had no influence in the results. I am thinking of buying Hiarcs 7.32 >>>>because it looks like it is the better engine but I know looks can be deceiving. >>>>Thanks again to you and Tanya for posting your results. >>> >>> >>>Another issue: two machines or one? With "pondering" disabled, Crafty will >>>have problems late in the game, because it assumes that it will save some time >>>by pondering, and it uses this time _before_ it actually saves it. As a result >>>it can run low on time near the time control... >> >>Paul, >> >>The platforms where the test were run: >>Hiarcs7.32-Fritz5.32 Hiarcs interface. Pentium 400 >>Hiarcs7.32-Junior5 Junior interface. Pentium 200 >>Hiarcs7.32-Crafty16.6 Fritz interface. Pentium 90 >> >>Hiarcs theoretical advantage was therefore only against Fritz, although I really >>don't believe that one or other opening books are decisive. No one single game >>has been won (advantage of over 1.00)immediately (5 or 6 moves)after the >>opening. >>If you want to buy Hiarcs7.32 I must recommend it. I don't see a better choice >>in the market today. >> >>Bob, >> >>I recognize obviously your authority in this matter. My tests were run in one >>single machine, each working overnight so that it is clear that they are not >>'technically perfect'. What is a bit strange is that Crafty's performance was >>excellent in the first eight games. Afterwards it lost all the games in a raw. I >>will see these games one by one to confirm what happened. >> >>Regards, >> >>Tomas. > > >There are a dozen things that come to mind. Does the engine get restarted for >each game, or is it just sent a 'new' command and expected to reset everything? >I don't test that very well, and it is possible that a "new" will leave things >in a bad state... As you probably know, the Fritz/Hiarcs/Junior environment allows to play one game after another non-stop for days,weeks or months, changing the opening every two games. This is the only way I can test the programmes because at this time of the year I am very bussy marking exams of my students :-) . This means that the engine is not restarted at all after each game. Probably this is the reason of Crafty's very poor performance in this test. I will let you know if I discover any other reason for it. Tomas.
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.