Author: Uri Blass
Date: 08:39:03 09/03/04
Go up one level in this thread
On September 03, 2004 at 11:20:37, Ed Schröder wrote: >On September 03, 2004 at 03:25:35, Frank Quisinsky wrote: > >>On September 02, 2004 at 20:19:39, Ed Schröder wrote: >> >>>On September 02, 2004 at 17:53:20, Frank Quisinsky wrote: >>> >>>>On September 02, 2004 at 17:28:18, Ed Schröder wrote: >>>> >>>>Hi, >>>> >>>>thanks Ed! >>>> >>>>Your words are nice to read for me: >>>> >>>>- A little bit against my bad conscience >>>> (Leiden 2000, or 2001). >>>> >>>>- mail to your beta testers in test time of Gandalf 5.1 >>>> >>>>But nobady is perfect! >>>>In this time I made to many things and loosed in a lot of cases the feeling for >>>>the situation. >>> >>>Okay Frank, let's forget about the past and make a fresh new start, will you? >>> >>>My best to you, >>> >>>Ed >> >>Hi Ed, >> >>yes, that would be great! >> >>I have a little question Ed! >>Not the right thread but important for me. >> >>Since over one year I come back to my first years of computer chess. I try to >>reproduce my collection of favorite chess computers and played a lot games in my >>free time this year. I like the Mephisto chassi of Mephisto Milano, Berlin, >>Modena, Atlanta and have here all six types. My favorite is Mephisto Milano >>(like the playing style and the display information). >> >>The question I have: >>You created at last the following programs: >> >>Mephisto MMV (very tactical) >>Mephisto Milano (combination tactic and positional playing style) >>Mephisto Polgar (positional playing style) >> >>After all my tests for many years and in this times :-) I believe the best one >>is the Mephisto Milano but not in the SSDF. This my opinion about it and it >>would nice to know what you are thinking about it. I believe the way you go with >>the program style in Mephisto Milano is the way you search for your first "PC" >>chess programs. Your experiments with GM human games in the following years are >>a second good example. Do you think that the Mephisto Milano is the strongest of >>this three chess computers? Is this right that you try to find out the >>combination of positional and tactic playing style in your first PC chess >>program up to today in Rebel / ProDeo? With time we can see that the combinated >>engines are stronger with the tactical only engines. The hardware is faster and >>in the first years of computer chess the tactical programs have an advantage. So >>it's a little bit a sensation that the Milano and not the MMV is stronger chess >>computer (in my opinion). >> >>Your answer will be great for a lot of chess computer fans I know becuase this >>question is a long time questions if we played our tournaments. Have this >>discuss with my chess friend each time we play our tourneys. Perhaps you find >>the time for an answer. > >I would rate the Milano as the best but between the 3 of them is not much >difference in strength, not more than 30 maximum 50 elo. The problem was that >since the Polgar I already had reached the limitations of the 8-bit processor >and its limited RAM and ROM and thus no big improvements could be expected. > >Requests from me to H&G for more RAM and speed were declined and so they more or >less forced me to find my own way and move on to a new processor instead, this >against the express wish of H&G. My choice fell on the Archimedes RISC chip >which later was produced by H&G after all (known as the Mephisto RISC) after a >lot of juridical tug-of-war. > > >>Today we have the discuss that Rebel can be stronger with "only" more tactic >>(opinion from my chess friend, not my). > >Tell your friend he is wrong, nowadays there is very little to gain with SEARCH >(in comp-comp) because even nowadays so-called amateur programs are real good in >that. Today the top programs win their games because of better chess knowledge, >this shift started 2-3 years ago. > >My best, > >Ed I do not agree. If you can do something in the search that is equivalent to being 2 times faster you still earn significant rating. Not as much as in the past but still significant. I know that in the world championbship Movei did mistakes that it could avoid if it only could search deeper. In the game against isichess it probably could get at least a draw by searching one ply deeper and in the game against Crafty it could probably also get a draw with searching one ply deeper. In the games against Fritz and Diep I do not know if searching deeper could change the final result but it could at least avoid tactical mistakes and after avoiding them there is a game to play. Uri
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.