Author: Don Dailey
Date: 09:24:07 09/08/98
Go up one level in this thread
On September 08, 1998 at 10:34:55, Christophe Theron wrote: >On September 08, 1998 at 07:31:12, odell hall wrote: > >>Could someone give me a technical reason why Chess genius plays so much better >>on 486 machines than all the other programs. I have been pleasantly surprised by >>results against others programs running on much faster hardware (pent pro 200). >>These results have been so surprising that I have decided not to upgrade to a >>pent. Among the victims of genius 5 on 486, is crafty 15.8 running on mmx233, >>Rebel 9 (pent pro 200) and CM555 (pent mmx233) all the games were played at game >>in 30. I realize that statistically i am suppose to win some of the games, but i >>have been winning almost half!!, against much faster machines. However when i >>play my rebel 9 on the 486 it never wins, but draws often. > >Genius is known to have a very good search algorithm, well fitted to slow >computers. In the last decade, Lang's programs were hard to beat because >computers were much slower. > >It seems that today this algorithm has lost its huge superiority. Genius is able >to find very quickly long tactical tricks. Other programs take often a little >bit longer. But given a fast computer and time controls longer than blitz, other >programs generally won't miss those tactical issues. > >This is not to say Genius algorithm sucks. I believe it is still superior, and I >would have been proud to write it myself. But other issues (long term positional >understanding for example) seems to be the real point nowadays, and maybe Genius >is not the best in this regard. > >Anyway, your Genius would be much stronger if you had a Pentium, there is no >possible doubt here. > > > Christophe Hi Christophe, This is a shrewd analysis and I agree in every way. I have always been in awe of the Lang algorithm. He seems to be doing stuff that is fundamentally different that what everyone else does, and it works, quite well. I don't know if it actually is much different, but it really seems to be. In the old days, even the psion program (by Richard Lang) could spit out 7 ply or more principle variations that were quite sound, and this was on pitiful hardware by todays standards. The kind of pruning most of us use seems to need a few ply to really get going, but Richard Langs algorithm didn't seem to care. His program is truly awesome and I have great respect for it. Not to mention its great positional play and endgame understanding. I think it is still fundamentally superior to the other programs. It may not actually be the very strongest currently, but this may be because Richard has not made any substantial effort to stay ahead and I also don't think the book is engineered as well as the top contenders, which could be a big part of the reason his program is not dominant right at the moment. - Don
This page took 0.01 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.