Author: Melvin S. Schwartz
Date: 21:30:10 07/20/99
Go up one level in this thread
On July 19, 1999 at 22:29:25, Mark Young wrote: >On July 19, 1999 at 21:23:32, Melvin S. Schwartz wrote: > >> >>Mark, I don't know what time control John is using but I played Hiarcs at 40/2 >>in a number of games and I've seen some real bad moves. In at least one game it >>overlooked a simple pawn advance and went from about a +1.50 to about even in >>one move. The game became a dead draw when Hiarcs should have won. I gave the >>critical position to both Fritz 5.32 and Rebel 10 and they both made the proper >>move which would have kept the score over a pawn advantage and with excellent >>winning chances. >> >>Another problem I believe is the evaluation score by Hiarcs. In this regard, I >>think Rebel 10 is more accurate. In fact, I have a game going now where Hiarcs >>thinks it has a plus score where I couldn't see it and so gave it to both Fritz >>and Rebel who scored it a plus for me - just coincidence? Maybe yes? Maybe >>Hiarcs knows something the programs don't - I will see what develops. Remember, >>these are games at 40/2. >> >>Regards, >>Mel > > >I understand what you are saying, but the same can be said about the other top >programs in some games. The positions you mention are of interest, and please >post them. What I find more importent is the win loss ___________ Mark, when I get a chance, I will post a few examples for you. Now please don't get me wrong, I am not saying that Hiarcs 7.32 is a bad program. It's just that I along with some others expected better. Perhaps we expected too much. Nevertheless, the program in my opinion is not head and shoulders above Fritz 5.32 for example. Soon there will be Shredder 4, Fritz 6, Junior 6, Ferret? a new Rebel, a Tiger 11 or 12, etc, and then there will certainly be lots more to cloud the issue of what is the strongest program. And there will probably even be a new Hiarcs with I hope a better opening book. Regards, Mel percentage against the >pool of player you are testing against, not how a program won or lost in this >game or that game. > >This reminds me of a short time ago when Fritz 5.0 was top dog and killing >everthing around. In Fritz 5's case people dismissed Fritz 5 because yes it won, >but in many peoples opinion it won ugly and/or it misevaluated this postion or >that. Fritz 5 and Fritz 5.32 play in the same style today as it did yesterday, >but now more people understand I think that it my not play in a "human style" >but it plays very sound chess never the less. It is very hard to argue against a >strong winning record of games against strong humans and computers players. > >When I see a losing win/loss record and a performence worse then Hiarcs 7.0, I >will agree that Hiarcs 7.32 is broken. As far as I know no one thinks that >Hiarcs 7.0 is broken or weak. > >It maybe of interest to see how well Hiarcs7.32 performs against Hiarcs 7.0. If >something is broken this may be a good way to spot it, as I find they play and >have the same eval and move selection in most cases.
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.