Author: Terry Ripple
Date: 01:17:45 09/22/99
Go up one level in this thread
On September 21, 1999 at 19:36:10, odell hall wrote: >On September 21, 1999 at 19:18:33, Dann Corbit wrote: > >>On September 21, 1999 at 19:12:07, odell hall wrote: >>[snip] >>> I considered this possibility you mentioned, the only thing that makes me >>>doubt that it is the chip is the fact, that when I play hiarcs7. in computer vs >>>computer games on my machine it "destroyes" everything. I beat a friend of mine >>>who has CM6000 in a six game match, this makes me think it is not the chip. >>That's a puzzle. If it "destroys" all the competition, why do you believe that >>their evals (which differ considerably) are better? If they really had a >>superior evaluation at crucial points of the program, wouldn't they win all/most >>of the time? Perhaps Hiarcs has a much greater positional understanding? >> >>I do know that (at least against computers) Hiarcs seems to fare very well. >> >>There may be some particular list of parameters where Hiarcs has a weakness. >>Perhaps if you can specify exactly the conditions of the experiment we can >>figure out what is going on. >>[snip] > > > Hi Dann > > > When I first got hiarcs7.32 Like everyonelse I was quite excited with high >expectations, infact hiarcs6 was one of my favourites, so I expected even more >from hiarcs7. I first noticed this flaky Queensidee castleing bug in hiarcs when >it lost a game against chessmaster, it castled uncessarily in a very dangerous >position, and knowing the tacticaly gifted CM, it went on to slaughter hiarcs, >however I was willing to write this off, as hiarcs began to beat rebel and other >programs, but then I started to see this queenside castling tendency over and >over. I am not saying that hiarcs is not strong, I am just saying that it has >some obvious weakness which needs some attention. It doesn't appear to be as >agressive as hiarcs6, I think that if money were on the line I could draw it >atleast 3 out of ten games. When you refer to "conditions of the experiement" I >am not sure I know what you mean, are you referring to the computer vs computer >games, or the conditions of the game that I played? Time control was 40 moves >in 1 hour, but when I play computer vs computer it is always at game\30. >I looked at the nodes per second, Hiarcs 7.32 is doing 32,000 to 36,000 nodes >per second on My K6-2 350. What's funny is that It destroys other programs, but >barely beats me, Maybe it is just my style of play , I don't know. All I know >is I get slaughtered by all my other programs, and although I have never beaten >hiarcs I have several draws and often have positions where I should have won, >which is never the case when I play Fritz, Chessmaster, Rebel, or Junior. I am >not saying Hiarcs7.32 is a lousy program, I am just saying that I am not >Impressed with it. Hi Odell, To get 3 draws out of 10 games(1.5 points) would give you about a 15% win - ratio, so this would compute to about a 300 point ELO difference between you and Hiarcs! Are you close to a 2250 to 2300 USCF rated player?? Regards,Terry
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.