Author: odell hall
Date: 17:18:07 09/21/99
Go up one level in this thread
On September 21, 1999 at 19:59:50, Ralf Elvsén wrote:
>On September 21, 1999 at 19:16:04, odell hall wrote:
>
>>On September 21, 1999 at 16:14:44, James T. Walker wrote:
>>
>>>On September 21, 1999 at 15:00:14, odell hall wrote:
>>>
>>>>Hi
>>>>
>>>> Has anyone noticed the often bizarre and irresponsible play of Hiarcs7.32?
>>>>It is hard to believe this program is so celebrated here at CCC, the program has
>>>>some obvious flaws that I think any Astute Human master could easily exploit.
>>>>Last night I played it on my AMD k6-2 350 40mb, at 40\1hr, and nearly won, after
>>>>the game I analyzed with fritz5.32, it pointed out atleast five missed wins on
>>>>my part. Some may some, "well what are you talking about you lost didn't you"?,
>>>>True but the point that I am making is that a player of my level should never of
>>>>had hiarcs in such a position (uscf 1804). Analyzing the game with junior,
>>>>fritz, and chessmaster they all made superior moves than hiarcs, and avoiding
>>>>the trouble. If I were a computer chess operator on icc i would not feel
>>>>comfortable playing hiarcs7.32 against titled players in anything over game\30.
>>>>Don't get me wrong I am not basing my accessment on just this one game, but
>>>>many. I think the biggest problem with hiarcs7 is the queenside Castling bug?
>>>>Here is the Game, to illustrate what I am talking about.
>>>>
>>>>
>>>>Event "Match game4"]
>>>>[Site "?"]
>>>>[Date "????.??.??"]
>>>>[Round "?"]
>>>>[White "Hiarcs7.32"]
>>>>[Black "O.hall"]
>>>>[Result "1-0"]
>>>>[ECO "E61"]
>>>>[WhiteElo "2595"]
>>>>[BlackElo "1805"]
>>>>[Annotator "ohall"
>>>>[PlyCount "69"]
>>>>
>>>>{39936kB, super.ctg. PentiumII
>>>>} 1. d4 {0} 1... Nf6 {15} 2. c4 {0} 2... g6 {7}
>>>>3. g3 {0} 3... Bg7 {8} 4. Bg2 {0} 4... O-O {10} 5. Nc3 {0} 5... c6 {9} 6. Qb3 {
>>>>0.58/9 68} 6... d6 {77} 7. Nf3 {0.56/9 20} 7... Nbd7 {41} 8. Bg5 {0.49/9 131}
>>>>8... h6 {26} 9. Bd2 {0.49/9 80} 9... e5 {60} 10. dxe5 {0.42/9 71} 10... Nxe5 {
>>>>73} 11. Nxe5 {0.36/9 59} 11... dxe5 {28} 12. O-O-O {0.34/9 74} 12... Qc7 {40}
>>>>13. Na4 {-0.03/9 232} 13... b5 {73} 14. cxb5 {0.51/9 173} 14... Be6 {39} 15.
>>>>Qc2 {0.94/8 68} 15... Rac8 {50} 16. bxc6 {1.07/8 73} 16... Nd5 {43} 17. Kb1 {
>>>>1.51/9 112} 17... f5 {113} 18. h4 {1.67/8 99} 18... e4 {25} 19. g4 {1.07/9 325}
>>>>19... Rb8 {24} 20. gxf5 {1.90/8 249} 20... gxf5 {21} 21. h5 {0.90/8 511} 21...
>>>>Nb4 {63} 22. Bxb4 {-0.40/9 370} 22... Rxb4 {4} 23. Kc1 {-0.87/8 196} 23... Rc4
>>>>{206} 24. Nc3 {-1.30/9 80} 24... Qxc6 {9} 25. Rdg1 {-1.41/9 94} 25... Kh7 {124}
>>>>26. Rh3 {-0.95/8 62} 26... Bd4 {67} 27. Rd1 {-0.74/7 25} 27... Bxf2 {39} 28. e3
>>>>{-0.54/8 31} 28... f4 {114} 29. Qxf2 {0.00/9 0} 29... Rxc3+ {177} 30. bxc3 {
>>>>-0.04/9 0} 30... Qxc3+ {29} 31. Qc2 {1.59/8 37} 31... Qa1+ {74} 32. Kd2 {
>>>>2.14/8 37} 32... Rd8+ {28} 33. Ke1 {2.60/9 10} 33... Rxd1+ {36} 34. Qxd1 {
>>>>2.60/9 6} 34... Qxa2 {32} 35. Bxe4+ {6.56/8 37} 1-0
>>>
>>>Hello Odell,
>>>All I can say is my Hiarcs 7.32 plays different from yours. Beginning with
>>>8.Bg5 mine wants to play 8.0-0. At 10 it wants to again play 10.0-0. This list
>>>continues with many different moves. I'm not sure what this means. I know that
>>>Hiarcs uses the hash tables different from most programs and my computer is
>>>different from yours but some of your moves are never considered by my Hiarcs
>>>7.32. Position learning could also account for some of this, I don't know. I
>>>agree the way yours played gives call for concern. The only thing I can say is
>>>my Hiarcs plays equal to Fritz and Junior on equal hardware if not slightly
>>>better. It mostly depends on the time control as to which one wins my matches.
>>>But as far as I'm concerned Hiarcs is one of the top programs and I have no
>>>explanation for the way your Hiarcs played.
>>>Jim Walker
>>
>>
>> I am using the hiarcs7.32 engine, which is incorporated into the fritz5.32
>>interface Perhaps this accounts for the discrepancy in moves, what speed is your
>>computer?
>
>I checked the moves on my computer (PIII 450) and except for a very
>small eval difference at move 10 which makes it choose
>another move (maybe the same as James found),
>all moves and evalauations are identical up to move 19, where the
>differences suddenly really start to become frequent. The strangest move
>(leaving the q-side castling aside) is in my opinion 21. h5 .
>Here my version wants to play 21. Rhg1 with an eval of +1.25 .
>
>I wish I could understand the reason for these differences...
>
>Nice game btw. You play the computers more bravely than I do.
>You'll get it next time :)
>
>So you say other programs crush you much more convincingly
>than Hiarcs does? Which is the hardest in your opinion? I was
>too very happy with Hiarcs first, but maybe I want a program that
>beats me 10 - 0 and not 8 - 2 :)
>
>Ralf
Hi Ralf
Yes I totally agree, I was very surprised by the move 21. H5? I have yet to
understand what it purpose was, I was thinking at the time it was one of the
deep positional master moves beyond my understanding? The thing took all day to
make the move (8:31). Fritz and Chessmaster consider 21Qc5!! to be much better
and I agree. Although that line only a computer could play, because it seems
very risky. My opinion the Hardest Program to beat is Fritz5.32 it has rock
solid defenses, I am not making any claims that I beat hiarcs7, It's play is not
just that impressive to me, I think hiarcs6 plays better.
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.