Author: Albert Silver
Date: 06:57:33 03/23/98
Go up one level in this thread
On March 22, 1998 at 21:22:38, Mark Young wrote:
>Had a chance today to play more games on ICC with Fritz 5. Here are the
>new results.
>
>18 games played avg. player rating of 2519.111 Fritz 5 has scored 78%
>for a rating of 2735.336 .
>
>I went into this testing thinking Fritz 5 was not any better then the
>other programs I have, but that was thinking based on the older version
>of Fritz 5. When this mess all started with Fritz 5 and SSDF I bought
>the upgrade to Fritz 5 for retesting. (Upgraded to the version that ssdf
>tested) I am coming to the conclusion that Fritz 5 is clearly better
>then the other programs I have tested. I am coming to this conclusion
>not just because of the results. I am looking at how the program is
>winning. Fritz 5 is not better because it is using some kind of book
>tricks or autoplayer cheats. Fritz 5 is winning because it is out
>playing the other programs in all phases of the game. In my mind we have
>reached the point were the critics of Fritz 5 are going to have to show
>that Fritz 5 is not the best commercal program out instead of just
>saying it.
I don't know. While I disagree with F5's critics that claim that Fritz 5
is a stupid program, I don't think it is the best out there either. I
agree that it is very strong, and is no longer notches below in playing
strength as opposed to other top programs. It continues to be a
tremendously fast searcher, that often renders it an advantage in depth,
and with it's much greater knowledge is really giving other programs a
hard time. It still plays many very strange moves though (as well as
brilliant ones, I'll gladly concede), and when I say strange, I mean
that look awful. It plays much more speculatively, and I often see it
playing odd sacrifices which when it is calculating it will give a
roughly equal eval for example, but right after it plays the move and
the sac is accepted, it's eval will start scraping the ground, and not
because of an extra ply showing it the error of it's ways. I use it
readily for analysis, but when it comes to actually playing a program,
my 3 favourites are MChess Pro 7.1 for it's whacky fun chess, Hiarcs 6
for it's beautiful positional play (which I believe is closely related
to it's understanding of the value of space) and superb endgame play,
and Rebel 9 for it's superb well-rounded play always fighting for the
initiative (not to mention that I can use Rebel with my PC Autoboard).
Still, if Fritz is able to improve it's play in it's next incarnation,
without sacrificing speed, as much as it has in this last, then it will
be a awesome opponent indeed. Of course the same could be said about
others as well: If Hiarcs can improve it's speed somehow, it's tactical
play and it's horrible opening book, it too would be even more fearsome
(I'd probably hesitate to turn on my computer after that ;-) ). And
Rebel? Possibly it's tactics which are irregular as well. Still, Ed is
probably hard at work porting it over to windows so Rebel 10 (or
whatever it is called) may not have so many improvements in it's engine.
>
>Here is an example of what I am saying. Fritz 5 played 2 games vs
>Wchessx. Fritz 5 won 2-0. Wchessx is one of the strongest programs on
>ICC and runs on very fast hardware. Now look at HOW Fritz 5 playing
>black for an even position in the Catalan opening totally out plays
>Wchessx. And in a few moves has itself a won game.
>
>
>[Event "?"]
>[Site "?"]
>[Date "1998.??.??"]
>[Round "?"]
>[White "Wchessx"]
>[Black "Fritz 5"]
>[Result "0-1"]
>[ECO "E06"]
>[Annotator "Young,M"]
>[PlyCount "122"]
>
>{32896kB, f5book.ctg
>} 1. d4 {0} 1... Nf6 {0} 2. Nf3 {6} 2... d5 {0} 3. c4 {6}
>3... e6 {0} 4. g3 {6} 4... Be7 {0} 5. Bg2 {8} 5... O-O {0} 6. Qc2 {6}
>6... c5 {
>0} 7. O-O {7} 7... cxd4 {0} 8. Nxd4 {6} 8... Nc6 {0} 9. Nxc6 {6} 9...
>bxc6 {0}
>10. Rd1 {7} 10... Ba6 {0} 11. cxd5 {67} 11... cxd5 {0.16/11 77} 12. Nc3
>{7}
>12... Bc5 {0.06/9 46} 13. Bf4 {66} 13... Rc8 {0.13/10 45} 14. e3 {13}
>14... Nd7
>{0.00/10 28} 15. Qa4 {64} 15... Bb7 {0.00/11 0} 16. Qb3 {58} 16... Ba8 {
>-0.38/9 22} 17. g4 {61} 17... Qh4 {-0.44/10 42} 18. h3 {60} 18... Nb6 {
>-0.25/10 0} 19. Bf1 {59} 19... f5 {-0.44/10 37} 20. Bg3 {55} 20... Qh6 {
>-0.41/10 0} 21. gxf5 {49} 21... Rxf5 {-0.50/10 34} 22. Qc2 {17} 22...
>Nc4 {
>-0.63/8 21} 23. Kh2 {47} 23... Rcf8 {-0.69/10 0} 24. Bxc4 {55} 24...
>dxc4 {
>-1.22/11 0} 25. Rd2 {46} 25... Rh5 {-1.31/11 0} 26. e4 {32} 26... Rxh3+
>{
>-1.63/10 50} 27. Kg2 {7} 27... Rh5 {-1.84/10 38} 28. Rad1 {20} 28... Qf6
>{
>-2.91/10 9} 29. f4 {35} 29... Be3 {-3.09/10 0} 30. Rh1 {41} 30... Rxh1 {
>-3.53/11 51} 31. Kxh1 {7} 31... Bxd2 {-3.44/11 27} 32. Qxd2 {13} 32...
>Qh6+ {
>-3.94/12 22} 33. Qh2 {42} 33... Qxh2+ {-3.88/15 0} 34. Bxh2 {44} 34...
>Rb8 {
>-3.56/13 0} 35. f5 {45} 35... Rxb2 {-3.72/13 0} 36. fxe6 {40} 36... Rc2
>{
>-3.66/12 10} 37. Be5 {51} 37... Kf8 {-3.75/12 0} 38. Kg1 {50} 38... Ke7
>{
>-4.09/12 45} 39. Bxg7 {8} 39... Kxe6 {-4.09/12 23} 40. Bh8 {19} 40... h5
>{
>-4.41/12 30} 41. Nb5 {37} 41... Kd7 {-4.47/12 0} 42. Nc3 {49} 42... h4 {
>-4.81/14 0} 43. Bg7 {42} 43... Ke6 {-4.63/12 0} 44. Kh1 {57} 44... h3 {
>-4.66/13 58} 45. Bh8 {38} 45... Bc6 {-4.63/13 15} 46. Kg1 {42} 46... Bd7
>{
>-6.16/11 18} 47. Nd5 {50} 47... Rxa2 {-6.88/11 14} 48. Bd4 {18} 48...
>h2+ {
>-7.31/11 24} 49. Kh1 {29} 49... a5 {-7.09/12 10} 50. Nc3 {38} 50... Rc2
>{
>-7.97/12 16} 51. Nd5 {19} 51... Bc6 {-8.28/12 0} 52. Bb6 {30} 52... Bxd5
>{
>-15.47/14 0} 53. exd5+ {27} 53... Kf5 {-11.72/12 1} 54. Bc7 {45} 54...
>a4 {
>-#8/13 0} 55. Bxh2 {49} 55... a3 {-#7/11 5} 56. Bg1 {25} 56... a2 {-#6/9
>0} 57.
>Bd4 {8} 57... c3 {-#5/7 0} 58. Bxc3 {7} 58... Rxc3 {-#4/5 0} 59. Kg2 {6}
>59...
>a1=Q {-#3/3 0} 60. d6 {8} 60... Qb2+ {-#2/1 0} 61. Kf1 {7} 61... Rc1#
>{-#1/1 0}
>0-1
Albert Silver
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.