Computer Chess Club Archives


Search

Terms

Messages

Subject: CSTal2.03 vs. Hiarcs7.32 so far

Author: Thorsten Czub

Date: 09:18:40 09/15/99


Hi - 4 games have been manually played so far between
cstal 2.03 and hiarcs7.32.

The 5th game is going on and is a very interesting attack
game where you can see very good the difference between
finding (using search = hiarcs) and dreaming-about (using
evaluations = cstal).
Hiarcs was out of book in move 11. playing f5 as the first
move it had to compute.
In this game CSTal TRIES to attack hiarcs, and hiarcs more or less
tries to find out ways not to lose. Within its search horizont
hiarcs manages to find a way out of the attack, but out of
the horizont it cannot stop tal from making the attacks.
Hiarcs evaluations are precise within its horizont.
Cstal evaluations are unprecise , but they don't need to be
precise, hiarcs has to show that it is dreaming, not cstal.
I have no idea who will win, but it is very interesting to see
both programs evaluate the position very different (e.g. in
the 22nd move hiarcs says +0.29 against it, meanwhile cstal
says +2.90 for itself.
Now Hiarcs says +0.54 against it. Why was it unable to see this
some plies before ?? Because NOW it gets into its horizont.
But CSTal was able to "guess" the consequences. Because CSTal
is not trying to compute anything PRECISELY.
After 2 Minutes Hiarcs now says +0.72 against itself having the
following mainline:
+(0.72) 22...Re7 23.Re3 b5 24.Rh3 Nf5 25.Nxh7 Nxh4 26.Nxf8 g5 27.Be4

The one program trusts its search, the other program trusts into
its evaluation. The one has a precise evaluation and a very
limited search-horizont, the other program has a rough evaluation
but a kind of unlimited search-window.


After 6 minutes hiarcs now says:
+(0.80) 22...Re7 23.Re3 Qg7 24.c3 Qf8 25.Rh3 Nf5 26.Nxh7 Nxh4 27.Nxf8
g5 and moves.

CStal instantly moves Re3. It is now again Hiarcs to play.
The evaluation decreases into +1.74 for white after having
seen that Qg7 c3 Nf7 Nxh7 Nh6 Ng5 Nf7 f4 Nh8 leads into
nothing.
It than finds Nf5, this is the ponder-move of CSTal (CStal
evaluates this move >+3.97 against hiarcs).
The next branch hiarcs considers long time is b5.
It seems hiarcs has problems to find a better move than Qg7.
It likes b5 and gets +1.72 score against it saying
23...b5 24.Rh3 Nf5 25.Nxh7 Nxh4 26.Nxf8 g5 27.Nh7 Nf5
28.Nxg5 b4 as the best main-line.

For another long time hiarcs considers about b6-branch 17/25.
Is this move a better choice ?
Really it plays b6 and gets only +1.03 against itself.

CSTals score jumps to +3.65 in the seventh search.
+3.42 in the 8th search. First it tries to play rh3, than
it changes into the very aggressive g4 !
Hiarcs pondered about rh3 and has to compute again.
It now scores +1.70 against itself. Hiarcs has now
incredible main-lines full of fire.
I am very nervous because i don't know how this continues.
E.g. hiarcs says:
24...Qf4 25.Qxh6 Ba6 26.Rf3 Qxg4+ 27.Rg3 Qd1+ 28.Kg2 Bxd3
29.cxd3 Qa1 30.Rh3 Qg7 31.Qh4 h5 and evaluates this +1.85
against itself.
The main problem for me to operate the programs is that
i have problems to stand the game. CStals main-lines and
evaluations are mainly unprecise so i look the whole time
on hiarcs7.32 screen, watching its evaluations.
26th move, hiarcs says Rg3 would be the best for white,
but cstal plays rc3. Hiarcs score drops my heart-beat
increases.
But hiarcs has problems to come deeper. it seems to have so many
extensions to follow that it cannot come out of the 8th search
with 25 plies selective. CStal expects Qxg4 and says +5.00.
28th move:
Hiarcs has not much time. Its clock says 32 minutes to go.
CSTals clock says 1h4' left. But the last moves were forced.
And the next moves too. So maybe it finds a way out without
having time trouble.
It now says +1.57 for white considering about the easy Bb7+.
CStal expects this and says +4.53 for itself.
We will see how the time-problem brings hiarcs into another
trouble in a few moves.
29th move: hiarcs considers about Bxe4+. It now evaluates +1.86
against itself and has 23' to move.
CStal expects this move and says +5.97.
The hiarcs mainline says 29... Bxe4+ 30.Nxe4 Qd5 31.Qh4 Rg7
32.Rcf3 Kh8 33.Nf6 Qb7 34.Rh3 with +1.86 against itself.
In an instant cstal takes back with the knight. again hiarcs
has to move. 30st move, and only 19' to go.
Both say Qd5 followed by Qh4 looks like the right thing.
Hiarcs score decreases into +2.11 for white.
After 5 minutes hiarcs runs through the 10 ply and the
score decreases again into +2.47 for cstal-side.
Kh8 (branch 5 from 34) looks again interesting but
there was not enough time to consider about it.
Hiarcs has only 11' to compute for the next 10 moves
and again cstal has pondered about Qd5 and moves
in an instant.
Hiarcs plays Rg7 very fast (only 1 minutes used for the
move) and cstal pondered Rf7. So now it is cstal computing
much having still 1h3' on the clock.
Hiarcs ponders Rcf3 and Cstal wants to play Rgf3.
Hiarcs clock says 10'46" awaiting the 32.move.
CStal likes Rgf3 and gives +6.16 pawns for it :-))
Because hiarcs now says +2.84 for white, my heartbeat
decreases. I am very confident that hiarcs will lose
this game.
Cstal plays Rgf3 and hiarcs has to compute again.
33rd move: Hiarcs has only 5'15" left. CSTal considers
about Kg1 and changes in the last minute
into Qd4 the move hiarcs ponders about.
Instantly hiarcs moves b5. So still 5'15" left for 6
moves.
Hiarcs expectes white to play Qd6 (+3.38 for white)
but CStal seems to play Ne4 to bring the f-rook on
the 7th rank (the g7-rook is pinned ).
Therefore hiarcs has to consider about the defense.
With 4'21" left hiarcs plays Qd5.
CSTals score says +7.16 and wants to exchange queens
and feels very happy.
This move is expected by hiarcs so again hiarcs will
not lose much time for taking back on d5 when Qxd5
happens.
CSTal has still enough time on its clock (38').


[Event "k6-400"]
[Site "?"]
[Date "1999.09.05"]
[Round "1"]
[White "CSTal2.03"]
[Black "Hiarcs7.32"]
[ECO "C49"]
[Result "1-0"]

1. e4 Nf6 2. Nc3 e5 3. Nf3 Nc6 4. Bb5 Bb4 5. Nd5 Nxd5 6. a3
Nf4 7. g3 Ba5 8. gxf4 exf4 9. b4 Bb6 10. Bb2 f6 11. d4 Qe7
12. Qe2 a5 13. c3 f5 14. exf5 Qxe2+ 15. Kxe2 O-O 16. Bc4+
Kh8 17. Kd2 axb4 18. axb4 Rxa1 19. Rxa1 Ne7 20. b5 Nxf5
21. Ra8 g6 22. Bb3 Kg7 23. c4 c6 24. Kc1 cxb5 25. c5 Bc7
26. d5+ Kh6 27. d6 Nxd6 28. Be5 Ne4 29. Bxc7 Nxc5 30. Bc2
Ne6 31. Bd6 Re8 32. Bd3 Kg7 33. h4 b4 34. Kd2 Kh6 35. Ra4
b3 36. Rb4 b2 37. Rxb2 Kh5 38. Rb1 Kg4 39. Ne5+ Kh5
40. Be2+ Kh6 41. Ng4+ Kg7 42. Be5+ Kf7 43. Nf6 Rd8 44. Nxh7
d6 45. Bf6 Rd7 46. h5 Nc5 47. Rb4 d5 48. Rxf4 Ne4+ 49. Kd3
gxh5 50. Bxh5+ Kg8 51. Ng5 Nxf6 52. Rxf6 Rc7 53. Bf7+ Kg7
54. Rg6+ Kf8 55. Bxd5 Bf5+ 56. Kd4 Bxg6 57. Ne6+ Ke7
58. Nxc7 b6 59. Bc4 Kd6 60. Nd5 Bh5 61. Nxb6 Kc6 62. Nd5
Bf3 63. Ke5 1-0

[Event "k6-400 "]
[Site "?"]
[Date "1999.09.05"]
[Round "2"]
[White "Hiarcs7.32"]
[Black "CSTal2.03"]
[ECO "A08"]
[Result "1/2-1/2"]

1. e4 c5 2. Nf3 e6 3. d3 Nc6 4. g3 d5 5. Nbd2 Nf6 6. Bg2
Be7 7. O-O O-O 8. Re1 b5 9. a4 b4 10. exd5 exd5 11. d4 Bg4
12. dxc5 Bxc5 13. Nb3 Bb6 14. Bg5 a5 15. Qd3 Rc8 16. h3 Be6
17. g4 Re8 18. Qb5 Qc7 19. Bxf6 gxf6 20. Re2 Bd7 21. Rxe8+
Bxe8 22. Qe2 Qf4 23. Qa6 Rb8 24. Qe2 Ne5 25. Ne1 Rc8
26. Qd2 Ng6 27. Qxf4 Nxf4 28. Bf1 f5 29. Nd3 Ne6 30. Ne1
fxg4 31. hxg4 Bd7 32. Bb5 Bxb5 33. axb5 Rc4 34. Nxa5 Rxg4+
35. Kf1 Rh4 36. Nc6 Kg7 37. Ra6 Nf4 38. Kg1 Rg4+ 39. Kf1
1/2-1/2

[Event "k6-400"]
[Site "?"]
[Date "1999.09.07"]
[Round "3"]
[White "CSTal2.03"]
[Black "Hiarcs7.32"]
[ECO "B77"]
[Result "0-1"]

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 g6 6. Be3
Bg7 7. f3 O-O 8. Qd2 Nc6 9. Bc4 Nxd4 10. Bxd4 Bd7 11. O-O
Rc8 12. Bd5 Qc7 13. a3 b6 14. Rfd1 Nh5 15. Bxg7 Nxg7
16. Qh6 Qc5+ 17. Kh1 e6 18. Rd2 f5 19. Bb3 Bc6 20. Re1 Rc7
21. exf5 Qxf5 22. Qe3 d5 23. Ne2 Qf6 24. c3 Rcc8 25. Nd4
Bd7 26. Rdd1 a5 27. Qd3 a4 28. Ba2 Rc5 29. Qa6 Rb8 30. Re3
Nf5 31. Nxf5 Qxf5 32. Ree1 Qf6 33. Qe2 Rf8 34. Rd4 Rcc8
35. Qe3 b5 36. Rb4 Qg7 37. Rh4 Rf7 38. Qg5 Rcf8 39. Bb1 Qf6
40. Qg3 Qe7 41. Rg4 Rf6 42. Qe5 Qf7 43. Rg3 Qg7 44. Qd6
R8f7 45. Qb8+ Qf8 46. Qa7 Qd6 47. Rg5 Qf4 48. Rge5 Qc4
49. Qb8+ Rf8 50. Qd6 Bc8 51. Bc2 R6f7 52. R5e3 Qc7 53. Qb4
Bd7 54. Qh4 Qc4 55. Qf2 Kg7 56. Rd3 Qc7 57. Rd4 Rf4 58. Qg3
Qd6 59. Red1 Bc6 60. Qg5 e5 61. Rxf4 Rxf4 62. Qg3 Qf8
63. Rg1 e4 64. fxe4 dxe4 65. Qe3 Qf6 66. h3 g5 67. Qa7+ Kh6
68. Qe3 Rf1 69. Rxf1 Qxf1+ 70. Qg1 Qxg1+ 71. Kxg1 g4
72. hxg4 Kg5 73. Bd1 Kf4 74. Kf1 Ke3 75. Ke1 Kd3 76. g5 Be8
77. g3 Bd7 78. Bh5 Kc2 79. g6 hxg6 80. Bxg6 Kxb2 81. Kd2
Kxa3 0-1

[Event "k6-400"]
[Site "?"]
[Date "1999.09.15"]
[Round "4"]
[White "Hiarcs7.32"]
[Black "CSTal2.03"]
[ECO "B45"]
[Result "1/2-1/2"]

1. e4 c5 2. Nc3 Nc6 3. Nge2 e6 4. d4 cxd4 5. Nxd4 Nf6
6. Nxc6 bxc6 7. e5 Nd5 8. Ne4 Qc7 9. f4 Qb6 10. Bd3 Ba6
11. Bxa6 Qxa6 12. Qe2 Qxe2+ 13. Kxe2 f5 14. exf6 Nxf6
15. Nxf6+ gxf6 16. Be3 d5 17. g3 Rg8 18. Rad1 Bd6 19. b3 a6
20. c4 O-O-O 21. f5 dxc4 22. Rc1 Rg4 23. fxe6 Re8 24. Rxc4
Rxc4 25. bxc4 Rxe6 26. Kd3 Bc7 27. Rb1 Rd6+ 28. Ke2 Re6
29. a4 f5 30. Kd3 Rd6+ 31. Kc3 Re6 32. Kd4 Kd7 33. Rf1 Re4+
34. Kd3 Ke6 35. Rb1 Kd7 36. Rf1 Ke6 37. Rb1 Kd7 38. Rb7 Kc8
39. Rb2 Kd7 40. Rb7 Kc8 41. Rb2 Kd7 42. a5 Kc8 43. Bb6 Re5
44. Rb1 Bd6 45. Rf1 Kd7 46. Ra1 Bb4 47. Ra2 Re1 48. Rf2
Rd1+ 49. Kc2 Rh1 50. Kb3 Rb1+ 51. Ka2 Rh1 52. Kb2 Ke6
53. Re2+ Kd6 54. Kb3 c5 55. Rf2 Ke5 56. Bc7+ Ke6 57. Bf4
Rb1+ 58. Ka2 Re1 59. Kb3 Rb1+ 60. Kc2 Re1 61. Kb3 Rb1+
1/2-1/2

[Event "k6-400"]
[Site "?"]
[Date "1999.09.15"]
[Round "5"]
[White "CSTal2.03"]
[Black "Hiarcs7.32"]
[ECO "B40"]
[Result "*"]

1. e4 c5 2. Nc3 e6 3. Nf3 Nf6 4. d4 cxd4 5. Nxd4 Bb4 6. e5
Nd5 7. Bd2 Nxc3 8. bxc3 Ba5 9. Qg4 O-O 10. Bd3 Nc6 11. Nf3
f5 12. exf6 Qxf6 13. O-O g6 14. Bh6 Bxc3 15. Rab1 Re8
16. Rb3 Ba5 17. Ng5 Ne5 18. Qg3 Bc7 19. Rc3 Nf7 20. Rxc7
Nxh6 21. Qh4 Qf8 22. Re1 Re7 23. Re3 b6 24. g4 Qf4 25.
Qxh6 Ba6 26. Rc3 Qxg4+ 27. Rg3 Qd1+ 28. Kg2 Bb7+ 29. Be4
Bxe4+ 30. Nxe4 Qd5 31. Qh4 Rg7 32. Rgf3 Kh8 33. Nf6 Qb7
34. Qd4 b5 35.Ne4 Qd5 36. Qxd5 exd5 37. *



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.