Computer Chess Club Archives


Search

Terms

Messages

Subject: A. STEEN vs. FRUIT 2.2.1 {Posted at request of Graham Banks & M. Monge}

Author: A. Steen

Date: 05:54:54 11/24/05


To the insolent trolls and abusive patzers - please first read:
http://www.talkchess.com/forums/1/message.html?464028
before flaming, thanks.  There, I was specifically requested by two members to
post at least one game.

We are all patzers at chess.  Just look at the optimal moves in the recently
calculated 7-man tablebases, see how incomprehensible they are and how hard it
is to differentiate between the winning, drawing and losing moves, and be
humble. Just like me.

The purpose of agreeing to Graham's request is to show how easy it is to beat
Fruit.  It walks punch drunk into losses in ways the Chessbase top engines do
not.  Even though it beats those same engines in head-to-head play!

I choose an exciting-looking (only in looks!) game from a controversial start,
to make it interesting for many readers. I was playing when in a relaxed,
confident frame of mind, and the annotations are similarly light-hearted.

Another factor influencing my selection of this game is that there are one or
two disturbing aspects to Fruit's play in this game, which may be of assistance
to the developer team to identify.

Thirdly, the whole thing was quite easy for me to play, as my attack almost
played itself some of the time, and the tightrope was easy to discern when it
was there.  I still had quite a lot of time on the clock at the end of the game.
This illustrates my motif - Fruit is much weaker against reasonable humans than
it is against other engines.   Most players here should be able to find most or
all of my moves, and so should most chess engines (but then why don't these
engines score better against Fruit?).

Fourthly, given the above the game is certainly no showcase for my abilities.
So I cannot be again accused by the trolls of boasting or whatever.

Fifthly, the game illustrates how one can boldly play against the "best"
computers, rather than indulge in "Father's" abusive shufflings and seek closed
positions.

Below is the whole game, but please scroll down for some light illustrated
commentary, aimed at the 1500-rated player.  As every one of my chessic
_critics_ on this board have demonstrated no more understanding than that of a
1500-player (if anything, I am being a little too generous here), that is an
appropriate level.  Also, as this is not a deep game by any means - there being
few strategic concepts within it - such a light dusting isn't inappropriate.

Lastly, since I would normally never comment on such a casual game, hopefully my
style of commentary won't reveal too much about who I might or might not be.

[Event "Event"]
[Date "2005.11.??"]
[Round "1"]
[White "A. Steen"]
[Black "Fruit 2.2.1"]
[Result "1-0"]
[PlyCount "171"]
[TimeControl "900"]

1. e4 Nc6 2. f4 Nf6 3. e5 Nd5 4. d4 d6 5. c4 Nb6 6. d5 Nb4 7. a3 Na6 8. Nf3 g6
9. Nc3 Bg7 10. b4 Bg4 11. c5 dxe5 12. Qb3 Bxf3 13. gxf3 Nxd5 14. Qxd5 e4 15.
Bb5+ c6 16. Bxc6+ bxc6 17. Qxc6+ Kf8 18. Bb2 Nxc5 19. Qxc5 Qd3 20. Na4 Bxb2 21.
Nxb2 Qxf3 22. Rf1 Qh3 23. Qd4 Rg8 24. Nc4 g5 25. Ra2 Rc8 26. Re2 a5 27. Qxe4
Qc3+ 28. Nd2 Qc1+ 29. Kf2 Qc7 30. Nb3 gxf4 31. Nc5 Rg6 32. Ke1 axb4 33. axb4 Re8
34. Qe5 Qxe5 35. Nd7+ Kg7 36. Nxe5 Rf6 37. Re4 Ra8 38. Rexf4 Rxf4 39. Rxf4 f6
40. Nc6 e5 41. Rg4+ Kh6 42. b5 Ra1+ 43. Ke2 Ra2+ 44. Ke3 Ra3+ 45. Ke4 Rb3 46.
Na7 Kh5 47. Rg3 Rb2 48. Kf5 Rf2+ 49. Ke6 Rb2 50. Rg7 e4 51. Kf5 Kh6 52. Rg3 Rxh2
53. Rb3 e3 54. Rxe3 Rf2+ 55. Ke6 Kg5 56. Rb3 Re2+ 57. Kd7 Rd2+ 58. Kc8 Rd6 59.
Nc6 f5 60. Kc7 Rf6 61. b6 Rf7+ 62. Kd8 h5 63. b7 h4 64. b8=Q Rf8+ 65. Ke7 Rxb8
66. Rxb8 Kf4 67. Kf6 h3 68. Rh8 Kg3 69. Nd4 f4 70. Ne2+ Kf3 71. Nxf4 Kxf4 72.
Rxh3 Ke4 73. Rh4+ Ke3 74. Kf5 Kd3 75. Ke5 Ke2 76. Kd4 Kf3 77. Kd3 Kg3 78. Ra4
Kf3 79. Kd2 Kf2 80. Ra3 Kg1 81. Rf3 Kh1 82. Ke2 Kg2 83. Ra3 Kg1 84. Kf3 Kh1 85.
Kg3 Kg1 86. Ra1# 1-0

Settings for Fruit 2.2.1
Pondering: ON, of course.
Book: Fruit_ASv3.ctg, which I am told is the best one for Fruit 2.2.x.  Learning
had been on and I had previously been trying to persuade the computer to play
Alekhine's Defence, which it would not.  It prefers 1. .. N-c6 to 1. .. N-f6,
which is noteworthy. Of course, I take it out of book as soon as possible so all
that is moot. :)
CPU: Intel running at 3.8GHz, more RAM than you would imagine
Hash Table: 1024 Mb
TableBases: About 300 Gb
Engine Parameters: All standard, and "OwnBook" ticked and present
Other settings: Resign OFF.
I could not see Fruit's evaluation or the bar graph, of course. And I did not
takeback any of my moves.


1. e4 Nc6 2. f4 Nf6 3. e5 Nd5 4. d4

[D]r1bqkb1r/pppppppp/2n5/3nP3/3P1P2/8/PPP3PP/RNBQKBNR b KQkq d3 0 4

At this point Fruit shocked me by playing 4. .. g5, which is not a sensible
move. Maybe this was a bug or a result of something that was in the hash.  I
switched on the appropriate pane, used takeback and pressed "+" to get MV mode,
and made Fruit play its second choice here, 4. .. d6.  Given more time, Fruit
switches to 4. .. e6, which is sound.  It is a little disturbing that the weaker
moves were considered.  I am doing Fruit a favour by not holding it to 4. .. g5,
or the game would have been Short Shrift. :)

d6 5. c4 Nb6 6. d5 Nb4 7. a3 Na6

[D]r1bqkb1r/ppp1pppp/nn1p4/3PP3/2P2P2/P7/1P4PP/RNBQKBNR w KQkq - 0 8

I consider this position winning for white.

** THE POINT IS, AGAINST FRITZ 9 OR SHREDDER 9.1 UCI I WOULD HAVE HAD TO FIGHT
FOR THE WIN FROM HERE ONWARDS.  FRUIT IS JUST TOO DECENT AND HONEST! :)

8. Nf3 g6 9. Nc3 Bg7

[D]r1bqk2r/ppp1ppbp/nn1p2p1/3PP3/2P2P2/P1N2N2/1P4PP/R1BQKB1R w KQkq - 0 10

I choose the most violent of the many alternatives here.

10. b4 Bg4 11. c5 dxe5 12. Qb3 Bxf3 13. gxf3

[D]r2qk2r/ppp1ppbp/nn4p1/2PPp3/1P3P2/PQN2P2/7P/R1B1KB1R b KQkq - 0 13

If 13. .. N-d7, 14. c6  1-0
If 13. .. N-c8, 14. B-b5+  1-0
I know you saw those :)

13. .. Nxd5! 14. Qxd5

[D]r2qk2r/ppp1ppbp/n5p1/2PQp3/1P3P2/P1N2P2/7P/R1B1KB1R b KQkq - 0 14

Here I was expecting 14. .. exf, at which point I would have traded queens and
gone for an easy endgame win.  Fruit's move (14. .. e4) is probably an
improvement.  Investigating the position after the game, Fruit expected 14. ..
e4 to be answered by 15. Q-d2 (weird!), which shows it has no understanding of
the dynamical potentialities of the position.

14. .. e4

Hereafter, I am seduced into attacking as wildly as I felt I could.  Perhaps
Fruit is a psychologist.

[D]r2qk2r/ppp1ppbp/n5p1/2PQ4/1P2pP2/P1N2P2/7P/R1B1KB1R w KQkq - 0 15

I hope you can find my move here.  However much time Fruit seems to be given in
post-mortem analysis, it does not find it for white (it oscillates from 15. Q-d2
to 15. Qxd8+).

15. Bb5+

I would have thought this move was totally obvious.

15. .. c6 16. Bxc6+

[D]r2qk2r/pp2ppbp/n1B3p1/2PQ4/1P2pP2/P1N2P2/7P/R1B1K2R b KQkq - 0 16

Here I expected 16. .. K-f8, against which 17. B-b2 is forced.  I have no idea
whether 16. .. K-f8 is better than the move Fruit played, which was to grab the
bishop.  Fruit's king safety algorithm or weighting may need review.

16. .. bxc6 17. Qxc6+ Kf8

[D]r2q1k1r/p3ppbp/n1Q3p1/2P5/1P2pP2/P1N2P2/7P/R1B1K2R w KQ - 0 18

Of course I cannot touch the black knight.

18. Bb2

[D]r2q1k1r/p3ppbp/n1Q3p1/2P5/1P2pP2/P1N2P2/1B5P/R3K2R b KQ - 0 18

What black now did really woke me up.  I had missed this entirely, and was
expecting 18. .. exf, which pawn I would doubtless pick up in due course along a
comfortable cruise to a win, view to a kill, etc.

18. .. Nxc5!

[D]r2q1k1r/p3ppbp/2Q3p1/2n5/1P2pP2/P1N2P2/1B5P/R3K2R w KQ - 0 19

To decline the sac would cost me the game.  The point of black's move is to
deflect my Q from the long diagonal, where it paralyses black, and to gain
dynamical counterchances.  Unfortunately for black, my moves are easy to find.

19. Qxc5 Qd3

[D]r4k1r/p3ppbp/6p1/2Q5/1P2pP2/P1Nq1P2/1B5P/R3K2R w KQ - 0 20

Post-mortem analysis reveals that Fruit expected 20. R-d1 or 20. N-d1 here (I
cannot tell which), and spent its ponder time analysing what followed one of
those.  This is the only move I spent more than a few seconds on, certainly
considering both those moves and 20. N-d5?, which might have been my choice
against a weak non-computer opponent and just for the hell of it.

At the end it was a toss-up between 20. N-d1 and what I played, and it would
have made no difference which one I chose.

20. Na4 Bxb2 21. Nxb2 Qxf3 22. Rf1 Qh3 23. Qd4

[D]r4k1r/p3pp1p/6p1/8/1P1QpP2/P6q/1N5P/R3KR2 b Q - 0 23

WQ is boss-lady. :)

23. .. Rg8 24. Nc4

[D]r4kr1/p3p2p/6p1/5p2/1PNQpP2/P6q/7P/R3KR2 w Q f6 0 25

Here I expected Fruit to play 24. .. f5 to hold the pawn, giving me time to
activate my rooks, eject the black queen and advance my Q-side pawn majority,
get an isolated passed pawn (probably on the b-file) and the rest blindfold.

Fruit, however, found something a little more interesting and probably better,
though it is all academic.

24. .. g5 25. Ra2 Rc8

[D]r4kr1/p3p2p/6p1/5p2/1PNQpP2/P6q/7P/R3KR2 w Q f6 0 25

It is utterly obvious to state that if white were to blunder here with 26.
Qxa7??, Black would win with 26. .. Rxc4 picking up the hanging knight. What is
a little less obvious is that black's attack would be absolutely overwhelming,
and white must then sacrifice his queen in the next couple of moves to avoid
checkmate (I've just checked this by computer).

26. Re2 a5

[D]2r2kr1/4pp1p/8/p5p1/1PNQpP2/P6q/4R2P/4KR2 w - a6 0 27

On post-mortem, Fruit insists here that PxRP is best for white, which is
ridiculous.

27. Qxe4 Qc3+ 28. Nd2 Qc1+ 29. Kf2 Qc7 30. Nb3 gxf4 31. Nc5 Rg6

[D]2r2k2/2q1pp1p/6r1/p1N5/1P2Qp2/P7/4RK1P/5R2 w - - 0 32

At this point I momentarily and irrationally decided to kill the f4 pawn (on the
grounds that it might later give black something to help him with a perpetual?)

Of course, correct was 32. Q-e5 forcing the exchange of queems; I never even saw
it until a few moves later.  Black should not have permitted a Q exchange, but
that's a common computer failing in lost endgames.  Once the Qs are out, all
chances of counterplay disappear.

32. Ke1 axb4 33. axb4 Re8 34. Qe5 Qxe5 35. Nd7+

[D]4rk2/3Npp1p/6r1/4q3/1P3p2/8/4R2P/4KR2 b - - 0 35

The game is now totally over, but my cruel no-resign setting gave me the
satisfaction of checkmating Fruit. No more annotation is needed from this move
onwards, and I concede in advance that I am certain I fouled up in the endgame
and could have mated Fruit 10 moves earlier. :)

32. .. Kg7 36. Nxe5 Rf6 37. Re4 Ra8 38. Rexf4 Rxf4 39. Rxf4 f6 40. Nc6 e5 41.
Rg4+ Kh6 42. b5 Ra1+ 43. Ke2 Ra2+ 44. Ke3 Ra3+ 45. Ke4 Rb3 46. Na7 Kh5 47. Rg3
Rb2 48. Kf5 Rf2+ 49. Ke6 Rb2 50. Rg7 e4 51. Kf5 Kh6 52. Rg3 Rxh2 53. Rb3 e3 54.
Rxe3 Rf2+ 55. Ke6 Kg5 56. Rb3 Re2+ 57. Kd7 Rd2+ 58. Kc8 Rd6 59. Nc6 f5 60. Kc7
Rf6 61. b6 Rf7+ 62. Kd8 h5 63. b7 h4 64. b8=Q Rf8+ 65. Ke7 Rxb8 66. Rxb8 Kf4 67.
Kf6 h3 68. Rh8 Kg3 69. Nd4 f4 70. Ne2+ Kf3 71. Nxf4 Kxf4 72. Rxh3 Ke4 73. Rh4+
Ke3 74. Kf5 Kd3 75. Ke5 Ke2 76. Kd4 Kf3 77. Kd3 Kg3 78. Ra4 Kf3 79. Kd2 Kf2 80.
Ra3 Kg1 81. Rf3 Kh1 82. Ke2 Kg2 83. Ra3 Kg1 84. Kf3 Kh1 85. Kg3 Kg1 86. Ra1
checkmate.

1-0, and almost a suicide-chess game by Fruit.  I am confident you, with or
without your engines, will find and agree with all or most of my moves.  Fruit
put no Laskeresque poison into the waters, ever.


Best,

A.S.



This page took 0.03 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.