The origins of Fruit 05/11/03 Gambit 006

Discussion of computer chess matches and engine tournaments.

Moderator: Ras

User avatar
Eelco de Groot
Posts: 4676
Joined: Sun Mar 12, 2006 2:40 am
Full name:   Eelco de Groot

The origins of Fruit 05/11/03 Gambit 006

Post by Eelco de Groot »

In games against Hiarcs 11.0 the first Gambit 005 version seemed not always to play very well so I started looking for some possible changes. In the following testgame I went looking for the point where actually the game was lost for Fruit from one of the basic Noomen positions, Noomen 05.


[Event "Test Fruit 05/11/03 Gambit 005"]
[Site "Groningen"]
[Date "2007.07.30"]
[Round "1.2"]
[White "Fruit 05/11/03"]
[Black "HIARCS 11 UCI"]
[Result "0-1"]

1. e4 c5 2. Nf3 Nc6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 d6 6. Bg5
e6 7. Qd2 a6 8. O-O-O h6 9. Be3 Be7 10. f4 O-O 11. Be2 Bd7
12. Kb1 b5 13. a3 Qb8 14. h4 Rc8 15. Nb3 b4 16. axb4 Nxb4
17. Bf3 e5 18. fxe5 dxe5 19. Qf2 Qc7 20. Qg3 Kh7 21. Rxd7
Nxd7 22. Bh5 Rf8 23. Nd4 Bf6 24. Nb3 a5 25. Nd4 Rab8
26. Bd1 Nxc2 27. Kxc2 Rfc8 28. Nde2 Qb7 29. Bc1 Qxe4+
30. Qd3 Qxd3+ 31. Kxd3 Nc5+ 32. Kc2 Nb3 33. Rf1 e4 34. Be3
a4 35. Rf5 Rc4 36. g4 Bxc3 37. Nxc3 a3 38. Kb1 Nd4 39. Bxd4
Rxb2+ 40. Kc1 Rxd4 41. Ra5 Rh2 42. Rxa3 h5 43. gxh5 Rc4
44. Bc2 f5 45. Kb2 Rd4 46. Kc1 Rdd2 47. Ra2 Kh6 48. Bb3
Rxa2 49. Bxa2 e3 50. Bb3 Rxh4 51. Bd1 f4 52. Nb5 Rxh5
53. Bxh5 Kxh5 54. Kd1 f3 55. Ke1 Kg4 0-1

[d]rq3rk1/3bbpp1/p1nppn1p/1p6/3NPP2/P1N1B3/1PPQB1PP/1K1R3R w - -
After 13... Qb8

I think in this position already the combination of 14. h4 and 15. Nb3 ?! are making the game very difficult for White, maybe it could have still easily saved the game once it knew it was in trouble, but for programs that is not so easy either, to change course and strategy in time when the opening or midgame plan goes bust. That makes these moves already critical. 14. h4 is not so bad at least it is an attacking move but 14. g4 , Hiarcs choice, would be better. But I have found no way to let Fruit 05/11/03 Gambit play this, whatever reasonable changes I made, mostly using -Ryan's I think?- pawn duo options. I could switch off the Pawn Duo settings but then the result was even worse. Main thing was trying to avoid the move Nb3 but I have not yet been succesful there, now Gambit 006 plays Nb3 in a slightly different line with h4 two moves earlier... But the situation in the above game improved a little if I carefully adjusted some parameters. The Pawn Duo settings were made a bit more aggressive over the whole range of file combinations a+b, b+c, c+d and d+e, then lowering the total weight in the midgame a bit. It is a very delicate balance showing the program has some real difficulty in this position, but eventually Gambit plays a combination of 14. h4 where Nb3 does not follow because Black immediately takes 14... Nxd4, then follows the choice 14. Rgh1 at 15 fifteen ply which I liked as being similar to 14. g4. It would never get the chance to play this 14. Rhg1 in a Blitz game but this was the only version so far that played this move. Unfortunately now the program plays 14. Rhe1 ?! at seventeen ply which is a rather senseless maneuver positioning the Rook behind all those Bishops, it recognizes this at the following ply and finally at eighteen ply Fruit Gambit 006 also plays 14. g4! Instead of 14. Rhe1 I wished that the version had opted for 14. Bf3 which is also okay and another version plays this, but this version had the downside of also playing Nb3 at 14 ply and never considering 14. Rhg1.

Move choices now in version Gambit 006:


00:00:00.0 1,21 1 85 g4
00:00:00.0 1,32 2 371 g4 d5
00:00:00.0 1,21 3 731 g4 d5 Nxc6 Bxc6
00:00:00.0 0,91 4 2045 g4 d5 exd5 exd5 Nxc6 Bxc6
00:00:00.0 0,25 5 7396 g4 Nxd4 Bxd4 e5 Bf2 Bxg4 Bxg4 Nxg4
00:00:00.0 0,60 5 11311 Nxc6 Bxc6 Bf3 Qb7 Qd3 a5
00:00:00.0 0,65 5 13909 h4 Nxd4 Qxd4 e5 Qb4 exf4 Bxf4
00:00:00.0 0,58 6 23266 h4 Nxd4 Qxd4 e5 Qd3 Rc8 g3
00:00:00.1 0,71 7 54421 h4 Nxd4 Qxd4 e5 Qd3 Rc8 f5 a5 Nxb5 Bxb5 Qxb5 Qxb5 Bxb5 Nxe4
00:00:00.2 0,63 8 101511 h4 Nxd4 Qxd4 e5 Qd3 Rc8 f5 Qb7 g4
00:00:00.4 0,63 9 247023 h4 Nxd4 Qxd4 e5 Qd3 Rc8 f5 Qb7 g4 a5
00:00:01.3 0,47 10 707604 h4 Rc8 g3 Nxd4 Qxd4 e5 Qd3 Be6 h5 b4 axb4 Qxb4
00:00:02.6 0,42 11 1499077 h4 Rc8 Nb3 e5 Qd3 Be6 f5 Bxb3 cxb3 b4 Nd5 Nxd5 Qxd5
00:00:05.3 0,27 12 3054104 h4 Rc8 Rhf1 Nxd4 Bxd4 e5 fxe5 dxe5 Be3 Be6 Nd5 Nxd5 exd5 Rd8 dxe6 Rxd2 exf7+ Kf8 Rxd2 Bxh4
00:00:13.1 0,17 13 7594286 h4 Rc8 Rhf1 Nxd4 Bxd4 Bc6 Bf3 a5 Bxf6 Bxf6 Qxd6 Qb7 e5 Bxh4 Bxc6 Rxc6
00:00:36.5 0,23 14 21730212 h4 Nxd4 Bxd4 e5 Be3 Qb7 fxe5 dxe5 Nd5 Bc6 Qd3 a5 Bf3 Nxd5 exd5 Bd7 Bh5
00:00:55.5 0,08 15 33409532 h4 Nxd4 Bxd4 e5 Be3 a5 fxe5 dxe5 Nd5 Nxd5 Qxd5 Rd8 Rhf1 Bf5 Qc6 Rc8 Qxb5 Bxe4
00:01:47.3 0,10 15 65216466 Rhg1 { :) }Nxd4 Qxd4 Bc6 Bf3 e5 fxe5 dxe5 Qb6 Qxb6 Bxb6 Rfc8 Rd3 Rab8 Ba7 Rb7 Be3 Rd7
00:02:35.4 0,08 16 94800069 Rhg1 Nxd4 Qxd4 Bc6 Bf3 e5 Qd3 b4 axb4 Qxb4 g3 Rab8 b3 a5 Ba7 Rb7 Bf2 Rfb8
00:04:53.6 0,01 17 181140544 Rhg1 Nxd4 Qxd4 Bc6 Bf3 e5 fxe5 dxe5 Qb6 Qc8 Nd5 Bxd5 exd5 e4 d6 Bd8 d7 Bxb6 dxc8Q Raxc8 Bxb6 exf3 gxf3
00:07:43.6 0,18 17 286013351 Rhe1 { :( ?!}Nxd4 Bxd4 e5 Be3 b4 axb4 Qxb4 Nd5 Nxd5 exd5 Qb7 fxe5 dxe5 Qd3 Rfc8 Rf1 Rab8 b3 f5 Qxa6
00:10:24.0 -0,16 18 384996107 Rhe1 Nxd4 Bxd4 e5 Be3 b4 axb4 Qxb4 Nd5 Nxd5 exd5 Qa4 fxe5 dxe5 d6 Bf6 Rf1 Rab8 b3 Rfc8 Rxf6 gxf6 Bxh6
00:14:48.0 0,02 18 546582770 Rhg1 Nxd4 Bxd4 e5 Be3 Qb7 fxe5 dxe5 Bxh6 gxh6 Qxh6 Nh7 Qh5 Bc5 Rgf1 f6 Bg4 Bxg4 Qxg4+ Qg7 Qe6+ Qf7 Qh3 Ng5
00:18:59.2 0,28 18 703697278 g4 Nxd4 Qxd4 e5 Qd3 Nxg4 Bxg4 Bxg4 Rdg1 Be6 fxe5 Kh8 Qd2 Rd8 Bxh6 g6 exd6 Qxd6 Qg2 Qe5 Rd1

As you can see, a real struggle and this was just one position! The final version still loses this Noomen 05 opening now by playing 12. h4, 13. a3 and 14. Nb3... :( Still the new parameters seem to create a bit more aggressive play against the first Hiarcs 11, this is a very solid version not using the Hypermodern settings. I also increased Fruit's time usage in the opening, it is a bit more similar to Hiarcs now and suits better the more aggressive set-ups I hope. Not using any Fischer timecontrols here Fruit Gambit still has to start playing very quickly in the later phases but the overall impression I had of time usage and Gambit 006 trying to set up some sort of attack in some games was favourable. But I only played a few games so far, I just wanted to tell something about how these settings came about.

These were the first games, as you can see in Noomen 05, Fruit 05/11/03 Gambit 006 is still going down aginst the strong Hiarcs when playing White but it won with the Black pieces! Both sides tried to race across with their pawns on either side of the board but Black was first using all its pawn duo knowledge to good effect this time :) :

[Event "Test Fruit 05/11/03 Gambit 006"]
[Site "Groningen"]
[Date "2007.07.30"]
[Round "1.1"]
[White "HIARCS 11.0 UCI"]
[Black "Fruit 05/11/03 Gambit 006"]
[Result "0-1"]

1. e4 c5 2. Nf3 Nc6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 d6 6. Bg5
e6 7. Qd2 a6 8. O-O-O h6 9. Be3 Be7 10. f4 O-O 11. Nxc6
bxc6 12. Bd3 d5 13. Bd4 c5 14. Bxf6 Bxf6 15. exd5 Rb8
16. d6 Qa5 17. a3 Bb7 18. Bc4 Rfd8 19. Bb3 Rxd6 20. Qxd6
Rd8 21. Qxd8+ Bxd8 22. Rhe1 Bxg2 23. f5 c4 24. Bxc4 Bg5+
25. Kb1 Qxf5 26. Bxa6 Bf6 27. Re2 Bc6 28. Rd3 g5

[d]6k1/5p2/B1b1pb1p/5qp1/8/P1NR4/1PP1R2P/1K6 w - -

Start of a pawn race that Fruit wins!

29. b4 g4 30. Kb2 h5 31. Kb3 h4 32. Bc4 Qf4 33. a4 g3 34. hxg3 hxg3
35. Bb5 Bb7 36. Re1 g2 37. Ne2 Qf2 38. Red1 Bf3 39. Rxf3
Qxf3+ 40. c3 e5 41. Kc2 e4 42. Rb1 Be5 43. Rd1 Kg7 44. a5
Qf1 45. Rd5 g1=Q 46. Nxg1 Qxg1 47. a6 Bf6 48. Rc5 e3 49. a7
Qa1 50. Rc7 Qa2+ 51. Kd3 Qd2+ 52. Ke4 e2 53. Bxe2 Qxe2+
54. Kd5 Bh4 55. a8=Q Qd3+ 0-1

[d]Q7/2R2pk1/8/3K4/1P5b/2Pq4/8/8 w - -


Second game was unfortunately lost in similar manner as in the analyzed testgame as described:

[Event "Test Fruit 05/11/03 Gambit 006"]
[Site "Groningen"]
[Date "2007.07.30"]
[Round "1.2"]
[White "Fruit 05/11/03 Gambit 006"]
[Black "HIARCS 11.0 UCI"]
[Result "0-1"]

1. e4 c5 2. Nf3 Nc6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 d6 6. Bg5
e6 7. Qd2 a6 8. O-O-O h6 9. Be3 Be7 10. f4 O-O 11. Be2 Bd7
12. h4 b5 13. a3 Qb8 14. Nb3 b4 15. axb4 Qxb4 16. Bf3 a5
17. e5 dxe5 18. fxe5 Nxe5 19. Bxa8 Rxa8 20. Nb1 Qg4 21. Qd4
Qh5 22. Bg5 Nc6 23. Qe3 Nd5 24. Rxd5 exd5 25. Bxe7 Re8
26. Nc3 Rxe7 27. Qc5 Re5 28. Nd4 Nxd4 29. Qxd4 f6 30. Qf2
Bf5 31. Rd1 Be4 32. g3 Qg4 33. Rf1 a4 34. Rd1 Rf5 35. Qe2
Qxe2 36. Nxe2 Rf3 37. Kd2 a3 38. bxa3 Rxa3 39. Rb1 h5
40. Rb7 Kh7 41. c3 d4 42. Re7 Ra2+ 43. Ke1 Ra1+ 44. Kd2
dxc3+ 45. Kxc3 Bf5 46. Nd4 Bg4 47. Kb2 Ra5 48. Rb7 Rc5
49. Kb3 Kh6 50. Kb4 Rc1 51. Ra7 g5 52. hxg5+ fxg5 53. Re7
h4 54. gxh4 gxh4 55. Ra7 h3 56. Ra2 Kg5 57. Nb5 Kf4 58. Nd6
Re1 59. Kc5 Kg3 60. Ra3+ Bf3 61. Ra2 Re5+ 62. Kb4 h2
63. Rxh2 Kxh2 64. Nc4 Rg5 65. Kc3 Kg3 66. Kd4 Kf4 67. Nd2
Rd5+ 68. Kc3 Be2 69. Nb3 Ke3 70. Kb4 Kd3 71. Nc5+ Kc2
72. Ne4 Rd4+ 0-1

[d]8/8/8/8/1K1rN3/8/2k1b3/8 w - -

In Noomen 06, the only other round played thus far, Fruit Gambit 006 managed to win another sharp game after drawing the first turn with Black:

[Event "Test Fruit 05/11/03 Gambit 006"]
[Site "Groningen"]
[Date "2007.07.30"]
[Round "1.1"]
[White "HIARCS 11.0 UCI"]
[Black "Fruit 05/11/03 Gambit 006"]
[Result "1/2-1/2"]

1. e4 c5 2. Nf3 e6 3. d4 cxd4 4. Nxd4 Nc6 5. Nc3 a6 6. Be2
Qc7 7. Qd3 Nf6 8. Bg5 Nxd4 9. Qxd4 Bc5 10. Qd3 b5 11. a3
Bd6 12. Bh4 Bb7 13. Bg3 Be5 14. O-O Bxg3 15. fxg3 O-O
16. Bf3 d5 17. exd5 Rfd8 18. Rf2 Nxd5 19. Nxd5 Bxd5 20. Be4
f5 21. Bxd5 Rxd5 22. Qe3 Re5 23. Qf3 Rd8 24. Raf1 g6
25. Qf4 Qc5 26. Kh1 Re4 27. Qg5 Rd7 28. Qc1 e5 29. c3 Qe3
30. Qc2 Kh8 31. Qb1 Rd8 32. Rc2 Qd3 33. Kg1 Kg7 34. Qc1 h6
35. Rcf2 a5 36. Rc2 Rd7 37. Rcf2 Rd5 38. Rc2 Rd8 39. Qb1
Qe3+ 40. Kh1 h5 41. Rcf2 Qd3 42. Qa1 Kh7 43. Qa2 Qd5
44. Qb1 Rd7 45. Kg1 Kg7 46. Qc1 Qd6 47. Kh1 Kh7 48. Qg5 Rd8
49. h3 Rc4 50. Rxf5 gxf5 51. Qxh5+ Kg7 52. Qg5+ Kh8
53. Qh5+ Kg8 54. Qg5+ Kh8 55. Qh5+ Kg7 56. Qg5+ Kh8
57. Qh5+ 1/2-1/2

[d] 3r3k/8/3q4/pp2pp1Q/2r5/P1P3PP/1P4P1/5R1K b - -

Draw by three times repetition


In a very long struggle, both sides down to fractions of seconds on the clock, Gambit 006 finally managed to bring Hiarcs to its knees with Rook, Knight and pawn in the return game which looked okay I think, considering it was only ten minutes for each side. Of course the quality of play in the endgame was probably not very high but it was an exciting Blitz struggle, both sides played more or less a tempo for the last sixty moves:

[Event "Test Fruit 05/11/03 Gambit 006"]
[Site "Groningen"]
[Date "2007.07.30"]
[Round "1.2"]
[White "Fruit 05/11/03 Gambit 006"]
[Black "HIARCS 11.0 UCI"]
[Result "1-0"]

1. e4 c5 2. Nf3 e6 3. d4 cxd4 4. Nxd4 Nc6 5. Nc3 a6 6. Be2
Qc7 7. Nxc6 dxc6 8. Qd3 Nf6 9. Bg5 Be7 10. O-O-O b5 11. Qg3
Qxg3 12. hxg3 e5 13. f4 h6 14. fxe5 Nd7 15. Bf4 Nc5 16. Bf3
O-O 17. Ne2 Ne6 18. c3 c5 19. Be3 Rd8 20. Nf4 Rxd1+
21. Rxd1 Bg5 22. Kc2 c4 23. Rd6 Be7 24. Rb6 Bg5 25. b3
cxb3+ 26. axb3 Bd8 27. Rd6 Bg5 28. Kb2 Kf8 29. Be2 Be7
30. Rd1 Rb8 31. Bh5 Bg5 32. Rd5 Kg8 33. Bf3 Ra8 34. Rd2 Nf8
35. Rd6 a5 36. Be2 Bd7 37. Bc5 a4 38. Rd5 Ne6 39. Be3 Be8
40. Bxb5 Bxb5 41. Rxb5 axb3 42. Rxb3 Ra5 43. Rb8+ Kh7
44. Nxe6 Bxe3 45. Nf8+ Kg8 46. e6 fxe6 47. Nxe6+ Kf7
48. Nd8+ Kg6 49. Nc6 Rg5 50. e5 Kf5 51. Rf8+ Ke6 52. Rf3
Bc5 53. Rd3 Kf5 54. Kc2 Rg6 55. Rf3+ Ke6 56. c4 Kd7 57. Na5
Bb6 58. Nb3 Ke6 59. Kd3 Rg4 60. c5 Bc7 61. Nd4+ Kd5 62. Nc2
Kxe5 63. Rf8 Rxg3+ 64. Ne3 Ke6 65. Re8+ Kd7 66. Re4 Rg5
67. Kc4 h5 68. Nd5 Bd8 69. Re2 h4 70. Rd2 Ba5 71. Rf2 Kc6
72. Ne7+ Kd7 73. Nf5 Rg4+ 74. Kb5 Bc7 75. Re2 Bd8 76. Rd2+
Kc7 77. Ne3 Rg3 78. Nd5+ Kc8 79. c6 Rg5 80. Kc4 Rh5 81. Ra2
Kb8 82. Rf2 Rh8 83. Rf7 g5 84. Kb5 Re8 85. Rg7 Re5 86. Kc5
Re8 87. Nb6 Be7+ 88. Kb5 g4 89. Rxg4 Bd8 90. Rd4 Bc7
91. Rxh4 Bd8 92. Rh7 Re5+ 93. Ka6 Re7 94. Rh8 Kc7 95. Nd5+
Kxc6 96. Rxd8 Re2 97. Nb4+ Kc5 98. Nd3+ Kc6 99. Rc8+ Kd7
100. Rg8 Ke7 101. g4 Rd2 102. Rg7+ Kf6 103. Rd7 Rg2
104. Rd4 Kg5 105. Ne5 Rg3 106. Nf7+ Kg6 107. Rf4 Re3
108. Rf5 Rb3 109. Ne5+ Kh7 110. g5 Rb8 111. Rf7+ Kg8
112. Re7 Rb3 113. g6 Rg3 114. Kb5 Kf8 115. Rf7+ Kg8
116. Rf5 Rb3+ 117. Kc5 Ra3 118. Rf7 Rg3 119. Rb7 Rg1
120. Kd4 Rf1 121. Rc7 Rf2 122. Ke4 Rg2 123. Kf5 Rf2+ 1-0


[d]6k1/2R5/6P1/4NK2/8/8/5r2/8 w - -
Won! :!:

Eelco