A complete 2000 lines of code engine

[MikeB]

Being a little tired of finding bugs in Weini, I started a week-end project last sunday. The idea was the following : "starting from a blank page, how far can I go in 2 days now I know a little more on chess programming".

An expected side effect was to get a simple engine to experiment with and maybe find some Weini's bug in it.

It is a mailbox negamax (pvs) engine, with classic pruning trick, a double bucket TT and only PST eval.

After 2 days the engine was working in xboard, validated in perft, finds mate in 8 from a collection I often use and was able to always win against me at fast TC (but loses against fairy-max).

Since then I corrected some stuff but the engine is still quite bad, but working

Here's the source code : https://github.com/tryingsomestuff/Mini ... r/minic.cc

Some of you probably can easily spot very bad things that makes it so weak.

Maybe, as a 2000 lines c++ code, it may also become an entry point for beginners. I can open the repository for write if some of you want to contribute.

Very nice- here's a macOS version based on the latest source
https://www.dropbox.com/s/envzgbn1uo2be ... S.zip?dl=1

under xboard - edit the xboard config file to something like this:
"/Users/xxxx/cluster.mfb/minic -xboard" -fd "/Users/xxxx/cluster.mfb"

seems to dominate fairy max now ...

[Event "Mac Pro x5690 3.46 Ghz 18 CPU"] [Site "Mac-Pro.local"] [Date "2018.10.30"] [Round "-"] [White "Fairy-Max 4.8O"] [Black "Minic 0.9"] [Result "0-1"] [TimeControl "40/60"] [Annotator "1. +0.15 1... -0.18"] 1. d4 {+0.15/8 +56} Nf6 {-0.18/12 +58} 2. f4 {+0.14/7 +55} e6 {+0.39/11 +57} 3. g3 {+0.10/7 +54} c5 {+0.59/11 +56} 4. c3 {+0.01/8 +52} d5 {+0.42/12 +54} 5. Nf3 {+0.01/8 +51} Bd6 {+0.37/12 +53} 6. Bh3 {-0.05/8 +46} Nc6 {+0.56/11 +52} 7. dxc5 {+0.14/7 +46} Bxc5 {+0.70/13 +50} 8. b4 {-0.08/8 +44} Bb6 {+0.65/12 +49} 9. Nbd2 {-0.19/8 +41} a6 {+0.75/11 +48} 10. Bb2 {-0.21/8 +40} O-O {+0.75/11 +47} 11. Qc2 {+0.04/7 +37} Be3 {+0.67/11 +45} 12. Nf1 {+0.19/7 +36} Bb6 {+0.76/10 +44} 13. O-O-O {+0.13/8 +35} Ne4 {+1.19/11 +43} 14. e3 {-0.08/7 +34} e5 {+0.85/10 +41} 15. Bxc8 {+0.04/7 +33} Rxc8 {+1.00/12 +40} 16. a3 {+0.13/7 +32} exf4 {+1.35/11 +39} 17. gxf4 {+0.09/8 +31} Ne7 {+1.17/11 +38} 18. Rd3 {-0.01/7 +30} Nf5 {+1.29/10 +36} 19. a4 {+0.06/7 +29} a5 {+1.65/11 +35} 20. Ba3 {-0.09/7 +27} Re8 {+1.59/11 +34} 21. Qb2 {-0.19/7 +26} Qf6 {+2.98/11 +32} 22. Ne5 {-0.46/7 +24} Nxe3 {+3.37/11 +31} 23. Nxe3 {-1.46/8 +23} Bxe3+ {+5.58/11 +30} 24. Kb1 {-3.59/7 +22} Bxf4 {+5.30/11 +29} 25. Nf3 {-3.42/8 +20} Nxc3+ {+5.80/14 +27} 26. Ka1 {-2.98/7 +18} Re2 {+6.20/13 +26} 27. Nd2 {-4.57/7 +17} Be5 {+7.36/12 +25} 28. Rc1 {-4.91/8 +16} Rc6 {+7.74/14 +23} 29. Qb3 {-4.87/7 +15} Ne4+ {+7.74/12 +22} 30. Ka2 {-5.08/8 +14} Rxd2+ {+6.13/13 +21} 31. Rxd2 {-5.12/9 +12} Nxd2 {+5.95/16 +20} 32. Rxc6 {-5.06/9 +12} Qxc6 {+7.95/17 +18} 33. b5 {-5.51/12 +11} Qg6 {+14.85/18 +17} 34. Qd1 {-10.44/11 +9} Qd3 {+16.67/17 +16} 35. Bb2 {-11.76/13 +9} Bxb2 {+17.60/19 +14} 36. Kxb2 {-13.76/12 +7} Nc4+ {+17.97/19 +13} 37. Kc1 {-13.75/11 +6} Qa3+ {+18.00/18 +12} 38. Kc2 {-13.74/10 +5} Qxa4+ {+18.00/16 +11} 39. Kc1 {-13.85/10 +5} Qa1+ {+18.22/16 +9} 40. Kc2 {-14.54/10 +64} Ne3+ {+18.42/15 +68} 41. Kd3 {-18.28/11 +63} Qxd1+ {+21.94/13 +67} 42. Kxe3 {-19.00/11 +61} a4 {+23.04/13 +65} 43. Kf4 {-19.78/11 +59} a3 {+23.39/14 +64} 44. Ke5 {-19.79/10 +59} a2 {+23.35/13 +63} 45. Kd6 {-20.02/10 +57} a1=Q {+24.49/12 +62} 46. Kc7 {-20.61/9 +56} Qa5+ {+24.63/10 +60} 47. Kxb7 {-79.97/16 +55} Qd4 {+89.95/25 +59} 48. Kc6 {-79.98/28 +55} Qd8 {+89.97/64 +59} 49. Kb7 {-79.99/28 +55} Q4b6# {+89.99/64 +59} {Black mates} 0-1

[xr_a_y]

Yes the 0.9 version includes a very small opening book that may help and singular extension for tt move.

[xr_a_y]

Minic 0.12 is released, partialy using HQ bitboard implementation, mainly taken from "Dumb/Amoeba" (thanks to Richard Delorme). The speed up is already incredible and the strength follows ...

For this release I am cheating a little as sloc=2127. I'll clean up the code soon.

Code: Select all

Rank Name                          Elo     +/-   Games   Score   Draws
   1 minic0.12                     256      73      67   81.3%   34.3%
   2 minic0.9                      -86      63      66   37.9%   45.5%
   3 fairymax                     -142      71      67   30.6%   34.3%

I am also going to finish using bitboard in move generation soon and then backport that to Weini.
Maybe i'll add some eval features in Minic next week.

[xr_a_y]

I am also going to finish using bitboard in move generation soon and then backport that to Weini.

Or maybe the other way around ...

[xr_a_y]

So 1on1 Minic0.14 (still PST only) is now +100 versus Weini0.0.24 .

[Ronald]

Nice!

[xr_a_y]

Still PST only (on a single thread here) but in the meantime Minic goes SMP.

[xr_a_y]

Two first Minic wins against Fruit2.1, even one with black pieces

The rest of the 10 short TC games gives 7 wins for Fruit, 2 for Minic and 1 draw ... glad with that ...

[Event "Computer Chess Game"] [Site "DESKTOP-17O7D54"] [Date "2018.12.19"] [Round "1"] [White "Minic dev"] [Black "Fruit 2.1"] [Result "1-0"] [TimeControl "40/20"] [Annotator "1. +0.13 1... -0.09"] 1. e4 {+0.13/12} d5 {-0.09/10 0.5} 2. exd5 {+0.31/14 0.5} Nf6 {-0.10/10 0.7} 3. d4 {+0.31/13 0.5} Nxd5 {-0.16/9 0.6} 4. Nf3 {+0.35/12 0.5} e6 {-0.36/9 0.6} 5. Be2 {+0.37/12 0.5} Bb4+ {-0.13/9 0.7} 6. c3 {+0.46/12 0.5} Bd6 {-0.19/10 0.5} 7. c4 {+0.23/12 0.5} Nf4 {-0.15/9 0.4} 8. O-O {+0.49/13 0.5} O-O {-0.05/10 0.6} 9. Nc3 {+0.49/13 0.5} Nc6 {-0.05/9 0.3} 10. c5 {+0.49/12 0.5} Nxe2+ {-0.16/8 0.1} 11. Qxe2 {+0.45/13 0.5} Be7 {-0.23/9 0.3} 12. Rd1 {+0.42/12 0.5} Nb4 {-0.20/9 0.6} 13. Bf4 {+0.38/12 0.5} b6 {-0.23/9 0.5} 14. Ne5 {+0.39/12 0.5} Bb7 {-0.03/8 0.3} 15. Rac1 {+0.57/11 0.5} Bh4 {-0.18/8 0.5} 16. Qg4 {+0.74/10 0.5} Kh8 {-0.22/8 0.6} 17. a3 {+0.88/11 0.5} Nd5 {-0.25/8 0.4} 18. c6 {+1.13/13 0.5} f5 {-0.67/9 0.5} 19. Qf3 {+1.48/14 0.5} Ba6 {-0.89/10 0.5} 20. Nxd5 {+2.06/14 0.5} exd5 {-1.13/10 0.5} 21. Nd7 {+1.98/13 0.5} Re8 {-0.91/10 0.5} 22. Be5 {+2.00/13 0.5} Qg5 {-0.94/9 0.3} 23. Re1 {+1.96/12 0.5} Qd2 {-0.96/8 0.6} 24. g3 {+2.46/12 0.5} Bg5 {-1.37/9 0.7} 25. Qxf5 {+2.45/14 0.5} Bd3 {-1.61/9 0.4} 26. Qf7 {+2.96/13 0.5} Rg8 {-1.25/8 0.3} 27. Qxd5 {+2.84/12 0.5} Qxb2 {-1.67/9 0.9} 28. f4 {+3.06/14 0.5} Bd8 {-1.81/9 0.4} 29. Qf3 {+2.89/13 0.5} Bb5 {-1.42/8 0.5} 30. f5 {+3.20/12 0.5} Bg5 {-1.46/8 0.4} 31. Rb1 {+4.56/13 0.5} Qd2 {-3.17/9 0.6} 32. f6 {+4.37/12 0.5} Bc4 {-3.53/8 0.6} 33. fxg7+ {+5.25/13 0.5} Rxg7 {-2.82/4} 34. Rbd1 {+4.38/13 0.5} Qa2 {-4.07/8 0.3} 35. Bxg7+ {+5.26/14 0.5} Kxg7 {-4.50/9 0.5} 36. Qf5 {+4.75/14 0.5} h6 {-3.68/9 0.5} 37. Qe5+ {+4.48/13 0.5} Kg8 {-4.90/10 0.9} 38. Nf6+ {+4.56/14 0.5} Bxf6 {-4.43/9 0.4} 39. Qxf6 {+4.82/14 0.5} Rf8 {-4.43/9 0.3} 40. Qg6+ {+4.87/13 0.5} Kh8 {-4.50/4} 41. Qxh6+ {+4.84/13 0.5} Kg8 {-4.50/4} 42. Qg6+ {+5.09/13 0.5} Kh8 {-4.50/4} 43. Qh5+ {+4.75/13 0.5} Kg8 {-4.64/10 0.7} 44. Re7 {+5.25/14 0.5} Qf2+ {-4.87/9 0.8} 45. Kh1 {+5.18/13 0.5} Qf3+ {-4.79/10 0.9} 46. Qxf3 {+5.78/14 0.5} Rxf3 {-4.89/10 0.2} 47. Rxc7 {+6.10/14 0.5} Bd5 {-5.19/11 0.5} 48. Kg1 {+6.03/14 0.5} Rc3 {-5.97/12 2.4} 49. Re1 {+5.64/14 0.5} Kf8 {-5.79/11 0.3} 50. Rf1+ {+5.62/13 0.5} Kg8 {-5.59/11 0.5} 51. Rf6 {+6.01/14 0.5} Rc1+ {-7.14/11 0.9} 52. Kf2 {+7.23/11 0.5} Rc2+ {-7.14/11 0.4} 53. Ke1 {+6.93/12 0.5} Bc4 {-7.62/10 0.5} 54. Rg6+ {+8.46/13 0.5} Kf8 {-7.62/11 0.5} 55. Rxa7 {+8.50/13 0.5} Re2+ {-7.75/10 0.3} 56. Kd1 {+9.13/13 0.5} Re8 {-7.75/10 0.4} 57. c7 {+9.42/13 0.5} Rc8 {-7.82/12 0.4} 58. Rxb6 {+8.94/13 0.5} Ke7 {-8.38/13 0.4} 59. Kd2 {+9.13/12 0.5} Kd7 {-7.82/11 0.3} 60. Rg6 {+9.09/12 0.5} Ke7 {-7.85/11 0.5} 61. h4 {+9.83/13 0.5} Bf7 {-8.52/11 0.6} 62. Rh6 {+9.86/12 0.5} Bg8 {-8.88/12 0.6} 63. Rc6 {+10.02/12 0.5} Bd5 {-8.46/12 0.4} 64. Rc5 {+10.35/12 0.5} Ba2 {-8.77/11 0.5} 65. h5 {+11.03/12 0.5} Kd7 {-9.12/12 0.5} 66. a4 {+11.06/13 0.5} Bb1 {-9.49/11 0.5} 67. h6 {+11.53/13 0.5} Be4 {-9.51/11 0.5} 68. Ke3 {+11.80/13 0.5} Bh7 {-10.22/11 0.5} 69. a5 {+11.78/13 0.5} Kd6 {-11.69/12 0.7} 70. a6 {+12.50/13 0.5} Re8+ {-15.71/14 2.2} 71. Kf4 {+12.81/13 0.5} Rf8+ {-11.36/11 0.2} 72. Kg5 {+14.14/14 0.5} Rg8+ {-16.95/12 1.1} 73. Kf6 {+14.26/14 0.5} Rc8 {-16.95/12 0.1} 74. Kg7 {+15.18/14 0.5} Be4 {-17.14/11 0.2} 75. h7 {+16.21/14 0.5} Bxh7 {-17.38/11 0.1} 76. Kxh7 {+16.57/13 0.5} Ke7 {-17.54/12 0.1} 77. Rb7 {+17.37/13 0.5} Kf7 {-23.91/9 0.3} 78. Rb8 {+22.47/12 0.5} Rxc7 {-99.88/8 0.1} 79. Rxc7+ {+22.47/10 0.5} Ke6 {-99.90/9 0.1} 80. a7 {+22.23/9 0.5} Kd6 {-99.92/33 0.1} 81. Rc5 {+99.92/10 0.5} Ke7 {-99.94/43 0.5} 82. a8=Q {+99.94/44 0.5} Kd6 {-99.96/42 0.5} 83. Re8 {+99.96/58 0.1} Kd7 {-99.98/4} 84. Qd8# {+99.98/58} {Xboard adjudication: Checkmate} 1-0

[Event "Computer Chess Game"] [Site "DESKTOP-17O7D54"] [Date "2018.12.19"] [Round "10"] [White "Fruit 2.1"] [Black "Minic dev"] [Result "0-1"] [TimeControl "40/20"] [Annotator "1. +0.23 1... -0.15"] 1. Nf3 {+0.23/11} Nf6 {-0.15/12 0.5} 2. Nc3 {+0.11/11 0.7} d5 {-0.13/12 0.5} 3. d4 {+0.12/10 0.4} e6 {-0.09/12 0.5} 4. Bg5 {+0.12/9 0.6} Be7 {-0.14/13 0.5} 5. e3 {+0.37/10 0.5} h6 {+0.04/12 0.5} 6. Bf4 {+0.30/9 0.3} O-O {-0.07/12 0.5} 7. Ne5 {+0.27/9 0.5} c5 {+0.10/13 0.5} 8. Bb5 {+0.12/8 0.4} c4 {+1.49/12 0.5} 9. O-O {-0.95/8 0.5} a6 {+1.52/14 0.5} 10. Bxc4 {-1.10/9 0.4} dxc4 {+1.81/14 0.5} 11. Qe2 {-1.21/9 1.1} b5 {+2.39/13 0.5} 12. a4 {-1.54/9 1.1} b4 {+2.31/13 0.5} 13. Nb1 {-1.44/9 0.5} Bb7 {+2.29/14 0.5} 14. Nxc4 {-1.50/9 0.4} Qd5 {+2.28/13 0.5} 15. f3 {-1.34/9 0.4} Nbd7 {+2.11/12 0.5} 16. e4 {-1.29/10 0.5} Qc6 {+1.72/13 0.5} 17. Na5 {-1.14/10 0.4} Qb6 {+1.56/14 0.5} 18. Nxb7 {-1.29/10 0.4} Qxb7 {+1.44/12 0.5} 19. Nd2 {-1.29/10 0.5} Qc6 {+1.80/11 0.5} 20. Rfc1 {-1.22/9 0.4} Rae8 {+1.64/11 0.5} 21. c3 {-0.92/9 0.3} bxc3 {+1.46/12 0.5} 22. Rxc3 {-0.75/9 0.4} Qa8 {+1.54/12 0.5} 23. Rac1 {-0.75/9 0.5} a5 {+1.38/11 0.5} 24. Rc7 {-0.65/9 0.4} Rd8 {+1.24/12 0.5} 25. Qe3 {-0.72/8 0.6} Nh5 {+2.54/13 0.5} 26. Bxh6 {-1.26/8 0.4} gxh6 {+2.83/14 0.5} 27. Qxh6 {-1.50/8 0.4} Ng7 {+2.99/13 0.5} 28. Nb3 {-1.51/8 0.3} Bd6 {+2.98/13 0.5} 29. R7c3 {-1.76/9 0.4} Nb6 {+2.88/12 0.5} 30. Ra1 {-1.78/9 0.4} Bb4 {+2.85/13 0.5} 31. Rc2 {-1.96/9 0.4} Rc8 {+2.88/13 0.5} 32. Qc1 {-2.50/9 0.9} Rxc2 {+3.24/14 0.5} 33. Qxc2 {-2.67/9 0.3} Qc8 {+2.96/13 0.5} 34. Qd3 {-2.37/10 0.4} Qc4 {+3.22/13 0.5} 35. Qd1 {-2.67/10 0.3} Qc6 {+3.26/13 0.5} 36. Qe2 {-2.78/10 0.7} Nxa4 {+3.34/12 0.5} 37. Rc1 {-2.84/9 0.3} Qd7 {+3.47/12 0.5} 38. g4 {-2.95/9 0.4} Rb8 {+3.93/11 0.5} 39. Qc2 {-3.28/8 0.2} Bd6 {+3.89/12 0.5} 40. Ra1 {-3.17/8 0.2} Rc8 {+3.74/12 0.5} 41. Qe2 {-3.44/10 0.6} Qb7 {+4.03/12 0.5} 42. Nxa5 {-3.70/10 0.7} Qb4 {+3.76/13 0.5} 43. Kf1 {-3.82/10 0.6} Qxa5 {+4.30/14 0.5} 44. b3 {-4.24/10 0.5} Qc3 {+4.50/15 0.5} 45. Rxa4 {-5.34/10 0.4} Qxb3 {+4.80/15 0.5} 46. Ra2 {-4.76/10 0.4} Bxh2 {+4.69/14 0.5} 47. Rb2 {-4.63/10 0.5} Qc3 {+4.89/14 0.5} 48. Rd2 {-4.94/10 0.4} Bf4 {+4.89/12 0.5} 49. Rd3 {-5.72/11 0.4} Qa5 {+4.92/13 0.5} 50. Rd1 {-5.35/10 0.4} Rd8 {+4.95/12 0.5} 51. Kg2 {-4.80/10 0.5} Qg5 {+4.85/12 0.5} 52. Kf1 {-4.70/10 0.5} Ne8 {+5.19/13 0.5} 53. Qb2 {-4.91/10 0.5} Qh4 {+5.19/13 0.5} 54. Qf2 {-5.03/10 0.4} Qh3+ {+5.31/12 0.5} 55. Qg2 {-5.35/11 0.5} Qxg2+ {+5.22/14 0.5} 56. Kxg2 {-5.37/12 0.2} Nc7 {+5.33/14 0.5} 57. Rd3 {-5.46/13 0.5} Be5 {+5.39/14 0.5} 58. g5 {-5.57/11 0.3} Bf4 {+5.68/14 0.5} 59. Rb3 {-5.72/12 0.5} Rxd4 {+6.14/13 0.5} 60. Rb7 {-6.20/12 0.5} Rc4 {+6.21/14 0.5} 61. Kh3 {-6.47/11 0.4} Bxg5 {+6.21/13 0.5} 62. Kg4 {-6.58/11 0.4} Be3 {+6.32/13 0.5} 63. Rb3 {-6.71/11 0.5} Nd5 {+6.43/13 0.5} 64. Kh3 {-6.83/11 0.5} Nc3 {+6.40/12 0.5} 65. Rb7 {-7.02/12 0.6} Ne2 {+6.46/11 0.5} 66. Rb3 {-6.91/10 0.4} Bf4 {+6.54/12 0.5} 67. Rb2 {-7.18/11 0.6} Nd4 {+6.61/13 0.5} 68. Rg2+ {-7.26/11 0.6} Kf8 {+7.34/12 0.5} 69. Rf2 {-7.97/12 0.6} Rc1 {+7.51/13 0.5} 70. Ra2 {-7.98/12 0.6} Rg1 {+8.41/13 0.5} 71. Rg2 {-8.29/12 0.4} Rh1+ {+9.59/13 0.5} 72. Kg4 {-10.91/12 0.4} e5 {+9.67/13 0.5} 73. Rg3 {-10.91/11 0.6} Bxg3 {+9.90/13 0.5} 74. Kxg3 {-11.17/10 0.6} Rf1 {+10.23/14 0.5} 75. Kg2 {-19.64/12 1.8} Rxf3 {+10.31/14 0.5} 76. Kh2 {-99.86/12 0.5} Rf4 {+10.65/13 0.5} 77. Kg1 {-11.84/9 0.2} Rxe4 {+10.98/12 0.5} 78. Kf2 {-19.56/11 0.6} Re2+ {+11.82/12 0.5} 79. Kg3 {-99.88/11 0.3} Kg7 {+12.18/13 0.5} 80. Kg4 {-99.90/21 0.2} f5+ {+18.81/13 0.5} 81. Kg3 {-99.90/37 0.4} f4+ {+19.75/11 0.5} 82. Kh3 {-99.92/45 0.5} f3 {+19.04/11 0.5} 83. Kg3 {-99.90/31 0.4} Rg2+ {+99.90/9 0.5} 84. Kh3 {-99.92/45 0.5} Kg6 {+99.92/12 0.5} 85. Kh4 {-99.94/4} f2 {+99.94/58 0.2} 86. Kh3 {-99.96/4} f1=Q {+99.96/58} 87. Kh4 {-99.98/4} Qh1# {+99.98/58} {Xboard adjudication: Checkmate} 0-1

[CMCanavessi]

Amazing progress, and still PST only?

[xr_a_y]

Amazing progress, and still PST only?

There is a little passer evaluation term (but only a passer bonus depending on rank (nothing else specific to pawn yet, as protected, unstoppable, double, isolated, ...)