Progress on Blunder

[Guenther]

Saved the PGN as http://rebel13.nl/b/blunder.pgn as I need the PC for another match.

I fear only a debug/log file will help here (I always saved debugs when using cutechess-cli).
The first game with no result which probably crashed Blunder for all other games was a game in which the other side
had its turn, which was Monolith?

But... we are already in a drawn 5men endgame R+N vs. R and around half of the moves for a 50 moves draw were already done.

BTW what concurrency is used here?

[pgn][Event "?"]
[Site "?"]
[Date "2021.12.11"]
[Round "142"]
[White "Monolith_0.3"]
[Black "Blunder_7.3.0"]
[Result "*"]
[ECO "D31"]
[GameDuration "00:13:38"]
[GameEndTime "2021-12-11T13:16:59.278 West-Europa (standaardtijd)"]
[GameStartTime "2021-12-11T13:03:21.207 West-Europa (standaardtijd)"]
[Opening "QGD"]
[PlyCount "276"]
[Termination "unterminated"]
[TimeControl "40/120"]
[Variation "Semi-Slav, Marshall gambit"]

1. d4 {book} d5 {book} 2. c4 {book} e6 {book} 3. Nc3 {book} c6 {book}
4. e4 {book} dxe4 {book} 5. Nxe4 {book} Bb4+ {book} 6. Bd2 {book} Qxd4 {book}
7. Bxb4 {book} Qxe4+ {book} 8. Be2 {book} Na6 {book} 9. Bd6 {book} e5 {book}
10. Nf3 {book} Bg4 {book} 11. O-O {book} O-O-O {book} 12. Bd3 {+0.83/12 2.9s}
Qf4 {-0.59/18 4.1s} 13. Bxe5 {+0.66/11 4.0s} Qxe5 {-0.61/18 4.1s}
14. Nxe5 {+0.86/13 4.0s} Bxd1 {-0.62/17 4.1s} 15. Bf5+ {+0.66/12 4.0s}
Kc7 {-0.27/16 4.1s} 16. Raxd1 {+0.61/12 2.5s} Nh6 {-0.22/18 4.1s}
17. Bc2 {+0.46/12 2.7s} Rxd1 {-0.25/17 4.1s} 18. Rxd1 {+0.53/13 2.7s}
Re8 {-0.27/17 4.1s} 19. Nf3 {+0.35/11 4.2s} Ng4 {-0.08/16 4.1s}
20. h3 {+0.33/10 4.2s} Nf6 {-0.14/17 4.1s} 21. g4 {+0.23/10 2.6s}
h6 {-0.09/17 4.1s} 22. Kg2 {+0.31/11 2.5s} Nb4 {0.00/18 4.1s}
23. Bb1 {+0.24/12 4.4s} c5 {-0.06/20 4.1s} 24. a3 {+0.16/12 2.7s}
Nc6 {-0.06/17 4.1s} 25. Bc2 {+0.17/11 3.3s} a5 {+0.03/16 4.1s}
26. Rd2 {+0.13/11 4.6s} Nd7 {0.00/18 4.1s} 27. Rd5 {+0.14/12 3.9s}
Re2 {+0.17/20 4.1s} 28. Bf5 {+0.14/13 2.7s} Nf6 {+0.11/20 4.1s}
29. Rxc5 {+0.20/13 3.5s} Rxb2 {+0.38/18 4.1s} 30. Rb5 {+0.21/13 3.4s}
Ra2 {+0.31/19 4.1s} 31. Rb3 {+0.20/14 5.2s} a4 {+0.37/19 4.1s}
32. Re3 {+0.22/13 4.2s} g6 {+0.40/16 4.1s} 33. Bb1 {+0.15/14 3.0s}
Rb2 {+0.40/20 4.1s} 34. Bd3 {+0.03/13 3.8s} Kd6 {+0.49/16 4.1s}
35. Kg3 {0.00/12 6.2s} Kc5 {+0.42/16 4.1s} 36. Be2 {+0.05/13 4.9s}
Nd7 {+0.41/15 4.1s} 37. h4 {-0.31/12 3.8s} Ra2 {+0.66/19 4.1s}
38. Kg2 {-0.34/13 5.7s} Nf6 {+0.54/19 4.1s} 39. Ne1 {-0.30/12 8.7s}
Nd4 {+0.69/16 4.1s} 40. Nd3+ {-0.76/13 4.8s} Kb6 {+0.54/15 3.6s}
41. Bd1 {-0.69/12 1.7s} Rd2 {+1.25/16 3.0s} 42. Bxa4 {-0.70/11 3.0s}
Ka5 {+0.99/17 3.0s} 43. g5 {-0.65/12 3.0s} hxg5 {+1.04/19 3.0s}
44. hxg5 {-0.73/13 2.6s} Ng4 {+1.13/20 3.0s} 45. Rg3 {-1.00/14 2.4s}
Nxf2 {+0.87/21 3.0s} 46. Nxf2 {-0.95/14 1.8s} Kxa4 {+1.03/21 3.0s}
47. Kf1 {-0.86/13 2.4s} Ne6 {+1.01/21 3.0s} 48. Ke1 {-0.97/12 3.1s}
Ra2 {+1.19/17 3.0s} 49. Nd3 {-1.13/12 3.1s} Rxa3 {+1.08/16 3.0s}
50. Kd2 {-1.10/13 3.1s} Ra1 {+1.13/17 3.0s} 51. Rg4 {-1.12/11 3.1s}
Rf1 {+1.52/19 3.0s} 52. Rh4 {-1.44/11 2.5s} Rf5 {+1.70/15 3.0s}
53. Rh1 {-1.52/12 2.1s} Kb3 {+1.77/19 3.0s} 54. Rc1 {-1.56/12 2.0s}
Nxg5 {+1.61/16 3.0s} 55. Rc3+ {-1.61/13 2.1s} Ka4 {+1.63/17 3.0s}
56. Rc1 {-1.80/12 3.3s} Ka5 {+1.70/19 3.0s} 57. Rb1 {-1.78/12 3.3s}
b5 {+1.70/17 3.0s} 58. Ra1+ {-1.81/12 3.3s} Kb6 {+1.55/16 3.0s}
59. c5+ {-1.85/12 2.1s} Kb7 {+1.78/17 3.0s} 60. Rc1 {-2.03/12 1.7s}
Ne6 {+1.78/17 3.0s} 61. c6+ {-1.94/13 2.2s} Kc7 {+1.78/15 3.0s}
62. Nb4 {-2.00/12 2.0s} Kd6 {+1.73/16 3.0s} 63. Ke3 {-1.85/11 2.0s}
g5 {+1.93/17 3.0s} 64. Na6 {-2.41/12 3.7s} Nc7 {+2.08/17 3.0s}
65. Nb4 {-2.53/13 3.1s} g4 {+2.08/16 3.0s} 66. Rg1 {-2.75/10 3.8s}
Re5+ {+2.15/17 3.0s} 67. Kf2 {-3.01/14 2.9s} Re4 {+2.50/18 3.0s}
68. Rd1+ {-2.89/14 3.9s} Kc5 {+2.42/17 3.0s} 69. Rd7 {-3.03/14 2.4s}
Kb6 {+2.71/18 3.0s} 70. Nd3 {-3.16/14 3.5s} Kxc6 {+2.50/17 3.0s}
71. Rxf7 {-3.30/14 3.0s} b4 {+2.58/18 3.0s} 72. Rf6+ {-2.97/13 2.4s}
Kb5 {+2.64/17 3.0s} 73. Rf5+ {-3.33/13 4.0s} Kc4 {+2.55/15 3.0s}
74. Nb2+ {-3.44/12 4.6s} Kc3 {+2.62/16 3.0s} 75. Na4+ {-3.46/14 4.4s}
Kd4 {+2.65/17 3.0s} 76. Rf8 {-3.31/13 3.6s} Ne6 {+2.78/17 3.0s}
77. Rb8 {-3.27/13 5.1s} Ke5 {+2.75/17 3.0s} 78. Nb2 {-3.43/13 4.1s}
Nc5 {+2.58/16 3.0s} 79. Kg3 {-3.32/13 3.7s} b3 {+2.58/14 3.0s}
80. Nd1 {-3.25/13 4.4s} Kf5 {+2.58/16 2.5s} 81. Rf8+ {-3.05/12 2.9s}
Ke6 {+2.59/17 3.0s} 82. Rb8 {-3.24/12 1.6s} Rd4 {+2.59/16 3.0s}
83. Rb6+ {-3.41/12 3.0s} Kd7 {+2.87/17 3.0s} 84. Nb2 {-3.53/12 1.5s}
Re4 {+3.05/17 3.0s} 85. Kh4 {-3.47/13 2.6s} Kc7 {+2.87/17 3.0s}
86. Rb5 {-3.38/14 1.8s} Kc6 {+2.81/16 3.0s} 87. Rb8 {-3.23/14 3.1s}
Kd6 {+2.81/14 3.0s} 88. Rb6+ {-3.44/13 2.0s} Kd5 {+2.78/15 3.0s}
89. Rb8 {-3.47/11 3.1s} Kd4 {+2.60/17 3.0s} 90. Nd1 {-3.11/13 2.7s}
Kc4 {+2.84/19 3.0s} 91. Rd8 {-3.19/12 1.8s} Re2 {+2.98/20 3.0s}
92. Kxg4 {-3.54/12 3.2s} b2 {+3.03/20 3.0s} 93. Rb8 {-3.52/13 2.1s}
Nb3 {+3.08/20 3.0s} 94. Rc8+ {-3.52/14 1.8s} Kd4 {+3.07/20 3.0s}
95. Rd8+ {-3.51/14 3.3s} Ke5 {+3.08/19 3.0s} 96. Nc3 {-3.52/11 1.9s}
Rg2+ {+3.13/14 3.0s} 97. Kh5 {-3.71/11 2.0s} Rc2 {+3.13/16 3.0s}
98. Nb1 {-3.79/12 1.7s} Nd4 {+3.10/15 3.0s} 99. Rb8 {-3.50/13 3.5s}
Kf5 {+3.03/15 3.0s} 100. Kh4 {-3.39/13 2.4s} Kf4 {+3.15/15 3.0s}
101. Kh3 {-3.33/13 2.1s} Nc6 {+3.28/16 3.0s} 102. Rb6 {-3.25/13 2.2s}
Ne5 {+3.25/16 3.0s} 103. Rb4+ {-3.50/13 2.6s} Kf3 {+3.16/18 3.0s}
104. Rb3+ {-3.51/13 3.7s} Ke4 {+3.11/18 3.0s} 105. Rb4+ {-3.62/13 2.2s}
Kd5 {+3.11/18 3.0s} 106. Rb5+ {-3.73/14 3.4s} Kd4 {+3.13/19 3.0s}
107. Rb4+ {-3.99/14 4.1s} Nc4 {+3.13/20 3.0s} 108. Kh4 {-4.12/14 4.1s}
Kc5 {+3.27/22 3.0s} 109. Rb3 {-4.08/15 2.7s} Rc1 {+3.27/20 3.0s}
110. Na3 {-4.14/14 2.4s} Nxa3 {+3.15/20 3.0s} 111. Rxb2 {-4.08/13 2.3s}
Rg1 {+3.18/20 3.0s} 112. Kh3 {-4.10/14 3.7s} Nc4 {+3.30/15 3.0s}
113. Rg2 {-4.12/13 3.3s} Rd1 {+3.20/17 3.0s} 114. Rg5+ {-4.07/13 3.6s}
Kd4 {+3.30/17 3.0s} 115. Rg4+ {-4.19/13 2.8s} Kd5 {+3.25/18 3.0s}
116. Kg3 {-4.22/12 4.8s} Ne5 {+3.27/18 3.0s} 117. Ra4 {-4.23/13 6.2s}
Rg1+ {+3.25/20 3.0s} 118. Kf2 {-4.25/13 4.1s} Rh1 {+3.25/19 3.0s}
119. Ra8 {-4.22/13 4.0s} Rh3 {+3.29/18 3.0s} 120. Ra5+ {-4.23/15 10.0s}
Kd4 {+3.29/20 2.5s} 121. Ra4+ {-4.22/14 2.2s} Nc4 {+3.29/19 3.0s}
122. Ra7 {-4.20/11 2.1s} Ke4 {+3.31/18 3.0s} 123. Ra4 {-4.22/13 2.2s}
Rf3+ {+3.32/21 3.0s} 124. Kg2 {-4.20/13 2.0s} Rc3 {+3.32/20 3.0s}
125. Kf2 {-4.21/13 2.4s} Kf4 {+3.31/22 3.0s} 126. Ke2 {-4.22/14 2.4s}
Kg3 {+3.31/21 3.0s} 127. Ra8 {-4.21/13 2.6s} Ne5 {+3.31/19 3.0s}
128. Kd2 {-4.01/13 3.1s} Rb3 {+3.31/22 3.0s} 129. Kc2 {-4.04/12 2.5s}
Re3 {+3.30/20 3.0s} 130. Rc8 {-4.02/12 3.1s} Kf4 {+3.31/18 3.0s}
131. Rc3 {-4.02/12 2.0s} Re2+ {+3.31/17 3.0s} 132. Kb3 {-4.03/12 1.8s}
Re1 {+3.30/19 3.0s} 133. Rc8 {-4.03/12 2.6s} Kg3 {+3.30/18 3.0s}
134. Rf8 {-3.99/12 2.8s} Rc1 {+3.30/19 3.0s} 135. Rd8 {-4.01/12 3.2s}
Kf3 {+3.29/20 3.0s} 136. Rf8+ {-4.01/14 2.7s} Ke4 {+3.29/18 3.0s}
137. Re8 {-3.99/14 3.3s} Rb1+ {+3.31/19 3.0s} 138. Kc2 {-3.82/13 2.2s}
Rb4 {+3.29/21 3.0s, No result} *[/pgn]

[algerbrex]

Blunder 7.4.0 has been released. It has about 50 Elo over 7.3.0 and 100 Elo over 7.1.0. Downloads and release notes can be found here, per usual: https://github.com/algerbrex/blunder/re ... tag/v7.4.0

My current estimate for 7.4.0's strength is 2500+ Elo.

As I've now created a blog for Blunder to have my own space for my ramblings and typos: https://algerbrex.github.io/

For now, I decide to use GitHub Pages as it seemed to be exactly what I needed right now. Nothing fancy, just a simple, static place to make blog posts.

[jtwright]

Exciting! Congratulations on a new release. Cool to see your thoughts and experience with king safety. I really appreciate when people document their process of improving parts of the engine (though I'm guilty of not doing this myself...), often you just see the final product and don't know how they got there.

[algerbrex]

Exciting! Congratulations on a new release. Cool to see your thoughts and experience with king safety. I really appreciate when people document their process of improving parts of the engine (though I'm guilty of not doing this myself...), often you just see the final product and don't know how they got there.

Thanks! I wholeheartedly agree with your sentiment, although I'm also pretty guilty since I only now just started documenting my progress on Blunder

Part of the reason I decided to talk about king safety is that it was one of the harder parts to even get working decently in Blunder, and it took many hours of tinkering and testing before I started to see any decent results. So hopefully, in some way, I can help the next person who comes along get to this same point quicker and less painfully than I did.

[Guenther]

Blunder 7.4.0 has been released. It has about 50 Elo over 7.3.0 and 100 Elo over 7.1.0. Downloads and release notes can be found here, per usual: https://github.com/algerbrex/blunder/re ... tag/v7.4.0

My current estimate for 7.4.0's strength is 2500+ Elo.

As I've now created a blog for Blunder to have my own space for my ramblings and typos: https://algerbrex.github.io/

For now, I decide to use GitHub Pages as it seemed to be exactly what I needed right now. Nothing fancy, just a simple, static place to make blog posts.

Thanks for the new release! I have added the blog link to the chronology.

[algerbrex]

Blunder 7.4.0 has been released. It has about 50 Elo over 7.3.0 and 100 Elo over 7.1.0. Downloads and release notes can be found here, per usual: https://github.com/algerbrex/blunder/re ... tag/v7.4.0

My current estimate for 7.4.0's strength is 2500+ Elo.

As I've now created a blog for Blunder to have my own space for my ramblings and typos: https://algerbrex.github.io/

For now, I decide to use GitHub Pages as it seemed to be exactly what I needed right now. Nothing fancy, just a simple, static place to make blog posts.

Thanks for the new release! I have added the blog link to the chronology.

Thanks, Guenther! Much appreciated.

[Rebel]

Let's try again, Blunder 7.4.0 for the GRL.

Games : 1600

Elo pool : 2499

http://rebel13.nl/a/grl.htm

What's the name of the logfile so I can check?

[Rebel]

Another crash, can't discover a logfile.

[algerbrex]

Another crash, can't discover a logfile.

An damnit. Knew I forgot something. I took that code out because I wanted to clean it up before I merged it into main and never added it back in. That’s why no log file. The code’s still sitting on my laptop.

Since this seems to be a consistent crash that keeps happening, I might just release a hotfix version with the logging code so we can finally get to the bottom of this.

In the meanwhile, would it not be possible to use the stderr option with cutechess to produce a log file for Blunder specifically? Cutechess should log whatever exception Blunder is crashing with there.

[amanjpro]

Another crash, can't discover a logfile.

An damnit. Knew I forgot something. I took that code out because I wanted to clean it up before I merged it into main and never added it back in. That’s why no log file. The code’s still sitting on my laptop.

Since this seems to be a consistent crash that keeps happening, I might just release a hotfix version with the logging code so we can finally get to the bottom of this.

In the meanwhile, would it not be possible to use the stderr option with cutechess to produce a log file for Blunder specifically? Cutechess should log whatever exception Blunder is crashing with there.

Ed is running the matches in 40/120 TC. Have you tried to run a selfplay match between with a similar TC with `-debug` flag enabled? Leave it overnight. A day, a week, you will catch it almost certainly.

You can make blunder very noisy and write debugging lines in a file. You are not interested in speed, you are interested in verbosity at this point