Komodo Elo = 3496 vs Berserk-9-dev after 20 games so far

Discussion of computer chess matches and engine tournaments.

Moderators: hgm, Dann Corbit, Harvey Williamson

Chessqueen
Posts: 5479
Joined: Wed Sep 05, 2018 2:16 am
Location: Moving
Full name: Jorge Picado

Komodo Elo = 3496 vs Berserk-9-dev after 20 games so far

Post by Chessqueen »

So far I have run Komodo Elo = 3496 vs Berserk-9-dev and after 20 games the score is dead even at T/C 15/10, I am NOT implying that Berserk at this time control is rated 3496, but since it takes the full time control to make all of its moves while Komodo Dragon 2.6.1 Elo = 3496 play instantly is leveling out the score so far. I will continue this match up to 100 games to get an idea, NOT to determine the strength of Berserk -9-dev, it could be that the real rating at T/C 15/10 lie between 3485 to 3490 after CCRL test it against different engines :roll:

[pgn][Event "Komodo Dragon Elo = 3496"]
[Site "DESKTOP-OFQ3C0P"]
[Date "2022.06.12"]
[Round "20"]
[White "Dragon-2.6.1-64bit-avx2"]
[Black "Berserk-9-dev-x64-avx2"]
[Result "1/2-1/2"]
[ECO "B90"]
[Opening "Sicilian"]
[Time "10:38:56"]
[Variation "Najdorf, 6.Be3 e5 7.Nb3 Be6 8.f3"]
[WhiteElo "3497"]
[TimeControl "900+10"]
[Termination "adjudication"]
[PlyCount "181"]


1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. f3 e5 7. Nb3 Be6 8.
Be3 h5 9. Qd2 Nbd7 10. Nd5 Nxd5 11. exd5 Bf5 12. Na5 Be7 13. Be2 Bh4+ 14.
g3 Be7 15. Nxb7 Qb8 16. Na5 Qxb2 17. Rc1 Qxa2 18. Bc4 Qa4 19. Bb3 Qb5 20.
c4 Qb8 21. Ba4 Qc7 22. Nc6 Bh3 23. Kf2 Bf6 24. Qa5 Qxa5 25. Nxa5 O-O 26.
Rb1 Rfc8 27. Bc6 Rab8 28. Ba7 Rxb1 29. Rxb1 Bd8 30. Bb7 Bf5 31. Rb2 Nc5 32.
Bxc8 Nd3+ 33. Ke2 Nxb2 34. Bxf5 Bxa5 35. c5 dxc5 36. Bb8 g6 37. Bc8 Na4 38.
d6 c4 39. Bxa6 c3 40. Kd1 Nc5 41. Bc4 Kf8 42. Bc7 Bb4 43. Bb5 Nb7 44. d7
Ke7 45. Bxe5 Nd8 46. Bd4 Ne6 47. Bb6 Bd6 48. Ba5 h4 49. f4 c2+ 50. Kxc2
hxg3 51. hxg3 Bc7 52. Bb4+ Bd6 53. Bc3 Bc7 54. Bc6 Nc5 55. Bb4 Bd6 56. Bxc5
Bxc5 57. Kd3 Bd6 58. Kd4 Bc7 59. Kd5 f5 60. Bb5 Bb6 61. Ba4 Bf2 62. Kc6 Kd8
63. Bb3 Bd4 64. Bg8 Bf2 65. Bb3 Bxg3 66. Bf7 Bxf4 67. Bxg6 Be3 68. Bxf5 Bf4
69. Kd5 Bc1 70. Ke5 Bb2+ 71. Kd5 Bc1 72. Ke5 Bg5 73. Bg4 Be7 74. Ke6 Bh4
75. Bh3 Bg5 76. Bf1 Bf4 77. Bb5 Bc7 78. Kf5 Ba5 79. Kf4 Ke7 80. Ke3 Bd8 81.
Ke2 Ke6 82. Kd1 Kd6 83. Ke2 Ke7 84. Ba4 Bc7 85. Kd1 Kd8 86. Ke2 Be5 87. Kd2
Bd4 88. Kd1 Bh8 89. Bb5 Be5 90. Bc6 Bg3 91. Ke2 {User Adjudication} 1/2-1/2[/pgn]
Forget about memorization of Opening Theories https://www.youtube.com/watch?v=DN3381sdcdY
Chessqueen
Posts: 5479
Joined: Wed Sep 05, 2018 2:16 am
Location: Moving
Full name: Jorge Picado

Re: Komodo Elo = 3496 vs Berserk-9-dev after 20 games so far

Post by Chessqueen »

So far the score after 30 games is still draw 15 to 15 at T/C 15/10
[pgn][Event "Komodo Elo = 3496"]
[Site "DESKTOP-OFQ3C0P"]
[Date "2022.06.12"]
[Round "28"]
[White "Berserk-9-dev-x64-avx2"]
[Black "Dragon-2.6.1-64bit-avx2"]
[Result "1/2-1/2"]
[BlackElo "3496"]
[ECO "D12"]
[Opening "Slav"]
[Time "12:18:33"]
[Variation "4.e3 Bf5 5.Nc3 e6 6.Nh4 Be4"]
[WhiteElo "3493"]
[TimeControl "900+10"]
[Termination "adjudication"]
[PlyCount "346"]


1. Nf3 Nf6 2. d4 d5 3. c4 c6 4. e3 Bf5 5. Nc3 e6 6. Nh4 Be4 7. f3 Bg6 8.
Qb3 Qc7 9. Bd2 Nbd7 10. cxd5 Nxd5 11. g3 Be7 12. e4 Nxc3 13. Nxg6 hxg6 14.
bxc3 b5 15. Be2 a6 16. Kf1 O-O 17. Kg2 g5 18. a4 Nb6 19. axb5 axb5 20. Rhb1
Rfb8 21. Be3 Ra4 22. Qd1 Rba8 23. d5 Rxa1 24. Rxa1 Rxa1 25. Qxa1 cxd5 26.
Bxb6 Qxb6 27. exd5 Qe3 28. Qd1 exd5 29. Bxb5 Qc5 30. Qd3 Bf6 31. c4 d4 32.
Qe4 g6 33. Qc6 Be7 34. Qxc5 Bxc5 35. Kf2 d3+ 36. Ke1 Be3 37. Ba4 f5 38. Kd1
Kf7 39. Bd7 Kf6 40. Be8 g4 41. fxg4 fxg4 42. Ke1 Bg1 43. Kd2 Bxh2 44. Kxd3
Bxg3 45. Ke2 Bc7 46. Ke3 Kf5 47. Bd7+ Kg5 48. Kf2 g3+ 49. Kg2 Kf4 50. Bc6
Ke3 51. Bb7 Be5 52. Kg1 Kd3 53. Bd5 Kd2 54. Bb7 Kc3 55. Bd5 Kb2 56. Kg2 Bc7
57. Be6 Kc2 58. Kh1 Kd2 59. Kg1 Bf4 60. Bd7 Bc7 61. Be6 g5 62. Kg2 Kc2 63.
Bc8 Kc3 64. Be6 Kb4 65. Bg8 Ka5 66. Bf7 Ka6 67. Bh5 Bf4 68. Bg6 Be5 69. Be4
Bd6 70. Kg1 Ka7 71. Kg2 Bb8 72. Bd5 Ka6 73. Be6 Kb6 74. Bh3 Kc6 75. Bc8 Kc5
76. Be6 Kb4 77. Bd5 Kb3 78. Bc6 Kxc4 79. Ba4 Kd5 80. Bd1 Be5 81. Bc2 Kc6
82. Ba4+ Kd6 83. Bb5 Ke7 84. Bc6 Ke6 85. Be4 Kf7 86. Bd5+ Ke8 87. Bc4 Bd6
88. Bd3 Bb8 89. Be2 Ke7 90. Bc4 Bf4 91. Ba6 Kf7 92. Be2 Bc7 93. Bb5 Bb8 94.
Bc4+ Ke7 95. Bf1 Bd6 96. Bd3 Ke8 97. Bc2 Kf8 98. Bd1 Kg8 99. Kg1 Kg7 100.
Bb3 Bc7 101. Ba4 Be5 102. Bb5 Kg8 103. Ba6 Kh7 104. Bb7 Bd6 105. Bc8 Bb8
106. Kg2 Be5 107. Bg4 Kg7 108. Be6 Bb8 109. Bh3 Kf8 110. Be6 Be5 111. Kg1
Ke7 112. Ba2 Bd6 113. Kg2 Kf6 114. Kf3 Bc7 115. Bd5 Kf5 116. Kg2 Bb8 117.
Bb3 g4 118. Bc4 Be5 119. Bg8 Bd6 120. Bh7+ Ke6 121. Bc2 Bc7 122. Ba4 Kf5
123. Bb3 Bf4 124. Bd5 Bd6 125. Bc4 Kg5 126. Bb5 Kh4 127. Bc6 Bc7 128. Be4
Kg5 129. Bc6 Bf4 130. Kg1 Kh4 131. Kh1 Bc7 132. Bg2 Kg5 133. Bf1 Bb8 134.
Bg2 Kf6 135. Kg1 Bd6 136. Bh1 Kg7 137. Bc6 Kh6 138. Bb7 Bc7 139. Bg2 Bd6
140. Kh1 Bb4 141. Bc6 Be1 142. Kg2 Kg5 143. Bb5 Kh4 144. Ba6 Kh5 145. Bc4
Kh4 146. Be6 Bb4 147. Bf5 Ba5 148. Bg6 Bd8 149. Bf5 Bb6 150. Bb1 Bc7 151.
Ba2 Kg5 152. Bd5 Kf6 153. Bg8 Be5 154. Bh7 Bc7 155. Bd3 Kg7 156. Bc4 Kh6
157. Bf7 Bb8 158. Bd5 Bf4 159. Bb3 Kh7 160. Be6 Bc7 161. Bd7 Bb8 162. Bc6
Kh6 163. Bd7 Bd6 164. Bxg4 Kh7 165. Bh3 Be5 166. Bf5+ Kh6 167. Bc2 Kh5 168.
Bf5 Bb8 169. Bd3 Kg5 170. Bb5 Kf5 171. Bc6 Bc7 172. Kg1 Kg6 173. Bf3 Be5
{User Adjudication} 1/2-1/2[/pgn]
Last edited by Chessqueen on Sun Jun 12, 2022 8:18 pm, edited 3 times in total.
Forget about memorization of Opening Theories https://www.youtube.com/watch?v=DN3381sdcdY
Chessqueen
Posts: 5479
Joined: Wed Sep 05, 2018 2:16 am
Location: Moving
Full name: Jorge Picado

Re: Komodo Elo = 3496 vs Berserk-9-dev after 20 games so far

Post by Chessqueen »

So far the score after 30 games is still draw 15 to 15 at T/C 15/10
[pgn][Event "Komodo Elo = 3496"]
[Site "DESKTOP-OFQ3C0P"]
[Date "2022.06.12"]
[Round "30"]
[White "Berserk-9-dev-x64-avx2"]
[Black "Dragon-2.6.1-64bit-avx2"]
[Result "1/2-1/2"]
[BlackElo "3496"]
[ECO "B10"]
[Opening "Caro-Kann"]
[Time "11:25:33"]
[Variation "English, Exchange, 4...Nf6"]
[TimeControl "900+10"]
[Termination "adjudication"]
[PlyCount "316"]


1. c4 c6 2. e4 d5 3. exd5 cxd5 4. cxd5 Nf6 5. Qa4+ Nbd7 6. Bc4 g6 7. Nc3
Bg7 8. d3 a6 9. Nf3 Rb8 10. Bb3 b5 11. Qh4 Nc5 12. O-O Bb7 13. d6 exd6 14.
Re1+ Kf8 15. Be3 Nxb3 16. axb3 h6 17. Qb4 Kg8 18. Nd2 Bf8 19. Qh4 Rh7 20.
b4 g5 21. Qd4 Nd7 22. Nde4 Ne5 23. f4 gxf4 24. Bxf4 Rg7 25. Rf1 Rg6 26. Qa7
Nxd3 27. Bg3 Bg7 28. Rad1 Ne5 29. Rf2 Qf8 30. Bxe5 Bxe5 31. Ng3 Qc8 32. Nf5
Re6 33. Kh1 Kh7 34. Ne3 Re8 35. Ncd5 Qe6 36. Kg1 Bg7 37. h3 Kh8 38. Kh2
Red8 39. Kh1 Rdc8 40. Nf4 Qd7 41. Ned5 Re8 42. Kg1 Ra8 43. Qb6 Rac8 44. Qa7
Rcd8 45. Qb6 Qc6 46. Qxc6 Bxc6 47. Rc2 Bd7 48. Nh5 Bf5 49. Rcd2 Be5 50. Ra1
Rc8 51. Rxa6 Rc4 52. Nhf6 Rb8 53. Kf2 Bc8 54. Ra3 Be6 55. Kf3 Kg7 56. Nh5+
Kh7 57. Nhf4 Bc8 58. b3 Rc1 59. Nd3 Rf1+ 60. Ke3 Bg3 61. Ke2 Rf5 62. Ne3
Rg5 63. Ra7 Rb7 64. Rxb7 Bxb7 65. Rc2 Kg7 66. Rc7 Be4 67. Nf2 Bg6 68. Rb7
h5 69. Kf3 h4 70. Ne4 Bh5+ 71. Ng4 Rf5+ 72. Ke3 Rf1 73. Nxg3 hxg3 74. Rd7
Bxg4 75. hxg4 Rf2 76. Rxd6 Rxg2 77. Kf3 Rb2 78. Kxg3 Rxb3+ 79. Kf2 Rxb4 80.
g5 Rb2+ 81. Kf3 Rb1 82. Rb6 b4 83. Kg3 Kf8 84. Rb7 b3 85. Kf2 b2 86. Kg2
Kg7 87. Rb6 Kh7 88. Kh2 Kh8 89. Rh6+ Kg8 90. Rb6 Kf8 91. Rb5 Ke7 92. Rb8
Ke6 93. Rb7 Ke5 94. Kg2 Kf5 95. Rxf7+ Kg4 96. Rb7 Kxg5 97. Rb8 Kf5 98. Rb6
Ke5 99. Rb8 Kd5 100. Rb7 Kd4 101. Rd7+ Ke4 102. Rb7 Kd4 103. Rd7+ Kc4 104.
Rb7 Kc3 105. Rc7+ Kd3 106. Rb7 Ke4 107. Rb5 Ke3 108. Rb8 Kd4 109. Rb3 Kc5
110. Rb8 Kd6 111. Rb6+ Kc7 112. Rb5 Kd7 113. Rb6 Ke7 114. Rb4 Kd7 115. Rb6
Ke7 116. Rb4 Kf6 117. Rb3 Kf5 118. Rb7 Kg5 119. Rb5+ Kg4 120. Rb4+ Kh5 121.
Kh2 Kg6 122. Kg2 Kf7 123. Rb6 Rd1 124. Rb7+ Kf6 125. Rb6+ Kg7 126. Rxb2 Kf7
127. Rb7+ Ke6 128. Rb3 Kd6 129. Kf3 Rd2 130. Re3 Rc2 131. Rd3+ Kc6 132. Ke3
Ra2 133. Rd1 Ra5 134. Kf2 Rf5+ 135. Ke3 Re5+ 136. Kf2 Kb5 137. Rc1 Re4 138.
Rd1 Rc4 139. Ke2 Rc3 140. Rd8 Kb6 141. Kd2 Rc5 142. Kd1 Ra5 143. Rc8 Rg5
144. Kd2 Rg1 145. Kd3 Kb7 146. Rc2 Rg3+ 147. Ke2 Rg2+ 148. Kd3 Rg3+ 149.
Kd2 Kb6 150. Rc4 Rg2+ 151. Ke1 Rb2 152. Rc1 Kb7 153. Rc5 Rb1+ 154. Ke2 Rg1
155. Rc3 Rg2+ 156. Ke1 Rb2 157. Rc5 Rh2 158. Rc4 Rh1+ {User Adjudication}[/pgn]
Forget about memorization of Opening Theories https://www.youtube.com/watch?v=DN3381sdcdY
Chessqueen
Posts: 5479
Joined: Wed Sep 05, 2018 2:16 am
Location: Moving
Full name: Jorge Picado

Re: Komodo Elo = 3496 vs Berserk-9-dev after 20 games so far

Post by Chessqueen »

I decided to stop the match at 50 games since the score is so close 26 to 24 in favor of Berserk -9-dev against Komodo Dragon Elo = 3496 at T/C 15/10. Now for anybody believing that those are too few games are welcome to continue watching game after game after game being ending in draws :roll:

[pgn][Event "Dragon Elo = 3496"]
[Site "DESKTOP-OFQ3C0P"]
[Date "2022.06.12"]
[Round "50"]
[White "Dragon-2.6.1-64bit-avx2"]
[Black "Berserk-9-dev-x64-avx2"]
[Result "1/2-1/2"]
[ECO "D47"]
[Opening "Semi-Slav"]
[Time "18:40:25"]
[Variation "Meran, Wade, Main Line, 12.O-O"]
[WhiteElo "3496"]
[TimeControl "900+10"]
[Termination "normal"]
[PlyCount "372"]


1. d4 d5 2. c4 c6 3. Nf3 Nf6 4. Nc3 e6 5. e3 Nbd7 6. Bd3 dxc4 7. Bxc4 b5 8.
Bd3 Bb7 9. e4 b4 10. Na4 c5 11. e5 Nd5 12. O-O cxd4 13. Re1 a6 14. Bd2 g6
15. Be4 Rb8 16. Qc2 Qa5 17. a3 Be7 18. Bh6 Bf8 19. Bd2 Be7 20. Bh6 bxa3 21.
bxa3 Rc8 22. Qd1 Bc6 23. Rc1 Nc3 24. Nxc3 dxc3 25. Bxc6 Rxc6 26. h3 Qc7 27.
Qb3 Bf8 28. Bxf8 Kxf8 29. Re3 Kg7 30. Rexc3 Rxc3 31. Rxc3 Qa5 32. Rc6 Rb8
33. Qd3 Nxe5 34. Qd4 Rb5 35. Rd6 Kf8 36. Nxe5 Qe1+ 37. Kh2 Qxe5+ 38. Qxe5
Rxe5 39. Rxa6 g5 40. f3 Re2 41. Ra5 h6 42. h4 gxh4 43. Rh5 Re3 44. Rxh6 Kg7
45. Rxh4 Rxa3 46. Rb4 e5 47. Kg3 Kg6 48. Kg4 f6 49. f4 exf4 50. Rxf4 Rb3
51. Ra4 Rc3 52. Ra8 Rc7 53. Rg8+ Kh7 54. Rb8 Kg7 55. Rb5 Kg6 56. Kf4 Re7
57. Rb6 Re2 58. Kf3 Re8 59. Rb3 Rh8 60. Rb6 Re8 61. Kg3 Rg8 62. Rb2 Rg7 63.
Rc2 Rb7 64. Rc3 Kg7 65. Kf4 Kg6 66. Kg4 Re7 67. Kf3 Kh6 68. Kf2 Kg7 69. Rb3
Re6 70. Rb5 Re8 71. Rb4 Kg6 72. Rg4+ Kf7 73. g3 Rc8 74. Ra4 Rc2+ 75. Ke3
Rb2 76. Kf3 Rb3+ 77. Kg2 Kg6 78. Ra2 Rc3 79. Rf2 Kg5 80. Re2 Rb3 81. Rc2
Rb6 82. Rc5+ Kg6 83. Rc4 Kg5 84. Rc7 Kg6 85. Kh3 Kh6 86. Rc4 Kg6 87. Rc2
Rb8 88. Re2 Rh8+ 89. Kg4 Rg8 90. Kf4 Rg7 91. Rc2 Re7 92. Rc3 Kg7 93. Rc4
Rb7 94. Ke3 Rd7 95. Rh4 Re7+ 96. Kf2 Rc7 97. Kg2 Rc6 98. Ra4 Kg6 99. Rb4
Rc2+ 100. Kh3 Rc8 101. Rg4+ Kf7 102. Rh4 Kg7 103. Rg4+ Kh7 104. Rd4 Ra8
105. Rc4 Ra7 106. Kg2 Re7 107. Rg4 Re5 108. Rc4 Re7 109. Rc8 Kg6 110. Rc2
Re4 111. Kf3 Ra4 112. Re2 Ra3+ 113. Kf2 Rc3 114. Rb2 Rc6 115. Rb3 Ra6 116.
Ke3 Kg5 117. Kf3 Ra2 118. Rb5+ f5 119. Rb8 Ra3+ 120. Kg2 Ra2+ 121. Kh3 Ra4
122. Rg8+ Kh5 123. Rf8 Kg5 124. Rf7 Ra8 125. Rg7+ Kf6 126. Rh7 Rc8 127. Rh4
Re8 128. Kg2 Rg8 129. Rb4 Kg5 130. Rb5 Ra8 131. Rb6 Re8 132. Kf2 Ra8 133.
Rb2 Kg4 134. Rb4+ Kg5 135. Rb3 f4 136. gxf4+ Kxf4 137. Rb4+ Kg5 138. Rb6
Re8 139. Kf3 Rg8 140. Re6 Rg7 141. Re5+ Kf6 142. Re8 Rf7 143. Rc8 Rh7 144.
Rc6+ Kf5 145. Rc5+ Kg6 146. Kg4 Re7 147. Rc6+ Kf7 148. Rc4 Kg6 149. Rc5 Kf7
150. Kg3 Re6 151. Kf2 Ke8 152. Rc8+ Kd7 153. Rb8 Ra6 154. Rb7+ Ke8 155. Kf3
Rf6+ 156. Ke2 Kf8 157. Rc7 Rh6 158. Rc8+ Ke7 159. Ra8 Rg6 160. Rc8 Rg5 161.
Kf2 Re5 162. Rb8 Ra5 163. Rb7+ Kf6 164. Ke1 Ke6 165. Ke2 Ra2+ 166. Kd3 Ra3+
167. Kc2 Ra2+ 168. Kb1 Rd2 169. Rb6+ Kd7 170. Kc1 Rd5 171. Rb7+ Kd8 172.
Rf7 Rg5 173. Kd2 Rb5 174. Ke2 Rb2+ 175. Kf3 Ke8 176. Rh7 Rb5 177. Rh8+ Ke7
178. Rc8 Re5 179. Rc7+ Kf6 180. Kf2 Rd5 181. Rc8 Rf5+ 182. Ke2 Rg5 183.
Rc6+ Ke7 184. Ra6 Re5+ 185. Kf3 Kf7 186. Ra7+ Kg8 1/2-1/2[/pgn]
Forget about memorization of Opening Theories https://www.youtube.com/watch?v=DN3381sdcdY
lkaufman
Posts: 5942
Joined: Sun Jan 10, 2010 6:15 am
Location: Maryland USA

Re: Komodo Elo = 3496 vs Berserk-9-dev after 20 games so far

Post by lkaufman »

Chessqueen wrote: Sun Jun 12, 2022 6:38 pm So far I have run Komodo Elo = 3496 vs Berserk-9-dev and after 20 games the score is dead even at T/C 15/10, I am NOT implying that Berserk at this time control is rated 3496, but since it takes the full time control to make all of its moves while Komodo Dragon 2.6.1 Elo = 3496 play instantly is leveling out the score so far. I will continue this match up to 100 games to get an idea, NOT to determine the strength of Berserk -9-dev, it could be that the real rating at T/C 15/10 lie between 3485 to 3490 after CCRL test it against different engines :roll:

[pgn][Event "Komodo Dragon Elo = 3496"]
[Site "DESKTOP-OFQ3C0P"]
[Date "2022.06.12"]
[Round "20"]
[White "Dragon-2.6.1-64bit-avx2"]
[Black "Berserk-9-dev-x64-avx2"]
[Result "1/2-1/2"]
[ECO "B90"]
[Opening "Sicilian"]
[Time "10:38:56"]
[Variation "Najdorf, 6.Be3 e5 7.Nb3 Be6 8.f3"]
[WhiteElo "3497"]
[TimeControl "900+10"]
[Termination "adjudication"]
[PlyCount "181"]


1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. f3 e5 7. Nb3 Be6 8.
Be3 h5 9. Qd2 Nbd7 10. Nd5 Nxd5 11. exd5 Bf5 12. Na5 Be7 13. Be2 Bh4+ 14.
g3 Be7 15. Nxb7 Qb8 16. Na5 Qxb2 17. Rc1 Qxa2 18. Bc4 Qa4 19. Bb3 Qb5 20.
c4 Qb8 21. Ba4 Qc7 22. Nc6 Bh3 23. Kf2 Bf6 24. Qa5 Qxa5 25. Nxa5 O-O 26.
Rb1 Rfc8 27. Bc6 Rab8 28. Ba7 Rxb1 29. Rxb1 Bd8 30. Bb7 Bf5 31. Rb2 Nc5 32.
Bxc8 Nd3+ 33. Ke2 Nxb2 34. Bxf5 Bxa5 35. c5 dxc5 36. Bb8 g6 37. Bc8 Na4 38.
d6 c4 39. Bxa6 c3 40. Kd1 Nc5 41. Bc4 Kf8 42. Bc7 Bb4 43. Bb5 Nb7 44. d7
Ke7 45. Bxe5 Nd8 46. Bd4 Ne6 47. Bb6 Bd6 48. Ba5 hBerserk 9 4 49. f4 c2+ 50. Kxc2
hxg3 51. hxg3 Bc7 52. Bb4+ Bd6 53. Bc3 Bc7 54. Bc6 Nc5 55. Bb4 Bd6 56. Bxc5
Bxc5 57. Kd3 Bd6 58. Kd4 Bc7 59. Kd5 f5 60. Bb5 Bb6 61. Ba4 Bf2 62. Kc6 Kd8
63. Bb3 Bd4 64. Bg8 Bf2 65. Bb3 Bxg3 66. Bf7 Bxf4 67. Bxg6 Be3 68. Bxf5 Bf4
69. Kd5 Bc1 70. Ke5 Bb2+ 71. Kd5 Bc1 72. Ke5 Bg5 73. Bg4 Be7 74. Ke6 Bh4
75. Bh3 Bg5 76. Bf1 Bf4 77. Bb5 Bc7 78. Kf5 Ba5 79. Kf4 Ke7 80. Ke3 Bd8 81.
Ke2 Ke6 82. Kd1 Kd6 83. Ke2 Ke7 84. Ba4 Bc7 85. Kd1 Kd8 86. Ke2 Be5 87. Kd2
Bd4 88. Kd1 Bh8 89. Bb5 Be5 90. Bc6 Bg3 91. Ke2 {User Adjudication} 1/2-1/2[/pgn]
If Komodo Dragon 2.6.1 is moving instantly with Elo set to 3496, something is terribly wrong, and if it is holding a strong engine like Berserk 9 roughly even without having to think that is totally amazing. I can't offer any explanation for either of these two issues, and since we are already on Dragon 3 we would only be concerned if the issue is still present in Dragon 3. Maybe I'll test than myself. If I had to guess, I would suspect that any setting above 3400 but below 3500 just doesn't work properly, perhaps it just reverts to 3400 or even 3300 for example. A setting of 3496 is supposed to be almost the same as just using one thread at full strength, where I would expect it to beat Berserk 9 quite convincingly, but not without taking a similar amount of time.
Komodo rules!
lkaufman
Posts: 5942
Joined: Sun Jan 10, 2010 6:15 am
Location: Maryland USA

Re: Komodo Elo = 3496 vs Berserk-9-dev after 20 games so far

Post by lkaufman »

lkaufman wrote: Mon Jun 13, 2022 4:43 am
Chessqueen wrote: Sun Jun 12, 2022 6:38 pm So far I have run Komodo Elo = 3496 vs Berserk-9-dev and after 20 games the score is dead even at T/C 15/10, I am NOT implying that Berserk at this time control is rated 3496, but since it takes the full time control to make all of its moves while Komodo Dragon 2.6.1 Elo = 3496 play instantly is leveling out the score so far. I will continue this match up to 100 games to get an idea, NOT to determine the strength of Berserk -9-dev, it could be that the real rating at T/C 15/10 lie between 3485 to 3490 after CCRL test it against different engines :roll:

[pgn][Event "Komodo Dragon Elo = 3496"]
[Site "DESKTOP-OFQ3C0P"]
[Date "2022.06.12"]
[Round "20"]
[White "Dragon-2.6.1-64bit-avx2"]
[Black "Berserk-9-dev-x64-avx2"]
[Result "1/2-1/2"]
[ECO "B90"]
[Opening "Sicilian"]
[Time "10:38:56"]
[Variation "Najdorf, 6.Be3 e5 7.Nb3 Be6 8.f3"]
[WhiteElo "3497"]
[TimeControl "900+10"]
[Termination "adjudication"]
[PlyCount "181"]


1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. f3 e5 7. Nb3 Be6 8.
Be3 h5 9. Qd2 Nbd7 10. Nd5 Nxd5 11. exd5 Bf5 12. Na5 Be7 13. Be2 Bh4+ 14.
g3 Be7 15. Nxb7 Qb8 16. Na5 Qxb2 17. Rc1 Qxa2 18. Bc4 Qa4 19. Bb3 Qb5 20.
c4 Qb8 21. Ba4 Qc7 22. Nc6 Bh3 23. Kf2 Bf6 24. Qa5 Qxa5 25. Nxa5 O-O 26.
Rb1 Rfc8 27. Bc6 Rab8 28. Ba7 Rxb1 29. Rxb1 Bd8 30. Bb7 Bf5 31. Rb2 Nc5 32.
Bxc8 Nd3+ 33. Ke2 Nxb2 34. Bxf5 Bxa5 35. c5 dxc5 36. Bb8 g6 37. Bc8 Na4 38.
d6 c4 39. Bxa6 c3 40. Kd1 Nc5 41. Bc4 Kf8 42. Bc7 Bb4 43. Bb5 Nb7 44. d7
Ke7 45. Bxe5 Nd8 46. Bd4 Ne6 47. Bb6 Bd6 48. Ba5 hBerserk 9 4 49. f4 c2+ 50. Kxc2
hxg3 51. hxg3 Bc7 52. Bb4+ Bd6 53. Bc3 Bc7 54. Bc6 Nc5 55. Bb4 Bd6 56. Bxc5
Bxc5 57. Kd3 Bd6 58. Kd4 Bc7 59. Kd5 f5 60. Bb5 Bb6 61. Ba4 Bf2 62. Kc6 Kd8
63. Bb3 Bd4 64. Bg8 Bf2 65. Bb3 Bxg3 66. Bf7 Bxf4 67. Bxg6 Be3 68. Bxf5 Bf4
69. Kd5 Bc1 70. Ke5 Bb2+ 71. Kd5 Bc1 72. Ke5 Bg5 73. Bg4 Be7 74. Ke6 Bh4
75. Bh3 Bg5 76. Bf1 Bf4 77. Bb5 Bc7 78. Kf5 Ba5 79. Kf4 Ke7 80. Ke3 Bd8 81.
Ke2 Ke6 82. Kd1 Kd6 83. Ke2 Ke7 84. Ba4 Bc7 85. Kd1 Kd8 86. Ke2 Be5 87. Kd2
Bd4 88. Kd1 Bh8 89. Bb5 Be5 90. Bc6 Bg3 91. Ke2 {User Adjudication} 1/2-1/2[/pgn]
If Komodo Dragon 2.6.1 is moving instantly with Elo set to 3496, something is terribly wrong, and if it is holding a strong engine like Berserk 9 roughly even without having to think that is totally amazing. I can't offer any explanation for either of these two issues, and since we are already on Dragon 3 we would only be concerned if the issue is still present in Dragon 3. Maybe I'll test than myself. If I had to guess, I would suspect that any setting above 3400 but below 3500 just doesn't work properly, perhaps it just reverts to 3400 or even 3300 for example. A setting of 3496 is supposed to be almost the same as just using one thread at full strength, where I would expect it to beat Berserk 9 quite convincingly, but not without taking a similar amount of time.
OK, I checked out Dragon 3 a bit on this issue, and the behavior is the same as in 2.6.1. It seems that everything works properly as long as Elo is set to a value from 1 to 3400, but 3400 is the highest value that is allowed for limited strength. So if you set it to 3500 it just reverts to full strength Dragon (as we say), but if you set it to 3401 to 3499 it just disregards the last two digits. So all the tests you have run with Elo 3490 or 3496 or anything else starting with 34 have actually been with Elo 3400. That explains the "instant" (more like one second) moves. So apparently Dragon 2.6.1 playing bullet chess (like 1 second per move) on one thread with no Ponder can score nearly even with Berserk 9 playing Rapid! That is quite surprising to me, but I see on the CCRL blitz list that Dragon 2.6 on one thread is rated about 20 elo above Berserk 8.5 on 8 threads, so perhaps it is true.
Komodo rules!
Chessqueen
Posts: 5479
Joined: Wed Sep 05, 2018 2:16 am
Location: Moving
Full name: Jorge Picado

Re: Komodo Elo = 3496 vs Berserk-9-dev after 20 games so far

Post by Chessqueen »

lkaufman wrote: Mon Jun 13, 2022 5:19 am
lkaufman wrote: Mon Jun 13, 2022 4:43 am
Chessqueen wrote: Sun Jun 12, 2022 6:38 pm So far I have run Komodo Elo = 3496 vs Berserk-9-dev and after 20 games the score is dead even at T/C 15/10, I am NOT implying that Berserk at this time control is rated 3496, but since it takes the full time control to make all of its moves while Komodo Dragon 2.6.1 Elo = 3496 play instantly is leveling out the score so far. I will continue this match up to 100 games to get an idea, NOT to determine the strength of Berserk -9-dev, it could be that the real rating at T/C 15/10 lie between 3485 to 3490 after CCRL test it against different engines :roll:

[pgn][Event "Komodo Dragon Elo = 3496"]
[Site "DESKTOP-OFQ3C0P"]
[Date "2022.06.12"]
[Round "20"]
[White "Dragon-2.6.1-64bit-avx2"]
[Black "Berserk-9-dev-x64-avx2"]
[Result "1/2-1/2"]
[ECO "B90"]
[Opening "Sicilian"]
[Time "10:38:56"]
[Variation "Najdorf, 6.Be3 e5 7.Nb3 Be6 8.f3"]
[WhiteElo "3497"]
[TimeControl "900+10"]
[Termination "adjudication"]
[PlyCount "181"]


1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. f3 e5 7. Nb3 Be6 8.
Be3 h5 9. Qd2 Nbd7 10. Nd5 Nxd5 11. exd5 Bf5 12. Na5 Be7 13. Be2 Bh4+ 14.
g3 Be7 15. Nxb7 Qb8 16. Na5 Qxb2 17. Rc1 Qxa2 18. Bc4 Qa4 19. Bb3 Qb5 20.
c4 Qb8 21. Ba4 Qc7 22. Nc6 Bh3 23. Kf2 Bf6 24. Qa5 Qxa5 25. Nxa5 O-O 26.
Rb1 Rfc8 27. Bc6 Rab8 28. Ba7 Rxb1 29. Rxb1 Bd8 30. Bb7 Bf5 31. Rb2 Nc5 32.
Bxc8 Nd3+ 33. Ke2 Nxb2 34. Bxf5 Bxa5 35. c5 dxc5 36. Bb8 g6 37. Bc8 Na4 38.
d6 c4 39. Bxa6 c3 40. Kd1 Nc5 41. Bc4 Kf8 42. Bc7 Bb4 43. Bb5 Nb7 44. d7
Ke7 45. Bxe5 Nd8 46. Bd4 Ne6 47. Bb6 Bd6 48. Ba5 hBerserk 9 4 49. f4 c2+ 50. Kxc2
hxg3 51. hxg3 Bc7 52. Bb4+ Bd6 53. Bc3 Bc7 54. Bc6 Nc5 55. Bb4 Bd6 56. Bxc5
Bxc5 57. Kd3 Bd6 58. Kd4 Bc7 59. Kd5 f5 60. Bb5 Bb6 61. Ba4 Bf2 62. Kc6 Kd8
63. Bb3 Bd4 64. Bg8 Bf2 65. Bb3 Bxg3 66. Bf7 Bxf4 67. Bxg6 Be3 68. Bxf5 Bf4
69. Kd5 Bc1 70. Ke5 Bb2+ 71. Kd5 Bc1 72. Ke5 Bg5 73. Bg4 Be7 74. Ke6 Bh4
75. Bh3 Bg5 76. Bf1 Bf4 77. Bb5 Bc7 78. Kf5 Ba5 79. Kf4 Ke7 80. Ke3 Bd8 81.
Ke2 Ke6 82. Kd1 Kd6 83. Ke2 Ke7 84. Ba4 Bc7 85. Kd1 Kd8 86. Ke2 Be5 87. Kd2
Bd4 88. Kd1 Bh8 89. Bb5 Be5 90. Bc6 Bg3 91. Ke2 {User Adjudication} 1/2-1/2[/pgn]
If Komodo Dragon 2.6.1 is moving instantly with Elo set to 3496, something is terribly wrong, and if it is holding a strong engine like Berserk 9 roughly even without having to think that is totally amazing. I can't offer any explanation for either of these two issues, and since we are already on Dragon 3 we would only be concerned if the issue is still present in Dragon 3. Maybe I'll test than myself. If I had to guess, I would suspect that any setting above 3400 but below 3500 just doesn't work properly, perhaps it just reverts to 3400 or even 3300 for example. A setting of 3496 is supposed to be almost the same as just using one thread at full strength, where I would expect it to beat Berserk 9 quite convincingly, but not without taking a similar amount of time.
OK, I checked out Dragon 3 a bit on this issue, and the behavior is the same as in 2.6.1. It seems that everything works properly as long as Elo is set to a value from 1 to 3400, but 3400 is the highest value that is allowed for limited strength. So if you set it to 3500 it just reverts to full strength Dragon (as we say), but if you set it to 3401 to 3499 it just disregards the last two digits. So all the tests you have run with Elo 3490 or 3496 or anything else starting with 34 have actually been with Elo 3400. That explains the "instant" (more like one second) moves.
Are you going to fix Dragon for the next version, where if you set it from 3401 to 3499 indeed it will play better than just setting it to 3400 :?: Now I will run another test match of 50 games Dragon Elo = 3401 vs Berserk-9-dev to compare it to the first match of 50 games where I set the Elo of Dragon 2.6.1 to 3496 :roll:
Forget about memorization of Opening Theories https://www.youtube.com/watch?v=DN3381sdcdY
Chessqueen
Posts: 5479
Joined: Wed Sep 05, 2018 2:16 am
Location: Moving
Full name: Jorge Picado

Re: Komodo Elo = 3496 vs Berserk-9-dev after 20 games so far

Post by Chessqueen »

lkaufman wrote: Mon Jun 13, 2022 5:19 am
lkaufman wrote: Mon Jun 13, 2022 4:43 am
Chessqueen wrote: Sun Jun 12, 2022 6:38 pm So far I have run Komodo Elo = 3496 vs Berserk-9-dev and after 20 games the score is dead even at T/C 15/10, I am NOT implying that Berserk at this time control is rated 3496, but since it takes the full time control to make all of its moves while Komodo Dragon 2.6.1 Elo = 3496 play instantly is leveling out the score so far. I will continue this match up to 100 games to get an idea, NOT to determine the strength of Berserk -9-dev, it could be that the real rating at T/C 15/10 lie between 3485 to 3490 after CCRL test it against different engines :roll:

OK, I checked out Dragon 3 a bit on this issue, and the behavior is the same as in 2.6.1. It seems that everything works properly as long as Elo is set to a value from 1 to 3400, but 3400 is the highest value that is allowed for limited strength. So if you set it to 3500 it just reverts to full strength Dragon (as we say), but if you set it to 3401 to 3499 it just disregards the last two digits. So all the tests you have run with Elo 3490 or 3496 or anything else starting with 34 have actually been with Elo 3400. That explains the "instant" (more like one second) moves. So apparently Dragon 2.6.1 playing bullet chess (like 1 second per move) on one thread with no Ponder can score nearly even with Berserk 9 playing Rapid! That is quite surprising to me, but I see on the CCRL blitz list that Dragon 2.6 on one thread is rated about 20 elo above Berserk 8.5 on 8 threads, so perhaps it is true.
I just started another 100 games Berserk vs Komodo Dragon full strength, With Berserk contempt = 16 ( Not the default contempt =12 ). I decided to use this opening, https://fauziakram.gumroad.com/l/fzabk, but I limit it to only 8 moves, to let both engines play a little during the middlegame, because if I let this huge opening plus a 6 pieces EGTB, both engines will only make a few moves in the middlegame and in my opinion this is NOT really testing engines Chess Knowledge.

Rank Engine Score Dr Be S-B
1 Dragon-2.6.1-64bit-avx2 22.0/40 · ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· · ====1========1=1===1==================== 396.00
2 Berserk-9-x64-avx2-pext 18.0/40 ====0========0=0===0==================== · ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· · 396.00


40 of 100 games played
Tournament start: 2022.07.18, 06:32:50
Latest update: 2022.07.18, 12:54:00
Site/ Country: DESKTOP-OFQ3C0P, United States
Level: Blitz 3/2
Hardware: intel(R) Core(TM) i7-4770 CPU @ 3.40GHz with 16.0 GB Memory
Operating system: Windows 10 Home Home Edition (Build 9200) 64 bit
Table created with: Arena 3.5.1
Forget about memorization of Opening Theories https://www.youtube.com/watch?v=DN3381sdcdY
Marcus91
Posts: 28
Joined: Sun Jul 10, 2022 11:00 am
Full name: Marco Giorgio

Re: Komodo Elo = 3496 vs Berserk-9-dev after 20 games so far

Post by Marcus91 »

lkaufman wrote: Mon Jun 13, 2022 5:19 am
lkaufman wrote: Mon Jun 13, 2022 4:43 am
Chessqueen wrote: Sun Jun 12, 2022 6:38 pm So far I have run Komodo Elo = 3496 vs Berserk-9-dev and after 20 games the score is dead even at T/C 15/10, I am NOT implying that Berserk at this time control is rated 3496, but since it takes the full time control to make all of its moves while Komodo Dragon 2.6.1 Elo = 3496 play instantly is leveling out the score so far. I will continue this match up to 100 games to get an idea, NOT to determine the strength of Berserk -9-dev, it could be that the real rating at T/C 15/10 lie between 3485 to 3490 after CCRL test it against different engines :roll:

[pgn][Event "Komodo Dragon Elo = 3496"]
[Site "DESKTOP-OFQ3C0P"]
[Date "2022.06.12"]
[Round "20"]
[White "Dragon-2.6.1-64bit-avx2"]
[Black "Berserk-9-dev-x64-avx2"]
[Result "1/2-1/2"]
[ECO "B90"]
[Opening "Sicilian"]
[Time "10:38:56"]
[Variation "Najdorf, 6.Be3 e5 7.Nb3 Be6 8.f3"]
[WhiteElo "3497"]
[TimeControl "900+10"]
[Termination "adjudication"]
[PlyCount "181"]


1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. f3 e5 7. Nb3 Be6 8.
Be3 h5 9. Qd2 Nbd7 10. Nd5 Nxd5 11. exd5 Bf5 12. Na5 Be7 13. Be2 Bh4+ 14.
g3 Be7 15. Nxb7 Qb8 16. Na5 Qxb2 17. Rc1 Qxa2 18. Bc4 Qa4 19. Bb3 Qb5 20.
c4 Qb8 21. Ba4 Qc7 22. Nc6 Bh3 23. Kf2 Bf6 24. Qa5 Qxa5 25. Nxa5 O-O 26.
Rb1 Rfc8 27. Bc6 Rab8 28. Ba7 Rxb1 29. Rxb1 Bd8 30. Bb7 Bf5 31. Rb2 Nc5 32.
Bxc8 Nd3+ 33. Ke2 Nxb2 34. Bxf5 Bxa5 35. c5 dxc5 36. Bb8 g6 37. Bc8 Na4 38.
d6 c4 39. Bxa6 c3 40. Kd1 Nc5 41. Bc4 Kf8 42. Bc7 Bb4 43. Bb5 Nb7 44. d7
Ke7 45. Bxe5 Nd8 46. Bd4 Ne6 47. Bb6 Bd6 48. Ba5 hBerserk 9 4 49. f4 c2+ 50. Kxc2
hxg3 51. hxg3 Bc7 52. Bb4+ Bd6 53. Bc3 Bc7 54. Bc6 Nc5 55. Bb4 Bd6 56. Bxc5
Bxc5 57. Kd3 Bd6 58. Kd4 Bc7 59. Kd5 f5 60. Bb5 Bb6 61. Ba4 Bf2 62. Kc6 Kd8
63. Bb3 Bd4 64. Bg8 Bf2 65. Bb3 Bxg3 66. Bf7 Bxf4 67. Bxg6 Be3 68. Bxf5 Bf4
69. Kd5 Bc1 70. Ke5 Bb2+ 71. Kd5 Bc1 72. Ke5 Bg5 73. Bg4 Be7 74. Ke6 Bh4
75. Bh3 Bg5 76. Bf1 Bf4 77. Bb5 Bc7 78. Kf5 Ba5 79. Kf4 Ke7 80. Ke3 Bd8 81.
Ke2 Ke6 82. Kd1 Kd6 83. Ke2 Ke7 84. Ba4 Bc7 85. Kd1 Kd8 86. Ke2 Be5 87. Kd2
Bd4 88. Kd1 Bh8 89. Bb5 Be5 90. Bc6 Bg3 91. Ke2 {User Adjudication} 1/2-1/2[/pgn]
If Komodo Dragon 2.6.1 is moving instantly with Elo set to 3496, something is terribly wrong, and if it is holding a strong engine like Berserk 9 roughly even without having to think that is totally amazing. I can't offer any explanation for either of these two issues, and since we are already on Dragon 3 we would only be concerned if the issue is still present in Dragon 3. Maybe I'll test than myself. If I had to guess, I would suspect that any setting above 3400 but below 3500 just doesn't work properly, perhaps it just reverts to 3400 or even 3300 for example. A setting of 3496 is supposed to be almost the same as just using one thread at full strength, where I would expect it to beat Berserk 9 quite convincingly, but not without taking a similar amount of time.
OK, I checked out Dragon 3 a bit on this issue, and the behavior is the same as in 2.6.1. It seems that everything works properly as long as Elo is set to a value from 1 to 3400, but 3400 is the highest value that is allowed for limited strength. So if you set it to 3500 it just reverts to full strength Dragon (as we say), but if you set it to 3401 to 3499 it just disregards the last two digits. So all the tests you have run with Elo 3490 or 3496 or anything else starting with 34 have actually been with Elo 3400. That explains the "instant" (more like one second) moves. So apparently Dragon 2.6.1 playing bullet chess (like 1 second per move) on one thread with no Ponder can score nearly even with Berserk 9 playing Rapid! That is quite surprising to me, but I see on the CCRL blitz list that Dragon 2.6 on one thread is rated about 20 elo above Berserk 8.5 on 8 threads, so perhaps it is true.
On my laptop dragon 2.6.1 at 3400 moves in about 5 seconds (1 Thread)