What are Stockfish and Komodo estimated rating for these 5 depths ?

Discussion of anything and everything relating to chess playing software and machines.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
lkaufman
Posts: 4286
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: What are Stockfish and Komodo estimated rating for these 5 depths ?

Post by lkaufman » Thu Jul 30, 2020 4:49 am

Chessqueen wrote:
Thu Jul 30, 2020 3:48 am
lkaufman wrote:
Sun Jul 26, 2020 1:48 am
Chessqueen wrote:
Sat Jul 25, 2020 6:12 pm
Chessqueen wrote:
Sat Jul 25, 2020 12:38 pm
Uri Blass wrote:
Sat Jul 25, 2020 9:51 am
Unfortunately almost nobody want to learn chess 1 hour per day.

The people that I taught chess wanted to learn chess only 1 hour per week but I will be surprised if one hour per day is enough to beat stockfish depth 1 after a week.
You are correct Uri, it is NOT enough. Chess is harder than what people think, but there are gifted child, but even if they are 1 week is NOT enough. Talking from my own experience I started to take lesson from an online IM instructor and after 10 months of him guiding me and telling me to get a different chess book every months and testing me on what I should have learned from those books, and providing me with guided tactical positions to solves every week as I progress, different endgames books, strategies books and testing my comprehension of those books every other week, and giving different positions and asking me to take my time and answer what would be the best move to play and explain the reason why to him, anyway after 10 long months all that I have manage to accomplish is to draw versus Komodo 13.3 at Depth 4 in time control of 10 + 5 minutes, sometime I feel like I have wasted my money, but my trainer was very happy that I drew versus Komodo at depth 4. He simply put it this way you started training with me and you were a low 1300 to 1350 at your best and you have learned so much in 10 months which is what he expected me, since I am NOT a genius or gifted person like so many. All that I have to say is that there is so much to learn in chess, that the more you learn the more there is to keep on learning. I will be happy when I could beat Komodo 14 at Depth 5 Time control 10+5 sec increment and it might be harder to do because with every version of Komodo the harder it becomes :roll:
Larry what depth do you think that Jorge Sammour should play versus Komodo 13.3 at time control of 10 + 5 Sec increment, Depth 6 or Depth 7 ?
Well, since Alex Lenderman was a close match (maybe a slight underdog) vs. Komodo Skill 21 at 3' + 2", and Skill 21 is basically depth 8 with some randomization and limited book depth, he would probably be favored slightly against depth 7 with full opening book on one thread. Jorge is about a class below Lenderman in rating, but 5x the time limit should offset this, so I would bet on Jorge vs depth 7 but I would bet on Komodo using depth 8 (with a good opening book in each case).
It would be a tough fight a depth 8 Komodo versus Jorge since Jorge has been playing versus engines for more than 20 years and he knows how to take the engine out of book form the very first 3 moves
I played two games tonite at 10' + 5" vs K14 Skill 20 (depth 7) with opening book on, one with each color, and I managed to win both games, though I needed all of my time and both were close games. With less time or one more ply for the engine it would be tough for me. I expect Jorge is at least one level (ply) stronger than I am now, so maybe he would be favored at Skill 21 (depth 8) at 10' + 5". That is probably the limit, I would guess. This is still much faster than game in one second for the engine (excluding GUI time), so at least a thousand to one time odds on one thread!
Komodo rules!

Chessqueen
Posts: 1032
Joined: Wed Sep 05, 2018 12:16 am
Full name: Nancy M Pichardo

Re: What are Stockfish and Komodo estimated rating for these 5 depths ?

Post by Chessqueen » Thu Jul 30, 2020 9:50 pm

lkaufman wrote:
Thu Jul 30, 2020 4:49 am
Chessqueen wrote:
Thu Jul 30, 2020 3:48 am
lkaufman wrote:
Sun Jul 26, 2020 1:48 am
Chessqueen wrote:
Sat Jul 25, 2020 6:12 pm
Chessqueen wrote:
Sat Jul 25, 2020 12:38 pm
Uri Blass wrote:
Sat Jul 25, 2020 9:51 am
Unfortunately almost nobody want to learn chess 1 hour per day.

The people that I taught chess wanted to learn chess only 1 hour per week but I will be surprised if one hour per day is enough to beat stockfish depth 1 after a week.
You are correct Uri, it is NOT enough. Chess is harder than what people think, but there are gifted child, but even if they are 1 week is NOT enough. Talking from my own experience I started to take lesson from an online IM instructor and after 10 months of him guiding me and telling me to get a different chess book every months and testing me on what I should have learned from those books, and providing me with guided tactical positions to solves every week as I progress, different endgames books, strategies books and testing my comprehension of those books every other week, and giving different positions and asking me to take my time and answer what would be the best move to play and explain the reason why to him, anyway after 10 long months all that I have manage to accomplish is to draw versus Komodo 13.3 at Depth 4 in time control of 10 + 5 minutes, sometime I feel like I have wasted my money, but my trainer was very happy that I drew versus Komodo at depth 4. He simply put it this way you started training with me and you were a low 1300 to 1350 at your best and you have learned so much in 10 months which is what he expected me, since I am NOT a genius or gifted person like so many. All that I have to say is that there is so much to learn in chess, that the more you learn the more there is to keep on learning. I will be happy when I could beat Komodo 14 at Depth 5 Time control 10+5 sec increment and it might be harder to do because with every version of Komodo the harder it becomes :roll:
Larry what depth do you think that Jorge Sammour should play versus Komodo 13.3 at time control of 10 + 5 Sec increment, Depth 6 or Depth 7 ?
Well, since Alex Lenderman was a close match (maybe a slight underdog) vs. Komodo Skill 21 at 3' + 2", and Skill 21 is basically depth 8 with some randomization and limited book depth, he would probably be favored slightly against depth 7 with full opening book on one thread. Jorge is about a class below Lenderman in rating, but 5x the time limit should offset this, so I would bet on Jorge vs depth 7 but I would bet on Komodo using depth 8 (with a good opening book in each case).
It would be a tough fight a depth 8 Komodo versus Jorge since Jorge has been playing versus engines for more than 20 years and he knows how to take the engine out of book form the very first 3 moves
I played two games tonite at 10' + 5" vs K14 Skill 20 (depth 7) with opening book on, one with each color, and I managed to win both games, though I needed all of my time and both were close games. With less time or one more ply for the engine it would be tough for me. I expect Jorge is at least one level (ply) stronger than I am now, so maybe he would be favored at Skill 21 (depth 8) at 10' + 5". That is probably the limit, I would guess. This is still much faster than game in one second for the engine (excluding GUI time), so at least a thousand to one time odds on one thread!
Larry can you post the last 2 games at Depth = 7 Skill 20 (depth) with opening book on, and which GUI did you used?

Jorge Called and told me that he is too busy at the moment, and sent me my $50.00 back. Probably you could challenge GM Petr Neuman again at Skill Level= 21 https://ratings.fide.com/card.phtml?event=305243 . Anyway, let Assume that Komodo 14 could possibly draw a few games versus top players like Nakamura or MVL at Depth 11 Skill Level = 24 with time control of 10 +5 Sec increment.

lkaufman
Posts: 4286
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: What are Stockfish and Komodo estimated rating for these 5 depths ?

Post by lkaufman » Fri Jul 31, 2020 2:50 am

Chessqueen wrote:
Thu Jul 30, 2020 9:50 pm
lkaufman wrote:
Thu Jul 30, 2020 4:49 am
Chessqueen wrote:
Thu Jul 30, 2020 3:48 am
lkaufman wrote:
Sun Jul 26, 2020 1:48 am
Chessqueen wrote:
Sat Jul 25, 2020 6:12 pm
Chessqueen wrote:
Sat Jul 25, 2020 12:38 pm
Uri Blass wrote:
Sat Jul 25, 2020 9:51 am
Unfortunately almost nobody want to learn chess 1 hour per day.

The people that I taught chess wanted to learn chess only 1 hour per week but I will be surprised if one hour per day is enough to beat stockfish depth 1 after a week.
You are correct Uri, it is NOT enough. Chess is harder than what people think, but there are gifted child, but even if they are 1 week is NOT enough. Talking from my own experience I started to take lesson from an online IM instructor and after 10 months of him guiding me and telling me to get a different chess book every months and testing me on what I should have learned from those books, and providing me with guided tactical positions to solves every week as I progress, different endgames books, strategies books and testing my comprehension of those books every other week, and giving different positions and asking me to take my time and answer what would be the best move to play and explain the reason why to him, anyway after 10 long months all that I have manage to accomplish is to draw versus Komodo 13.3 at Depth 4 in time control of 10 + 5 minutes, sometime I feel like I have wasted my money, but my trainer was very happy that I drew versus Komodo at depth 4. He simply put it this way you started training with me and you were a low 1300 to 1350 at your best and you have learned so much in 10 months which is what he expected me, since I am NOT a genius or gifted person like so many. All that I have to say is that there is so much to learn in chess, that the more you learn the more there is to keep on learning. I will be happy when I could beat Komodo 14 at Depth 5 Time control 10+5 sec increment and it might be harder to do because with every version of Komodo the harder it becomes :roll:
Larry what depth do you think that Jorge Sammour should play versus Komodo 13.3 at time control of 10 + 5 Sec increment, Depth 6 or Depth 7 ?
Well, since Alex Lenderman was a close match (maybe a slight underdog) vs. Komodo Skill 21 at 3' + 2", and Skill 21 is basically depth 8 with some randomization and limited book depth, he would probably be favored slightly against depth 7 with full opening book on one thread. Jorge is about a class below Lenderman in rating, but 5x the time limit should offset this, so I would bet on Jorge vs depth 7 but I would bet on Komodo using depth 8 (with a good opening book in each case).
It would be a tough fight a depth 8 Komodo versus Jorge since Jorge has been playing versus engines for more than 20 years and he knows how to take the engine out of book form the very first 3 moves
I played two games tonite at 10' + 5" vs K14 Skill 20 (depth 7) with opening book on, one with each color, and I managed to win both games, though I needed all of my time and both were close games. With less time or one more ply for the engine it would be tough for me. I expect Jorge is at least one level (ply) stronger than I am now, so maybe he would be favored at Skill 21 (depth 8) at 10' + 5". That is probably the limit, I would guess. This is still much faster than game in one second for the engine (excluding GUI time), so at least a thousand to one time odds on one thread!
Larry can you post the last 2 games at Depth = 7 Skill 20 (depth) with opening book on, and which GUI did you used?

OK, here they are. Komodo 14 (same as Fritz) GUI. Very sloppy technique in first game (115 moves), but the last portion played pretty much with just 5 seconds per move. Second game very strange bishop sac, but it almost worked, I got in huge time pressure and it should have probably been a draw.

[Event "10min+5sec vs Skill 20"]
[Site "Bethesda"]
[Date "2020.07.30"]
[Round "?"]
[White "Kaufman, Larry"]
[Black "Komodo 14 64-bit bmi2"]
[Result "1-0"]
[ECO "C54"]
[Annotator "Kaufman,Larry"]
[PlyCount "229"]
[TimeControl "600+5"]

{256MB, Hiarcs14ziBook.ctg, alienware laptop} 1. e4 {0} e5 {B 0} 2. Nf3 {3} Nc6
{B 0} 3. Bc4 {4} Nf6 {B 0} 4. d3 {3} Bc5 {B 0} 5. O-O {3} d6 {B 0} 6. c3 {7} a6
{B 0} 7. a4 {3} h6 {B 0} 8. h3 {7} a5 {B 0} 9. Re1 {6} O-O {B 1} 10. Nbd2 {7}
Be6 {B 0} 11. Bb5 {17} Ne7 {B 0} 12. d4 {17} exd4 {B 0} 13. cxd4 {2} Bb4 {
-0.37/7 0} 14. Bd3 {60} Ng6 {-0.52/7 0} 15. Qc2 {20} Nf4 {-0.74/7 0} 16. Bf1 {5
} c6 {-0.76/7 0} 17. Rd1 {5} Re8 {-0.37/7 0} 18. Nc4 {36} d5 {-0.47/7 0} 19.
Bxf4 {35} dxc4 {0.30/7 0} 20. Ne5 {31} Qc7 {0.29/7 0} 21. Bg3 {88} Bd6 {
-0.25/7 0} 22. Bxc4 {9} Bxc4 {-0.27/7 0} 23. Qxc4 {13} Nxe4 {-0.53/7 0} 24. Bh2
{13} Nf6 {-0.80/7 0} 25. Rac1 {17} Nd5 {-0.80/7 0} 26. Rd3 {58} Re6 {-1.04/7 0}
27. Rf3 {9} Rae8 {-1.46/7 0} 28. Bg3 {21} f6 {-0.76/7 0} 29. Ng6 {13} Kh7 {
-0.94/7 0} 30. Qd3 {25} f5 {-0.50/7 0} 31. Bxd6 {42} Qxd6 {-0.43/7 0} 32. Ne5 {
6} g6 {-0.29/7 0} 33. g4 {14} f4 {-0.39/7 0} 34. h4 {15} Rxe5 {-0.56/7 0} 35.
dxe5 {3} Qxe5 {-0.06/7 0} 36. h5 {7} Rg8 {0.01/7 0} 37. Rc4 {29} Qg5 {-0.14/7 0
} 38. Kh2 {58} Qh4+ {-1.32/7 0} 39. Rh3 {20} Qxf2+ {0.00/7 0} 40. Kh1 {1} Qe1+
{0.00/7 0} 41. Kh2 {4} Rg7 {0.00/7 0} 42. Re4 {20} Qf2+ {-0.22/7 0} 43. Kh1 {3}
Qxb2 {-0.22/7 0} 44. hxg6+ {20} Rxg6 {0.00/7 0} 45. Re6 {2} Qa1+ {0.80/7 0} 46.
Kh2 {10} Qb2+ {0.00/7 0} 47. Kh1 {5} Qa1+ {0.80/7 0} 48. Kg2 {12} Qg7 {0.98/7 0
} 49. g5 {12} Nb4 {1.00/7 0} 50. Qb1 {24} Qb2+ {5.01/7 0} 51. Qxb2 {2} Rxe6 {
5.76/7 0} 52. Rxh6+ {3} Rxh6 {4.34/7 0} 53. gxh6 {2} Kxh6 {4.17/7 0} 54. Kf3 {3
} Kg6 {3.82/7 0} 55. Qd2 {4} b5 {3.76/7 0} 56. Kxf4 {6} bxa4 {3.27/7 0} 57.
Qg2+ {8} Kf7 {3.25/7 0} 58. Kf5 {7} a3 {2.83/7 0} 59. Qg6+ {3} Ke7 {2.83/7 0}
60. Qg7+ {6} Kd8 {2.77/7 0} 61. Ke6 {4} Kc8 {3.65/7 0} 62. Kd6 {8} Na6 {
5.56/7 0} 63. Qd7+ {7} Kb8 {5.56/5 0} 64. Qe8+ {3} Ka7 {3.71/7 0} 65. Qxc6 {4}
a2 {3.92/7 0} 66. Qc1 {5} Nb4 {2.43/7 0} 67. Qa3 {7} Kb6 {2.79/7 0} 68. Qa4 {5}
Ka6 {3.25/7 0} 69. Kc5 {4} Nd3+ {3.97/7 0} 70. Kd6 {4} Nb4 {2.39/7 0} 71. Kc5 {
5} Nd3+ {3.97/7 0} 72. Kd4 {3} a1=R {3.97/7 0} 73. Qxa1 {6} Nb4 {5/1 1} 74. Qa4
{4} Kb6 {4/1 0} 75. Kc4 {3} Nc6 {3/1 0} 76. Qb5+ {3} Kc7 {2/1 0} 77. Kc5 {1}
Ne5 {1/0 0} 78. Qxa5+ {2} Kd7 {6/1 0} 79. Qa7+ {4} Ke6 {11/1 0} 80. Qa6+ {2}
Kf5 {11/1 0} 81. Kd5 {2} Nf3 {10/1 0} 82. Qe6+ {5} Kf4 {9/1 0} 83. Qe4+ {2} Kg3
{8/1 0} 84. Qe3 {3} Kg4 {9/1 0} 85. Ke4 {5} Ng5+ {10/1 0} 86. Ke5 {4} Nf3+ {
9/1 0} 87. Kf6 {4} Nh4 {8/1 0} 88. Qe4+ {3} Kg3 {9/1 0} 89. Qe5+ {6} Kf2 {9/0 0
} 90. Qh2+ {5} Ng2 {9/1 0} 91. Ke5 {5} Kf3 {8/1 0} 92. Qh3+ {5} Kf2 {7/1 0} 93.
Ke4 {3} Ne1 {6/1 0} 94. Qh4+ {5} Ke2 {6/1 1} 95. Qh2+ {4} Kd1 {5/1 0} 96. Ke3 {
5} Nc2+ {4/1 0} 97. Kd3 {3} Ne1+ {3/0 0} 98. Kc3 {2} Nf3 {2/0 0} 99. Qg2 {6}
Ke1 {3/1 0} 100. Kd3 {17} Ne5+ {3/1 1} 101. Ke4 {5} Nc4 {3/1 0} 102. Qc2 {10}
Nd2+ {2/0 0} 103. Kd3 {3} Kf2 {1/0 0} 104. Qxd2+ {2} Kf3 {4/0 0} 105. Qe2+ {2}
Kf4 {5/1 0} 106. Qf2+ {2} Ke5 {7/1 0} 107. Qf3 {3} Kd6 {6/0 0} 108. Qf5 {2} Kc6
{6/0 0} 109. Qe5 {2} Kb7 {6/1 0} 110. Qd6 {2} Kc8 {5/0 0} 111. Qe7 {1} Kb8 {
4/1 0} 112. Kc4 {2} Kc8 {3/0 0} 113. Kc5 {1} Kb8 {2/1 0} 114. Kb6 {3} Ka8 {
1/0 0} 115. Qb7# {5} 1-0

[Event "Skill 20 vs 10min+5sec"]
[Site "Bethesda"]
[Date "2020.07.30"]
[Round "?"]
[White "Komodo 14 64-bit bmi2"]
[Black "Kaufman, Larry"]
[Result "0-1"]
[ECO "B42"]
[PlyCount "124"]

1. e4 c5 2. Nf3 e6 3. d4 cxd4 4. Nxd4 a6 5. Nc3 Qc7 6.
Bd3 Nf6 7. O-O Be7 8. Be3 d6 9. Nf3 b5 10. Bxb5+ axb5 11. Nxb5 Qd7 12. Qd3 Ba6
13. c4 Bxb5 14. cxb5 O-O 15. a4 Qb7 16. a5 Qxe4 17. Qxe4 Nxe4 18. b6 Na6 19.
Rfb1 Nb4 20. Nd2 Nxd2 21. Bxd2 d5 22. Bxb4 Bxb4 23. a6 Bc5 24. b7 Rab8 25. b4
Bb6 26. Rb3 g6 27. h3 Rfd8 28. Ra4 d4 29. Ra2 e5 30. Rba3 Ba7 31. Kf1 e4 32. b5
d3 33. Ke1 Kg7 34. Rb3 Rd6 35. Rb4 f5 36. Rc4 Rb6 37. Ra5 Kf6 38. Rc8 Ke5 39.
Kd2 h5 40. g3 h4 41. gxh4 Kf4 42. Kd1 Kf3 43. h5 gxh5 44. Rc5 f4 45. Rc8 Kxf2
46. Ra4 Kf3 47. Rcc4 Rxb5 48. Rxe4 Be3 49. Rxf4+ Bxf4 50. Rxf4+ Kxf4 51. Kd2
Ra5 52. Kxd3 Rxa6 53. Kd4 Rxb7 54. Kd5 Rb4 55. Kc5 Re4 56. h4 Ra5+ 57. Kc6 Rf5
58. Kc7 Re6 59. Kd7 Ke5 60. Kc8 Rf7 61. Kd8 Kd6 62. Kc8 Re8# 0-1





Jorge Called and told me that he is too busy at the moment, and sent me my $50.00 back. Probably you could challenge GM Petr Neuman again at Skill Level= 21 https://ratings.fide.com/card.phtml?event=305243 . Anyway, let Assume that Komodo 14 could possibly draw a few games versus top players like Nakamura or MVL at Depth 11 Skill Level = 24 with time control of 10 +5 Sec increment.
Skill level 24 is actually depth 12 (we skipped depth 11). I did this specifically to make it likely that level 24 would be an even match in fast rapid with those top guys.
Komodo rules!

Chessqueen
Posts: 1032
Joined: Wed Sep 05, 2018 12:16 am
Full name: Nancy M Pichardo

Re: What are Stockfish and Komodo estimated rating for these 5 depths ?

Post by Chessqueen » Fri Jul 31, 2020 10:33 pm

lkaufman wrote:
Fri Jul 31, 2020 2:50 am
Chessqueen wrote:
Thu Jul 30, 2020 9:50 pm
lkaufman wrote:
Thu Jul 30, 2020 4:49 am
Chessqueen wrote:
Thu Jul 30, 2020 3:48 am
lkaufman wrote:
Sun Jul 26, 2020 1:48 am
Chessqueen wrote:
Sat Jul 25, 2020 6:12 pm
Chessqueen wrote:
Sat Jul 25, 2020 12:38 pm
Uri Blass wrote:
Sat Jul 25, 2020 9:51 am
Unfortunately almost nobody want to learn chess 1 hour per day.

The people that I taught chess wanted to learn chess only 1 hour per week but I will be surprised if one hour per day is enough to beat stockfish depth 1 after a week.
You are correct Uri, it is NOT enough. Chess is harder than what people think, but there are gifted child, but even if they are 1 week is NOT enough. Talking from my own experience I started to take lesson from an online IM instructor and after 10 months of him guiding me and telling me to get a different chess book every months and testing me on what I should have learned from those books, and providing me with guided tactical positions to solves every week as I progress, different endgames books, strategies books and testing my comprehension of those books every other week, and giving different positions and asking me to take my time and answer what would be the best move to play and explain the reason why to him, anyway after 10 long months all that I have manage to accomplish is to draw versus Komodo 13.3 at Depth 4 in time control of 10 + 5 minutes, sometime I feel like I have wasted my money, but my trainer was very happy that I drew versus Komodo at depth 4. He simply put it this way you started training with me and you were a low 1300 to 1350 at your best and you have learned so much in 10 months which is what he expected me, since I am NOT a genius or gifted person like so many. All that I have to say is that there is so much to learn in chess, that the more you learn the more there is to keep on learning. I will be happy when I could beat Komodo 14 at Depth 5 Time control 10+5 sec increment and it might be harder to do because with every version of Komodo the harder it becomes :roll:
Larry what depth do you think that Jorge Sammour should play versus Komodo 13.3 at time control of 10 + 5 Sec increment, Depth 6 or Depth 7 ?
Well, since Alex Lenderman was a close match (maybe a slight underdog) vs. Komodo Skill 21 at 3' + 2", and Skill 21 is basically depth 8 with some randomization and limited book depth, he would probably be favored slightly against depth 7 with full opening book on one thread. Jorge is about a class below Lenderman in rating, but 5x the time limit should offset this, so I would bet on Jorge vs depth 7 but I would bet on Komodo using depth 8 (with a good opening book in each case).
It would be a tough fight a depth 8 Komodo versus Jorge since Jorge has been playing versus engines for more than 20 years and he knows how to take the engine out of book form the very first 3 moves
I played two games tonite at 10' + 5" vs K14 Skill 20 (depth 7) with opening book on, one with each color, and I managed to win both games, though I needed all of my time and both were close games. With less time or one more ply for the engine it would be tough for me. I expect Jorge is at least one level (ply) stronger than I am now, so maybe he would be favored at Skill 21 (depth 8) at 10' + 5". That is probably the limit, I would guess. This is still much faster than game in one second for the engine (excluding GUI time), so at least a thousand to one time odds on one thread!
Larry can you post the last 2 games at Depth = 7 Skill 20 (depth) with opening book on, and which GUI did you used?

OK, here they are. Komodo 14 (same as Fritz) GUI. Very sloppy technique in first game (115 moves), but the last portion played pretty much with just 5 seconds per move. Second game very strange bishop sac, but it almost worked, I got in huge time pressure and it should have probably been a draw.



Skill level 24 is actually depth 12 (we skipped depth 11). I did this specifically to make it likely that level 24 would be an even match in fast rapid with those top guys.
I performed several Depth Analysis at Depth 7 , 8 and 9 and none selected the move 10. Bxb5+ all the Depth that I tested on selected 10. a4 NOT 10. Bxb5+ ? as the best move. I do NOT know what happened with your Komodo 14, but it should be investigated. You can continue the 2nd second game and see how it will turn out but, there must be a Bug either on the GUI or Komodo 14 Software :roll:


lkaufman
Posts: 4286
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: What are Stockfish and Komodo estimated rating for these 5 depths ?

Post by lkaufman » Fri Jul 31, 2020 11:43 pm

Chessqueen wrote:
Fri Jul 31, 2020 10:33 pm
lkaufman wrote:
Fri Jul 31, 2020 2:50 am
Chessqueen wrote:
Thu Jul 30, 2020 9:50 pm
lkaufman wrote:
Thu Jul 30, 2020 4:49 am
Chessqueen wrote:
Thu Jul 30, 2020 3:48 am
lkaufman wrote:
Sun Jul 26, 2020 1:48 am
Chessqueen wrote:
Sat Jul 25, 2020 6:12 pm
Chessqueen wrote:
Sat Jul 25, 2020 12:38 pm
Uri Blass wrote:
Sat Jul 25, 2020 9:51 am
Unfortunately almost nobody want to learn chess 1 hour per day.

The people that I taught chess wanted to learn chess only 1 hour per week but I will be surprised if one hour per day is enough to beat stockfish depth 1 after a week.
You are correct Uri, it is NOT enough. Chess is harder than what people think, but there are gifted child, but even if they are 1 week is NOT enough. Talking from my own experience I started to take lesson from an online IM instructor and after 10 months of him guiding me and telling me to get a different chess book every months and testing me on what I should have learned from those books, and providing me with guided tactical positions to solves every week as I progress, different endgames books, strategies books and testing my comprehension of those books every other week, and giving different positions and asking me to take my time and answer what would be the best move to play and explain the reason why to him, anyway after 10 long months all that I have manage to accomplish is to draw versus Komodo 13.3 at Depth 4 in time control of 10 + 5 minutes, sometime I feel like I have wasted my money, but my trainer was very happy that I drew versus Komodo at depth 4. He simply put it this way you started training with me and you were a low 1300 to 1350 at your best and you have learned so much in 10 months which is what he expected me, since I am NOT a genius or gifted person like so many. All that I have to say is that there is so much to learn in chess, that the more you learn the more there is to keep on learning. I will be happy when I could beat Komodo 14 at Depth 5 Time control 10+5 sec increment and it might be harder to do because with every version of Komodo the harder it becomes :roll:
Larry what depth do you think that Jorge Sammour should play versus Komodo 13.3 at time control of 10 + 5 Sec increment, Depth 6 or Depth 7 ?
Well, since Alex Lenderman was a close match (maybe a slight underdog) vs. Komodo Skill 21 at 3' + 2", and Skill 21 is basically depth 8 with some randomization and limited book depth, he would probably be favored slightly against depth 7 with full opening book on one thread. Jorge is about a class below Lenderman in rating, but 5x the time limit should offset this, so I would bet on Jorge vs depth 7 but I would bet on Komodo using depth 8 (with a good opening book in each case).
It would be a tough fight a depth 8 Komodo versus Jorge since Jorge has been playing versus engines for more than 20 years and he knows how to take the engine out of book form the very first 3 moves
I played two games tonite at 10' + 5" vs K14 Skill 20 (depth 7) with opening book on, one with each color, and I managed to win both games, though I needed all of my time and both were close games. With less time or one more ply for the engine it would be tough for me. I expect Jorge is at least one level (ply) stronger than I am now, so maybe he would be favored at Skill 21 (depth 8) at 10' + 5". That is probably the limit, I would guess. This is still much faster than game in one second for the engine (excluding GUI time), so at least a thousand to one time odds on one thread!
Larry can you post the last 2 games at Depth = 7 Skill 20 (depth) with opening book on, and which GUI did you used?

OK, here they are. Komodo 14 (same as Fritz) GUI. Very sloppy technique in first game (115 moves), but the last portion played pretty much with just 5 seconds per move. Second game very strange bishop sac, but it almost worked, I got in huge time pressure and it should have probably been a draw.



Skill level 24 is actually depth 12 (we skipped depth 11). I did this specifically to make it likely that level 24 would be an even match in fast rapid with those top guys.
I performed several Depth Analysis at Depth 7 , 8 and 9 and none selected the move 10. Bxb5+ all the Depth that I tested on selected 10. a4 NOT 10. Bxb5+ ? as the best move. I do NOT know what happened with your Komodo 14, but it should be investigated. You can continue the 2nd second game and see how it will turn out but, there must be a Bug either on the GUI or Komodo 14 Software :roll:

Yes, I think so too, I also couldn't replicate the unsound sacrifice at the skill level selected. If I can't replicate it, I can't do anything about it. Since I know that based on many thousands of games Skill 20 is not measurably weaker than depth 7, I'll assume it is GUI-specific and not a Komodo issue. There is some randomness in skill level 20 but it's not enough to trigger a blunder or unsound sac like this, it just may choose a slightly less optimal move at times. On low skill levels it could easily make this bad sac, but it was clearly playing at a good level thruout the game other than this one move.
Komodo rules!

lkaufman
Posts: 4286
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: What are Stockfish and Komodo estimated rating for these 5 depths ?

Post by lkaufman » Sat Aug 01, 2020 3:31 am

Chessqueen wrote:
Fri Jul 31, 2020 10:33 pm
lkaufman wrote:
Fri Jul 31, 2020 2:50 am
Chessqueen wrote:
Thu Jul 30, 2020 9:50 pm
lkaufman wrote:
Thu Jul 30, 2020 4:49 am
Chessqueen wrote:
Thu Jul 30, 2020 3:48 am
lkaufman wrote:
Sun Jul 26, 2020 1:48 am
Chessqueen wrote:
Sat Jul 25, 2020 6:12 pm
Chessqueen wrote:
Sat Jul 25, 2020 12:38 pm
Uri Blass wrote:
Sat Jul 25, 2020 9:51 am
Unfortunately almost nobody want to learn chess 1 hour per day.

The people that I taught chess wanted to learn chess only 1 hour per week but I will be surprised if one hour per day is enough to beat stockfish depth 1 after a week.
You are correct Uri, it is NOT enough. Chess is harder than what people think, but there are gifted child, but even if they are 1 week is NOT enough. Talking from my own experience I started to take lesson from an online IM instructor and after 10 months of him guiding me and telling me to get a different chess book every months and testing me on what I should have learned from those books, and providing me with guided tactical positions to solves every week as I progress, different endgames books, strategies books and testing my comprehension of those books every other week, and giving different positions and asking me to take my time and answer what would be the best move to play and explain the reason why to him, anyway after 10 long months all that I have manage to accomplish is to draw versus Komodo 13.3 at Depth 4 in time control of 10 + 5 minutes, sometime I feel like I have wasted my money, but my trainer was very happy that I drew versus Komodo at depth 4. He simply put it this way you started training with me and you were a low 1300 to 1350 at your best and you have learned so much in 10 months which is what he expected me, since I am NOT a genius or gifted person like so many. All that I have to say is that there is so much to learn in chess, that the more you learn the more there is to keep on learning. I will be happy when I could beat Komodo 14 at Depth 5 Time control 10+5 sec increment and it might be harder to do because with every version of Komodo the harder it becomes :roll:
Larry what depth do you think that Jorge Sammour should play versus Komodo 13.3 at time control of 10 + 5 Sec increment, Depth 6 or Depth 7 ?
Well, since Alex Lenderman was a close match (maybe a slight underdog) vs. Komodo Skill 21 at 3' + 2", and Skill 21 is basically depth 8 with some randomization and limited book depth, he would probably be favored slightly against depth 7 with full opening book on one thread. Jorge is about a class below Lenderman in rating, but 5x the time limit should offset this, so I would bet on Jorge vs depth 7 but I would bet on Komodo using depth 8 (with a good opening book in each case).
It would be a tough fight a depth 8 Komodo versus Jorge since Jorge has been playing versus engines for more than 20 years and he knows how to take the engine out of book form the very first 3 moves
I played two games tonite at 10' + 5" vs K14 Skill 20 (depth 7) with opening book on, one with each color, and I managed to win both games, though I needed all of my time and both were close games. With less time or one more ply for the engine it would be tough for me. I expect Jorge is at least one level (ply) stronger than I am now, so maybe he would be favored at Skill 21 (depth 8) at 10' + 5". That is probably the limit, I would guess. This is still much faster than game in one second for the engine (excluding GUI time), so at least a thousand to one time odds on one thread!
Larry can you post the last 2 games at Depth = 7 Skill 20 (depth) with opening book on, and which GUI did you used?

OK, here they are. Komodo 14 (same as Fritz) GUI. Very sloppy technique in first game (115 moves), but the last portion played pretty much with just 5 seconds per move. Second game very strange bishop sac, but it almost worked, I got in huge time pressure and it should have probably been a draw.



Skill level 24 is actually depth 12 (we skipped depth 11). I did this specifically to make it likely that level 24 would be an even match in fast rapid with those top guys.
I performed several Depth Analysis at Depth 7 , 8 and 9 and none selected the move 10. Bxb5+ all the Depth that I tested on selected 10. a4 NOT 10. Bxb5+ ? as the best move. I do NOT know what happened with your Komodo 14, but it should be investigated. You can continue the 2nd second game and see how it will turn out but, there must be a Bug either on the GUI or Komodo 14 Software :roll:

Since that game was suspect, I played another one as Black, same conditions, and this time won quite easily:
[Event "Skill 20 vs 10min+5sec"]
[Site "Bethesda"]
[Date "2020.07.31"]
[Round "?"]
[White "Komodo 14 64-bit bmi2"]
[Black "Kaufman, Larry"]
[Result "0-1"]
[ECO "B12"]
[Annotator "Kaufman,Larry"]
[PlyCount "126"]
[TimeControl "600+5"]

{512MB, Hiarcs14ziBook.ctg, alienware laptop} 1. e4 {0.01/1 0} c6 {6} 2. d4 {
0.01/1 0} d5 {3} 3. f3 {0.01/1 0} e6 {3} 4. Nc3 {0.01/1 0} Bb4 {5} 5. Bf4 {
0.01/1 0} Nf6 {28} 6. Qd3 {0.01/1 0} O-O {18} 7. O-O-O {0.52/7 0} b6 {21} 8. e5
{0.23/7 0} Nfd7 {20} 9. Qe3 {0.29/7 0} Ba6 {24} 10. Bxa6 {0.17/7 0} Nxa6 {3}
11. Nge2 {0.33/7 0} Be7 {11} 12. a3 {0.51/7 0} Nc7 {13} 13. Kb1 {0.28/7 0} c5 {
5} 14. g4 {0.31/7 0} c4 {45} 15. h4 {0.64/7 0} b5 {3} 16. h5 {0.62/7 0} a5 {19}
17. h6 {-0.15/7 0} g6 {4} 18. a4 {-0.15/7 0} b4 {12} 19. Nb5 {-0.33/7 0} Nxb5 {
3} 20. axb5 {-0.71/7 0} a4 {18} 21. c3 {-1.02/7 0} b3 {16} 22. g5 {-2.14/7 0}
Qa5 {22} 23. Kc1 {-2.12/7 0} a3 {3} 24. bxa3 {-2.17/7 0} Qxa3+ {83} 25. Kd2 {
-1.84/7 0} Rfb8 {17} 26. Ke1 {-2.19/7 0} Rxb5 {8} 27. Kf2 {-1.89/7 0} b2 {4}
28. Qd2 {-1.59/7 0} Qb3 {77} 29. Be3 {-1.46/7 0} Ra3 {19} 30. Rb1 {-1.51/7 0}
Nb6 {18} 31. Qd1 {-2.06/7 0} Na4 {37} 32. Qxb3 {-1.91/7 0} Rbxb3 {5} 33. Bd2 {
-1.85/7 0} Bd8 {22} 34. Ke1 {-2.56/7 0} Ba5 {15} 35. Kd1 {-2.95/7 0} Bxc3 {68}
36. Bxc3 {-2.47/7 0} Rxc3 {17} 37. Nxc3 {-2.35/7 0} Nxc3+ {2} 38. Kc2 {-2.06/7
0} Nxb1 {3} 39. Rxb1 {-1.78/7 0} c3 {10} 40. Rd1 {-4.95/7 0} Kf8 {8} 41. Rh1 {
-4.81/7 0} Ke7 {4} 42. Rb1 {-5.13/7 0} Kd7 {4} 43. f4 {-2.88/7 0} Kc6 {4} 44.
f5 {-4.40/7 0} exf5 {4} 45. Kd3 {-5.86/7 0} Kd7 {16} 46. Kc2 {-5.57/7 0} Ke6 {6
} 47. Kd3 {-7.66/7 0} f4 {4} 48. Kc2 {-6.20/7 0} Kf5 {5} 49. Re1 {-9.69/7 0}
Ra1 {24} 50. Kxc3 {-10.18/7 0} Rxe1 {6} 51. Kxb2 {-12.00/7 0} Re4 {8} 52. Kc3 {
-11.66/7 0} f3 {2} 53. Kd3 {-13.14/7 0} f2 {2} 54. Kc2 {-20.05/7 0} Rxd4 {4}
55. e6 {-17.74/7 0} fxe6 {3} 56. Kc3 {-22.61/7 0} e5 {2} 57. Kb3 {-23.37/7 0}
Rc4 {3} 58. Ka3 {-#5/7 1} f1=Q {6} 59. Kb2 {-#3/7 0} Qc1+ {4} 60. Ka2 {-#2/7 0}
Ra4+ {9} 61. Kb3 {-#3/5 0} Qc4+ {3} 62. Kb2 {-#2/5 0} Ra2+ {2} 63. Kb1 {-#1/5 0
} Qc2# {3} 0-1

It seems my estimated ratings for the levels in our Komodo readme text are too high. Maybe it's partly just a matter of getting used to playing against the computer. Based on 5' + 2" games, most fairly strong GMs struggled with what is now level 22 (9 ply), and only Hikaru beat level 23 (10 ply) I believe, and yet at 10' +5", roughly double the time limit, it looks like I should at least be playing vs. level 21 (8 ply), and I'm at least 3 classes below Hikaru in fast chess. Around 30 years ago a 6 ply search was strong enough to get a US master (2200+) rating, then equivalent to 2100 FIDE, at standard time control, and a 7 ply engine got over 2300 US (2200 FIDE). Komodo is much more selective than those engines but also has more knowledge and extensions, so it's not clear which would be stronger at 7 ply or so fixed depth. But it's obvious that Skill 20 (7 ply) would not get a 2200 FIDE rating in standard chess now, if I can beat it at 10' + 5" and I'm only around US 2300/FIDE 2200 level now. I can't really reconcile these facts. Surely standards haven't risen that much; are FIDE 2200 players really that much stronger than 30 years ago, with most people saying that FIDE ratings have INFLATED? Bottom line: we need more data on humans vs. current engines (either ones with ratings in human range or crippled levels like Komodo Skill 20 to 24).
Komodo rules!

carldaman
Posts: 1952
Joined: Sat Jun 02, 2012 12:13 am

Re: What are Stockfish and Komodo estimated rating for these 5 depths ?

Post by carldaman » Sat Aug 01, 2020 3:44 am

My theory is that FIDE Elo inflation eventually trickles up toward the top echelon, and the 2200 human of today could indeed be uninflated and much stronger than the counterpart of 20-30 years ago, when there was inflation being fed into the system through the very same 2000-2200 FIDE lower layer, with those players getting FIDE ratings based on only peak (and not average) performances.

Speaking of FIDE rating inflation, it appears to me that it has slowed down during the past decade.

Chessqueen
Posts: 1032
Joined: Wed Sep 05, 2018 12:16 am
Full name: Nancy M Pichardo

Re: What are Stockfish and Komodo estimated rating for these 5 depths ?

Post by Chessqueen » Sat Aug 01, 2020 4:59 am

Chessqueen wrote:
Thu Jul 30, 2020 3:48 am
lkaufman wrote:
Sun Jul 26, 2020 1:48 am
Chessqueen wrote:
Sat Jul 25, 2020 6:12 pm
Chessqueen wrote:
Sat Jul 25, 2020 12:38 pm
Uri Blass wrote:
Sat Jul 25, 2020 9:51 am
Unfortunately almost nobody want to learn chess 1 hour per day.

The people that I taught chess wanted to learn chess only 1 hour per week but I will be surprised if one hour per day is enough to beat stockfish depth 1 after a week.
You are correct Uri, it is NOT enough. Chess is harder than what people think, but there are gifted child, but even if they are 1 week is NOT enough. Talking from my own experience I started to take lesson from an online IM instructor and after 10 months of him guiding me and telling me to get a different chess book every months and testing me on what I should have learned from those books, and providing me with guided tactical positions to solves every week as I progress, different endgames books, strategies books and testing my comprehension of those books every other week, and giving different positions and asking me to take my time and answer what would be the best move to play and explain the reason why to him, anyway after 10 long months all that I have manage to accomplish is to draw versus Komodo 13.3 at Depth 4 in time control of 10 + 5 minutes, sometime I feel like I have wasted my money, but my trainer was very happy that I drew versus Komodo at depth 4. He simply put it this way you started training with me and you were a low 1300 to 1350 at your best and you have learned so much in 10 months which is what he expected me, since I am NOT a genius or gifted person like so many. All that I have to say is that there is so much to learn in chess, that the more you learn the more there is to keep on learning. I will be happy when I could beat Komodo 14 at Depth 5 Time control 10+5 sec increment and it might be harder to do because with every version of Komodo the harder it becomes :roll:
Larry what depth do you think that Jorge Sammour should play versus Komodo 13.3 at time control of 10 + 5 Sec increment, Depth 6 or Depth 7 ?
Well, since Alex Lenderman was a close match (maybe a slight underdog) vs. Komodo Skill 21 at 3' + 2", and Skill 21 is basically depth 8 with some randomization and limited book depth, he would probably be favored slightly against depth 7 with full opening book on one thread. Jorge is about a class below Lenderman in rating, but 5x the time limit should offset this, so I would bet on Jorge vs depth 7 but I would bet on Komodo using depth 8 (with a good opening book in each case).
It would be a tough fight a depth 8 Komodo versus Jorge since Jorge has been playing versus engines for more than 20 years and he knows how to take the engine out of book form the very first 3 moves
I played two games tonite at 10' + 5" vs K14 Skill 20 (depth 7) with opening book on, one with each color, and I managed to win both games, though I needed all of my time and both were close games. With less time or one more ply for the engine it would be tough for me. I expect Jorge is at least one level (ply) stronger than I am now, so maybe he would be favored at Skill 21 (depth 8) at 10' + 5". That is probably the limit, I would guess. This is still much faster than game in one second for the engine (excluding GUI time), so at least a thousand to one time odds on one thread!
[/quote]

Larry can you post the last 2 games at Depth = 7 Skill 20 (depth) with opening book on, and which GUI did you used?

OK, here they are. Komodo 14 (same as Fritz) GUI. Very sloppy technique in first game (115 moves), but the last portion played pretty much with just 5 seconds per move. Second game very strange bishop sac, but it almost worked, I got in huge time pressure and it should have probably been a draw.



Jorge Called and told me that he is too busy at the moment, and sent me my $50.00 back. Probably you could challenge GM Petr Neuman again at Skill Level= 21 https://ratings.fide.com/card.phtml?event=305243 . Anyway, let Assume that Komodo 14 could possibly draw a few games versus top players like Nakamura or MVL at Depth 11 Skill Level = 24 with time control of 10 +5 Sec increment.
[/quote]

Skill level 24 is actually depth 12 (we skipped depth 11). I did this specifically to make it likely that level 24 would be an even match in fast rapid with those top guys.
[/quote]

I performed several Depth Analysis at Depth 7 , 8 and 9 and none selected the move 10. Bxb5+ all the Depth that I tested on selected 10. a4 NOT 10. Bxb5+ ? as the best move. I do NOT know what happened with your Komodo 14, but it should be investigated. You can continue the 2nd second game and see how it will turn out but, there must be a Bug either on the GUI or Komodo 14 Software :roll:


Since that game was suspect, I played another one as Black, same conditions, and this time won quite easily:

Last edited by Chessqueen on Sat Aug 01, 2020 5:23 am, edited 4 times in total.

lkaufman
Posts: 4286
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: What are Stockfish and Komodo estimated rating for these 5 depths ?

Post by lkaufman » Sat Aug 01, 2020 5:04 am

carldaman wrote:
Sat Aug 01, 2020 3:44 am
My theory is that FIDE Elo inflation eventually trickles up toward the top echelon, and the 2200 human of today could indeed be uninflated and much stronger than the counterpart of 20-30 years ago, when there was inflation being fed into the system through the very same 2000-2200 FIDE lower layer, with those players getting FIDE ratings based on only peak (and not average) performances.

Speaking of FIDE rating inflation, it appears to me that it has slowed down during the past decade.
That's also my theory, and it's clear that FIDE ratings have not been inflating in the last few years even at the top. I am convinced (as you are) that 2200 FIDE players today are indeed stronger than 30 years ago, the question is how much stronger? I've dropped about 200 elo during that period, and while some of it is probably due to this deflation, I don't claim to be as strong now at age 72 as I was 30 years ago (or more). Maybe half is aging and half is deflation?
Komodo rules!

Chessqueen
Posts: 1032
Joined: Wed Sep 05, 2018 12:16 am
Full name: Nancy M Pichardo

Re: What are Stockfish and Komodo estimated rating for these 5 depths ?

Post by Chessqueen » Sat Aug 01, 2020 5:32 am

lkaufman wrote:
Sat Aug 01, 2020 5:04 am
carldaman wrote:
Sat Aug 01, 2020 3:44 am
My theory is that FIDE Elo inflation eventually trickles up toward the top echelon, and the 2200 human of today could indeed be uninflated and much stronger than the counterpart of 20-30 years ago, when there was inflation being fed into the system through the very same 2000-2200 FIDE lower layer, with those players getting FIDE ratings based on only peak (and not average) performances.

Speaking of FIDE rating inflation, it appears to me that it has slowed down during the past decade.
That's also my theory, and it's clear that FIDE ratings have not been inflating in the last few years even at the top. I am convinced (as you are) that 2200 FIDE players today are indeed stronger than 30 years ago, the question is how much stronger? I've dropped about 200 elo during that period, and while some of it is probably due to this deflation, I don't claim to be as strong now at age 72 as I was 30 years ago (or more). Maybe half is aging and half is deflation?
Larry you are a very strong player for your age 72, https://www.denkerchess.com/wp-content/ ... ndings.pdf who seems to be the strongest of this tournament this year ? ==> https://new.uschess.org/senior

Post Reply