What level is a 300 rating?

[lkaufman]

I have a student who has a chess.com Rapid rating (based on hundreds of ten minute games) of 320. I always thought of a rating like 300 as being a virtual beginner; a USCF rating of 300 meant (at least about 15 years ago) a seven or eight year old kid who gave away pieces almost every other move. But this student was able to win (playing at a similar speed to chess.com Rapid games) against Dragon up thru Elo 1000 level, only losing at 1100. She does make many blunders, but maybe only every ten moves or so, not half the time. I'm sure that she would win almost every game against novices (meaning people who have played only a dozen games or something like that). While chess.com Rapid and FIDE (standard) ratings seem to be comparable above 2000 level, perhaps they are not at all close around the 1000 (FIDE floor) level? Does anyone have any comments regarding three digit ratings and whether a chess.com three digit rapid rating is a much stronger player than a USCF (or perhaps other national ratings if they go that low) comparable rating? Is a 1000 chess.com Rapid player much stronger than a FIDE 1000? I'm having a tough time getting the low end Elo ratings for Dragon right because of this confusion.

[Uri Blass]

I have a student who has a chess.com Rapid rating (based on hundreds of ten minute games) of 320. I always thought of a rating like 300 as being a virtual beginner; a USCF rating of 300 meant (at least about 15 years ago) a seven or eight year old kid who gave away pieces almost every other move. But this student was able to win (playing at a similar speed to chess.com Rapid games) against Dragon up thru Elo 1000 level, only losing at 1100. She does make many blunders, but maybe only every ten moves or so, not half the time. I'm sure that she would win almost every game against novices (meaning people who have played only a dozen games or something like that). While chess.com Rapid and FIDE (standard) ratings seem to be comparable above 2000 level, perhaps they are not at all close around the 1000 (FIDE floor) level? Does anyone have any comments regarding three digit ratings and whether a chess.com three digit rapid rating is a much stronger player than a USCF (or perhaps other national ratings if they go that low) comparable rating? Is a 1000 chess.com Rapid player much stronger than a FIDE 1000? I'm having a tough time getting the low end Elo ratings for Dragon right because of this confusion.

I have also a student with chess.com Rapid rating 390 and I do not believe she is close to 1000 Fide rating.
I could beat her with handicap of Queen and 2 rooks and I do not believe that I have a chance to do it against players who are close to Fide rating of 1000.

She can also win almost every game against novices but 1000 fide elo is a lot more than it.

I would like to test her against dragon elo 1000 and 1100 to see what she can do against them but I do not know how to set up dragon elo rating.

When I configure Dragon2.6.1 with arena gui I can see a lot of parameters but elo is not one of them
I can see only auto skill but I do not see a skill parameter.

Edit:I guess that the parameter I need to change is uci elo that I can see with Banksia gui but for some reason not with arena gui

[Uri Blass]

This is certainly not close to the level of fide elo 1000 otherwise almost every child who play in tournaments for rating in israel and has a israeli rating could get also fide rating and the fact is that many children in israel are not strong enough to get fide rating.

I do not claim that there are not levels of children that are weaker than 1000 elo of dragon2.6.1 but they almost never go to play for israeli rating OTB and if they go to play for israeli rating they usually lose all games and quit chess later.

[pgn][Event "Computer event"]
[Site "Somewhere on Earth"]
[Date "2022.04.25"]
[Round "1"]
[White "Dragon 2.6.1 by Komodo Chess 64-bit 1000 elo"]
[Black "Dragon 2.6.1 by Komodo Chess 64-bit"]
[Result "0-1"]
[TimeControl "depth: 24"]
[Time "06:48:13"]
[Board "2"]
[Termination "mate"]
[FEN "1nb1kbn1/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQ - 0 1"]
[SetUp "1"]

1. g4 {+22.29/1 1195 20} Nc6 {+18.83/19 4091 5953706} 2. f4 {+22.12/1 10 21} Nf6 {+18.34/19 2403 3729612}
3. g5 {+21.95/1 10 25} Nd5 {+18.91/18 1222 1850554} 4. d4 {+21.91/1 10 23} Nb6 {+19.11/20 5453 8704770}
5. e4 {+22.03/1 10 26} e6 {+19.17/19 2968 4885874} 6. c4 {+22.05/1 10 36} Bb4+ {+19.28/18 2139 3176512}
7. Nc3 {+21.97/1 10 11} d6 {+19.61/19 2438 3777831} 8. a4 {+22.19/1 10 32} Bd7 {+19.10/19 2129 3133124}
9. h3 {+22.15/1 10 36} a6 {+18.52/19 3692 5462097} 10. h4 {+22.16/1 10 34} Na5 {+18.67/21 4057 6030570}
11. b3 {+18.56/1 10 91} Bxc3+ {+16.38/19 2115 2953178} 12. Kf2 {+10.46/1 10 11} Bxa1 {+12.02/18 753 940776}
13. Kg3 {+13.88/1 10 40} Nc6 {+11.71/19 2554 3212844} 14. d5 {+11.83/1 10 41} Nd8 {+12.41/18 832 1070271}
15. h5 {+13.02/1 10 41} Bc3 {+12.15/18 2482 3099591} 16. Kg4 {+12.98/1 10 54} exd5+ {+9.49/19 1537 1887729}
17. Kh4 {+10.88/1 10 5} dxc4 {+9.78/20 1555 1882785} 18. Bxc4 {+13.34/1 10 37} h6 {+9.69/21 2890 3569941}
19. Kg3 {+11.39/1 10 49} hxg5 {+9.27/19 1583 1922202} 20. Kf2 {+8.24/1 10 52} gxf4 {+8.13/19 1588 1863379}
21. Bd3 {+7.79/1 10 45} Bd4+ {+7.64/19 1971 2287208} 22. Kf1 {+6.77/1 10 6} Bf6 {+7.69/19 2527 2933771}
23. Qe1 {+8.46/1 10 31} Ne6 {+7.53/22 4764 5425871} 24. a5 {+9.38/1 10 35} Nc5 {+7.50/21 1118 1251687}
25. Ke2 {+6.41/1 10 50} Bg4+ {+6.40/19 1571 1911556} 26. Nf3 {+8.80/1 10 19} Bxf3+ {+6.87/21 5389 6562531}
27. Kf2 {+5.81/1 10 13} Nxd3+ {+0.00/25 1485 1683822} 28. Kxf3 {-0.54/1 10 15} Nxe1+ {+0.00/26 1343 1595822}
29. Ke2 {-0.54/1 10 5} Nd7 {-0.63/27 3228 3984611} 30. Kxe1 {+0.43/1 10 19} Nc5 {-2.93/23 1135 1483936}
31. Kf2 {-1.63/1 10 54} Nxb3 {-4.31/20 1151 1514492} 32. Bxf4 {-0.03/1 10 22} Nxa5 {-4.57/20 1216 1649179}
33. Kf1 {+0.93/1 10 30} Nc4 {-5.28/21 1635 2257865} 34. Kf2 {+0.35/1 10 45} b5 {-5.76/19 1613 2181147}
35. Rh3 {-0.82/1 10 41} a5 {-6.74/20 1539 2037967} 36. Rh2 {-0.18/1 10 34} b4 {-7.73/21 1816 2285135}
37. Ke2 {-0.25/1 10 24} b3 {-8.20/21 902 1272817} 38. Bc1 {-1.13/1 10 23} a4 {-8.85/22 1576 2263940}
39. Kf1 {-2.79/1 10 27} a3 {-11.48/22 1636 2786912} 40. Bxa3 {-4.84/1 10 32} Nxa3 {-11.59/22 1038 1937626}
41. Kg2 {-3.43/1 10 26} b2 {-12.78/23 985 2184073} 42. h6 {-6.75/1 10 15} b1=Q {-250.00/17 386 857256}
43. h7 {-9.64/1 10 43} Qc2+ {M+10/16 212 389898} 44. Kh1 {-9.45/1 10 39} Qd1+ {M+6/18 194 410830}
45. Kg2 {-8.49/1 10 12} Qe2+ {M+5/19 243 511830} 46. Kg1 {-7.16/1 10 16} Bd4+ {M+3/23 169 353583}
47. Rf2 {-14.79/1 10 20} Qxf2+ {M+2/27 181 489113} 48. Kh1 {M-1/1 10 6} Qg1# {M+1/99 11 4059}
0-1

[/pgn]

[Uri Blass]

I can add that 7 or 8 year old children who compete in the qualifier for the israeli championship for children under 8 that is for israeli rating are usually stronger than the level of dragon with 1000 elo
Not only most of these children are too weak to get a fide rating but also
most of the children who play in the israeli final under 8 are too weak to get a fide rating, and the israeli final under 8 included this year 26 players who played in 7 round swiss at time control of 90+30.

Only one of these players who is the israeli champion under 8 has a fide rating.

Second place is registered in fide but did not get fide rating(Maybe today he is strong enough for it)

Here is the calculation of his games in one tournament and even at that time he was better than significant part of the players who has a low israeli rating(he started with israeli rating 1200 like every child at age 9 or smaller and got progress so
he had israeli rating of 1283 in that tournament when he played against players with higher israeli rating so earn israeli rating points inspite of getting 3.5 out of 10 when you do not see most of his points in the fide report because they were against players with no fide rating)

https://ratings.fide.com/calculations.p ... 1&rating=0

[Uri Blass]

1200 elo even lost faster

[pgn][Event "Computer event"]
[Site "Somewhere on Earth"]
[Date "2022.04.25"]
[Round "1"]
[White "1200 elo Dragon 2.6.1 by Komodo Chess 64-bit dup dup"]
[Black "Dragon 2.6.1 by Komodo Chess 64-bit contempt100"]
[Result "0-1"]
[TimeControl "depth: 24"]
[Time "12:07:36"]
[Board "1"]
[Termination "mate"]
[FEN "1nb1kbn1/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQ - 0 1"]
[SetUp "1"]

1. g4 {+22.29/1 1731 20} d6 {+18.55/21 18919 30768317} 2. e4 {+22.30/1 10 23} Nf6 {+18.72/21 16512 27012437}
3. f3 {+22.13/1 10 38} a6 {+18.75/21 18634 30103897} 4. d4 {+22.15/1 10 28} Nfd7 {+18.78/22 32868 52171961}
5. c4 {+22.22/1 10 35} b6 {+18.91/21 27430 42208245} 6. b4 {+22.23/1 10 34} Bb7 {+18.61/22 16847 25146867}
7. a4 {+22.19/1 10 35} e6 {+18.53/22 26163 40622079} 8. g5 {+22.19/1 10 35} d5 {+17.79/24 10934 15614277}
9. Bf4 {+15.96/1 10 294} dxe4 {+15.99/22 11196 15083270} 10. Be5 {+20.79/1 10 77} Bxb4+ {+13.66/21 4454 6134055}
11. Kf2 {+11.78/1 10 25} Nxe5 {+13.66/22 4720 6501582} 12. f4 {+16.70/1 10 40} Ng6 {+11.26/24 18817 23492835}
13. Kg3 {+16.72/1 10 35} h6 {+11.20/23 11061 13761357} 14. Nd2 {+11.45/1 10 104} Bd6 {+9.93/23 6751 8759759}
15. Bg2 {+10.66/1 10 164} Bxf4+ {+10.03/22 7219 9106469} 16. Kg4 {+7.39/1 10 6} hxg5 {+10.14/24 24271 31330091}
17. Nxe4 {+13.75/1 10 32} Nd7 {+9.11/22 6939 8896352} 18. Kh5 {+13.44/1 10 48} Bxe4 {M+3/33 802 2140041}
19. Kg4 {-1.11/1 10 66} Bxg2 {M+2/36 834 2509246} 20. Kh5 {M-1/1 10 82} Nf6# {M+1/99 10 3663}
0-1

[/pgn]

[Uri Blass]

1300 elo also lost
[pgn][Event "Computer event"]
[Site "Somewhere on Earth"]
[Date "2022.04.25"]
[Round "1"]
[White "1300 elo Dragon 2.6.1 by Komodo Chess 64-bit dup dup dup"]
[Black "Dragon 2.6.1 by Komodo Chess 64-bit contempt100"]
[Result "0-1"]
[TimeControl "depth: 24"]
[Time "12:18:51"]
[Board "1"]
[Termination "mate"]
[FEN "1nb1kbn1/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQ - 0 1"]
[SetUp "1"]

1. g4 {+22.29/1 965 20} d6 {+18.55/21 21814 30768317} 2. e4 {+22.30/1 10 23} Nf6 {+18.72/21 17103 27012437}
3. f3 {+22.13/1 10 39} a6 {+18.75/21 18192 30103897} 4. d4 {+22.15/1 10 28} Nfd7 {+18.78/22 32640 52171961}
5. c4 {+22.22/1 10 35} b6 {+18.91/21 27403 42208245} 6. b4 {+22.23/1 10 34} Bb7 {+18.61/22 16822 25146867}
7. a4 {+22.19/1 10 35} e6 {+18.53/22 26178 40622079} 8. g5 {+22.19/1 10 35} d5 {+17.79/24 11172 15614277}
9. Qd3 {+15.58/1 10 292} Bxb4+ {+16.60/22 13765 18467492} 10. Kf2 {+15.09/1 10 10} dxe4 {+15.96/22 14066 19310858}
11. Qe3 {+18.21/1 10 43} c5 {+14.43/22 9711 13449584} 12. Ne2 {+13.92/1 10 120} exf3 {+13.95/22 10268 13177067}
13. d5 {+15.99/1 10 47} fxe2 {+15.19/24 23619 33098548} 14. Qxe2 {+15.99/1 10 41} Kd8 {+13.10/21 5733 7261626}
15. Bh3 {+12.54/1 10 113} exd5 {+13.39/21 12744 18648124} 16. Bg4 {+11.32/1 10 57} d4 {+11.89/21 9891 13404176}
17. Kg1 {+13.53/1 10 60} Nc6 {+12.63/23 39375 55372569} 18. h4 {+13.74/1 10 43} Nde5 {+11.35/20 7650 9726580}
19. Bh5 {+13.48/1 10 43} Nd7 {+11.92/22 20672 28569417} 20. Bg4 {+13.74/1 10 43} Nde5 {+11.31/20 5766 7511829}
21. Bh5 {+13.48/1 10 41} Bc8 {+12.63/24 18154 24977255} 22. Qc2 {+13.14/1 10 45} g6 {+10.72/23 12758 15817993}
23. Bd1 {+13.58/1 10 36} Be6 {+11.24/23 10504 13140314} 24. Be2 {+13.44/1 10 66} Bf5 {+9.70/21 5613 6936787}
25. Qb3 {+13.30/1 10 38} d3 {+9.27/22 7488 8937803} 26. Bf1 {+13.48/1 10 48} Nd4 {+9.22/21 9291 11775176}
27. Qa2 {+13.24/1 10 54} a5 {+9.89/21 15942 20710260} 28. Rh2 {+13.36/1 10 75} Be1 {+9.99/20 10074 12415278}
29. Rh1 {+13.24/1 10 57} Ndf3+ {+8.47/22 13194 16650415} 30. Kg2 {+7.19/1 10 6} Be4 {+9.15/21 4463 5690977}
31. Na3 {+9.88/1 10 165} Nxh4+ {M+3/28 449 1028190} 32. Kh2 {+7.75/1 10 18} Nef3+ {M+2/34 438 1166896}
33. Kh3 {M-1/2 10 13} Bf5# {M+1/99 12 2871} 0-1

[/pgn]

[Uri Blass]

game took more moves but 1400 also lost when 1500 was the first version to get a draw and 1600 also lost and I wonder if I do something wrong because I see no improvement in the playing strength.

Maybe I did a mistake when I marked limit uci
I will try without it.

[pgn][Event "Computer event"]
[Site "Somewhere on Earth"]
[Date "2022.04.25"]
[Round "1"]
[White "1400 elo Dragon 2.6.1 by Komodo Chess 64-bit dup dup dup dup"]
[Black "Dragon 2.6.1 by Komodo Chess 64-bit contempt100"]
[Result "0-1"]
[TimeControl "depth: 24"]
[Time "12:31:05"]
[Board "1"]
[Termination "mate"]
[FEN "1nb1kbn1/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQ - 0 1"]
[SetUp "1"]

1. g4 {+22.29/1 231 20} d6 {+18.55/21 19242 30768317} 2. e4 {+22.30/1 10 23} Nf6 {+18.72/21 16513 27012437}
3. f3 {+22.13/1 10 40} a6 {+18.75/21 18211 30103897} 4. d4 {+22.15/1 10 28} Nfd7 {+18.78/22 32614 52171961}
5. c4 {+22.22/1 10 35} b6 {+18.91/21 27272 42208245} 6. b4 {+22.23/1 10 34} Bb7 {+18.61/22 16803 25146867}
7. a4 {+22.19/1 10 35} e6 {+18.53/22 26208 40622079} 8. g5 {+22.19/1 10 35} d5 {+17.79/24 10874 15614277}
9. Bf4 {+14.71/1 10 299} dxe4 {+15.99/22 11143 15083270} 10. Bg3 {+20.75/1 10 79} Bxb4+ {+14.38/22 12398 16585824}
11. Nc3 {+17.28/1 10 20} Bxc3+ {+12.42/20 5623 7191625} 12. Kf2 {+9.31/1 10 11} e5 {+12.54/20 4616 5732665}
13. d5 {+15.00/1 10 68} h6 {+12.64/21 13622 17223643} 14. h4 {+12.39/1 10 235} Nc5 {+12.55/20 14584 17494852}
15. d6 {+12.30/1 10 252} Bd4+ {+9.80/22 8182 8861023} 16. Ke1 {+12.09/1 10 29} cxd6 {+9.88/21 3968 3919347}
17. fxe4 {+11.11/1 10 167} Bxe4 {+8.91/22 9964 10382354} 18. a5 {+9.45/1 10 134} Bxh1 {+7.47/21 9673 11704168}
19. Bf2 {+7.09/1 10 73} Bc3+ {+7.69/22 18201 22264515} 20. Ke2 {+2.64/1 10 10} Ke7 {+7.87/23 9061 11841586}
21. Ra3 {+6.86/1 10 152} Bb4 {+7.19/22 7018 8927683} 22. Rh3 {+6.69/1 10 169} Be4 {+7.64/24 27917 34444726}
23. axb6 {+7.32/1 10 46} Nc6 {+7.92/23 17193 21230206} 24. Qa1 {+7.50/1 10 116} a5 {+7.63/20 6661 7875930}
25. Qb2 {+7.54/1 10 103} Ne6 {+7.60/23 31309 37643848} 26. Rg3 {+7.78/1 10 118} hxg5 {+7.35/20 6939 8190114}
27. h5 {+7.73/1 10 95} Nf4+ {+7.21/20 13424 15641064} 28. Ke3 {+7.45/1 10 4} Bh7 {+6.93/20 9847 11986347}
29. Rg4 {+7.44/1 10 34} Ke8 {+6.80/20 17943 21771140} 30. Rg3 {+6.69/1 10 39} Kf8 {+6.67/20 10555 12804041}
31. Nf3 {+7.22/1 10 35} f6 {+6.83/19 6056 7170366} 32. Rxg5 {+6.85/1 10 47} Bc5+ {+7.14/21 11343 13113152}
33. Kd2 {+4.70/1 10 8} Bb4+ {+7.64/18 2381 2884608} 34. Ke3 {+6.85/1 10 6} fxg5 {+7.22/21 9421 11101870}
35. Nxg5 {+6.85/1 10 38} Bf5 {+8.06/20 3811 4508229} 36. b7 {+7.15/1 10 38} Bg4 {+8.42/21 9719 11444760}
37. c5 {+6.48/1 10 93} Bxc5+ {+9.18/19 3973 4864498} 38. Kd2 {+6.00/1 10 10} Bb4+ {+9.89/21 7095 9025954}
39. Kc1 {+6.30/1 10 18} Ke7 {+10.18/22 9730 12779672} 40. Qc2 {+6.98/1 10 69} Nb8 {+11.24/18 14735 20629405}
41. Ba7 {+7.07/1 10 73} Nd7 {+10.67/18 4706 5794090} 42. Be3 {+6.77/1 10 71} Nd5 {+9.73/19 6706 8508728}
43. Qg6 {+7.24/1 10 83} Ba3+ {+10.99/22 3365 4609751} 44. Kd2 {+8.00/1 10 5} Bb4+ {+11.03/22 7810 11516063}
45. Kc1 {+7.24/1 10 17} Nxe3 {+11.26/19 2217 3272264} 46. Bb5 {+8.48/1 10 62} Nb8 {+11.68/19 8820 14190910}
47. Qe8+ {+8.40/1 10 80} Kf6 {+7.96/5 10 816} 48. Qg6+ {+8.83/1 10 97} Ke7 {+11.68/5 10 161}
49. Qe8+ {+8.40/1 10 35} Kf6 {+7.96/5 10 636} 50. Qd8+ {+7.82/1 10 67} Kf5 {+6.99/5 10 439}
51. Kb1 {+7.67/1 10 60} Kf4 {+9.96/18 2038 2316118} 52. Ne6+ {+6.34/1 10 84} Bxe6 {+9.09/18 1666 1958313}
53. Qxb8 {+6.86/1 10 43} Bf5+ {+9.20/18 2884 3684725} 54. Ka2 {+6.11/1 10 9} Be6+ {+8.90/17 3217 4045595}
55. Kb1 {+6.86/1 10 9} Bf5+ {+9.41/19 2471 3256075} 56. Ka2 {+6.11/1 10 5} Bc5 {+9.41/18 759 999301}
57. Qf8 {+5.88/1 10 28} Ba7 {+9.53/19 1031 1882786} 58. Qxg7 {+7.66/1 10 37} Nd1 {+8.89/16 1641 1926678}
59. Ka1 {+8.68/1 10 108} Be4 {+8.60/16 1581 1927362} 60. Be2 {+9.08/1 10 70} Bd4+ {+5.09/19 757 1037164}
61. Ka2 {+5.18/1 10 3} Nc3+ {+5.99/18 669 970562} 62. Kb3 {+5.83/1 10 6} Bxb7 {+6.14/21 1010 1592711}
63. Bd3 {+5.39/1 10 105} Bf3 {+4.99/20 1609 2084957} 64. h6 {+5.49/1 10 74} Ke3 {+6.52/19 1705 2314373}
65. Kc2 {+4.96/1 10 59} Bd1+ {+0.00/24 1436 2113535} 66. Kb2 {-2.75/1 10 14} Ne2+ {-0.53/28 3221 4967391}
67. Ka3 {-1.30/1 10 6} Kxd3 {-0.51/31 1414 2267930} 68. Qb7 {-0.55/1 10 24} Nc3 {-1.55/24 1006 1619529}
69. h7 {-0.64/1 10 67} e4 {-1.69/22 709 1254703} 70. Qa6+ {-0.27/1 10 47} Kc2 {-1.77/22 758 1154191}
71. Qxa5 {-1.21/1 10 31} Bf3 {-2.73/24 1190 1872585} 72. Qa6 {+0.60/1 10 40} Bc5# {M+1/99 33 2871}
0-1

[/pgn]

[pgn][Event "Computer event"]
[Site "Somewhere on Earth"]
[Date "2022.04.25"]
[Round "1"]
[White "1500 elo Dragon 2.6.1 by Komodo Chess 64-bit "]
[Black "Dragon 2.6.1 by Komodo Chess 64-bit contempt100"]
[Result "1/2-1/2"]
[TimeControl "depth: 24"]
[Time "12:42:44"]
[Board "1"]
[Termination "repetition"]
[FEN "1nb1kbn1/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQ - 0 1"]
[SetUp "1"]

1. g4 {+22.29/1 216 20} d6 {+18.55/21 19711 30768317} 2. e4 {+22.30/1 10 23} Nf6 {+18.72/21 16766 27012437}
3. f3 {+22.13/1 10 40} a6 {+18.75/21 18483 30103897} 4. d4 {+22.15/1 10 28} Nfd7 {+18.78/22 33276 52171961}
5. c4 {+22.22/1 10 35} b6 {+18.91/21 27927 42208245} 6. b4 {+22.23/1 10 34} Bb7 {+18.61/22 17072 25146867}
7. a4 {+22.19/1 10 35} e6 {+18.53/22 26532 40622079} 8. g5 {+22.19/1 10 35} d5 {+17.79/24 11041 15614277}
9. Kd2 {+14.66/1 10 340} dxe4 {+15.15/22 7762 10494102} 10. h4 {+18.17/1 10 153} Bxb4+ {+14.55/21 13465 17199385}
11. Ke3 {+18.17/1 10 5} exf3 {+14.49/22 7564 10086668} 12. h5 {+18.17/1 10 32} c5 {+13.76/22 14793 19370092}
13. Rh4 {+14.58/1 10 188} cxd4+ {+13.16/20 3689 4523128} 14. Kxd4 {+13.61/1 10 10} Nc6+ {+11.42/21 10803 12404191}
15. Ke4 {+8.44/1 10 23} f2 {+10.64/23 6677 8365112} 16. Be3 {+11.80/1 10 401} fxg1=Q {+9.72/23 12535 15857718}
17. Bxg1 {+10.33/1 10 66} f6 {+9.99/20 3346 4159789} 18. Kf4 {+13.28/1 10 205} fxg5+ {+8.93/19 3566 4706552}
19. Kxg5 {+13.28/1 10 6} Nce5 {+9.09/21 7745 9266772} 20. Kf4 {+13.07/1 10 61} g5+ {+9.05/23 9959 12381938}
21. Kg3 {+10.04/1 10 37} gxh4+ {+7.14/20 6412 7342377} 22. Kf4 {+6.17/1 10 16} Bf8 {+5.85/22 4620 5687041}
23. Ke3 {+9.17/1 10 45} Bc5+ {+6.51/22 11647 14066492} 24. Kd2 {+5.64/1 10 33} Bxg1 {+6.63/23 9863 12430601}
25. a5 {+4.25/1 10 51} Bd4 {+5.49/21 23309 27298679} 26. Nc3 {+4.54/1 10 77} Bf3 {+5.88/21 8779 10275484}
27. Qe1 {+4.31/1 10 128} Bc5 {+6.75/21 20252 21653894} 28. Kc2 {+4.85/1 10 50} Bxh5 {+7.71/21 25802 29848707}
29. Qxh4 {+5.35/1 10 40} Bg6+ {+7.25/19 3080 3743164} 30. Kb3 {+4.65/1 10 10} Be7 {+7.09/21 14441 17089748}
31. Qe1 {+4.71/1 10 127} Nc5+ {+6.20/20 2609 3302857} 32. Ka2 {+4.75/1 10 5} Nc6 {+7.03/22 2884 3819568}
33. Kb2 {+4.31/1 10 85} bxa5 {+6.85/22 15406 19432004} 34. Rd1 {+4.37/1 10 58} a4 {+6.66/21 8629 10422898}
35. Qd2 {+4.25/1 10 69} a3+ {+6.23/22 10560 13168804} 36. Ka2 {+4.94/1 10 43} Nb4+ {+6.69/19 2551 3229539}
37. Kxa3 {+4.94/1 10 7} Kf7 {+7.33/23 13309 16646107} 38. Qf2+ {+5.89/1 10 116} Bf5 {+6.64/20 4239 5289571}
39. Qf3 {+4.61/1 10 130} Nc2+ {+8.47/20 8797 12305334} 40. Kb2 {+4.51/1 10 6} Nb4 {+8.61/23 13752 22187313}
41. Qh5+ {+5.45/1 10 79} Bg6 {+7.85/20 8597 11625215} 42. Qf3+ {+4.45/1 10 79} Kg7 {+8.02/22 6348 8974546}
43. Rd4 {+4.39/1 10 79} a5 {+7.70/21 10705 13420923} 44. Qg4 {+4.38/1 10 69} Bf6 {+6.21/20 3910 4939281}
45. Qg1 {+3.46/1 10 68} a4 {+6.73/22 8935 11381676} 46. Nxa4 {+3.53/1 10 98} Nc2 {+0.00/30 2693 3648696}
47. Ka2 {+0.90/1 10 152} Bxd4 {+0.00/35 1928 2742040} 48. Qg5 {-1.06/1 10 52} Nxa4 {+0.00/36 2129 2893569}
49. Qe7+ {-0.32/1 10 52} Kh6 {+0.00/42 2289 3199492} 50. Qh4+ {+0.00/1 10 55} Kg7 {+0.00/43 2056 2740255}
51. Qe7+ {-0.32/1 10 30} Kh6 {+0.00/44 2163 2971623} 52. Qh4+ {+0.00/1 10 30} Kg7 {+0.00/48 3927 5092720}
53. Qe7+ {+0.00/1 10 24} 1/2-1/2

[/pgn]

[pgn][Event "Computer event"]
[Site "Somewhere on Earth"]
[Date "2022.04.25"]
[Round "1"]
[White "1600 elo Dragon 2.6.1 by Komodo Chess 64-bit "]
[Black "Dragon 2.6.1 by Komodo Chess 64-bit contempt100"]
[Result "0-1"]
[TimeControl "depth: 24"]
[Time "12:51:16"]
[Board "1"]
[Termination "mate"]
[FEN "1nb1kbn1/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQ - 0 1"]
[SetUp "1"]

1. g4 {+22.29/1 828 20} d6 {+18.55/21 19444 30768317} 2. e4 {+22.30/1 10 23} Nf6 {+18.72/21 16842 27012437}
3. f3 {+22.13/1 10 40} a6 {+18.75/21 19249 30103897} 4. d4 {+22.15/1 10 28} Nfd7 {+18.78/22 33366 52171961}
5. c4 {+22.22/1 10 35} b6 {+18.91/21 27577 42208245} 6. b4 {+22.23/1 10 34} Bb7 {+18.61/22 17148 25146867}
7. a4 {+22.19/1 10 35} e6 {+18.53/22 26727 40622079} 8. g5 {+22.19/1 10 35} d5 {+17.79/24 11213 15614277}
9. Kd2 {+14.23/1 10 351} dxe4 {+15.15/22 7969 10494102} 10. h4 {+18.17/1 10 154} Bxb4+ {+14.55/21 13507 17199385}
11. Ke3 {+18.17/1 10 5} exf3 {+14.49/22 7704 10086668} 12. h5 {+18.17/1 10 32} c5 {+13.76/22 15134 19370092}
13. Nxf3 {+14.86/1 10 182} cxd4+ {+12.57/19 4186 5389959} 14. Kxd4 {+13.83/1 10 11} h6 {+12.92/21 6878 8882942}
15. g6 {+14.93/1 10 123} Nc6+ {+12.30/20 6539 8703618} 16. Ke3 {+12.35/1 10 95} Bc5+ {+12.50/21 8275 10732025}
17. Kf4 {+12.19/1 10 46} fxg6 {+12.46/21 3909 5243716} 18. Be3 {+12.40/1 10 76} g5+ {+9.08/21 7965 9631959}
19. Kg3 {+9.53/1 10 103} Bxe3 {+9.69/22 3601 4287368} 20. Kh3 {+11.91/1 10 54} e5 {+8.70/23 28384 35232597}
21. Kg3 {+11.99/1 10 40} Nf6 {+8.32/22 15828 19925824} 22. Kg2 {+11.79/1 10 45} g4 {+7.23/23 8840 10921115}
23. Kg3 {+9.33/1 10 345} Ne4+ {+5.40/26 23958 29601493} 24. Kh4 {+9.20/1 10 25} Bf2+ {M+3/30 710 1698629}
25. Kxg4 {+6.39/1 10 3} Bc8+ {M+2/35 669 1559365} 26. Qd7+ {M-1/1 10 2} Bxd7# {M+1/99 19 297}
0-1

[/pgn]

[Uri Blass]

I guess that there is a real problem with the Dragon uci elo level in handicap games because it does not play so weak in normal games based on my testing.

I am sure humans who at the level of Dragon 1600 are not going to lose with queen and 2 rooks handicap against nobody.

[Uri Blass]

I have a student who has a chess.com Rapid rating (based on hundreds of ten minute games) of 320. I always thought of a rating like 300 as being a virtual beginner; a USCF rating of 300 meant (at least about 15 years ago) a seven or eight year old kid who gave away pieces almost every other move. But this student was able to win (playing at a similar speed to chess.com Rapid games) against Dragon up thru Elo 1000 level, only losing at 1100. She does make many blunders, but maybe only every ten moves or so, not half the time. I'm sure that she would win almost every game against novices (meaning people who have played only a dozen games or something like that). While chess.com Rapid and FIDE (standard) ratings seem to be comparable above 2000 level, perhaps they are not at all close around the 1000 (FIDE floor) level? Does anyone have any comments regarding three digit ratings and whether a chess.com three digit rapid rating is a much stronger player than a USCF (or perhaps other national ratings if they go that low) comparable rating? Is a 1000 chess.com Rapid player much stronger than a FIDE 1000? I'm having a tough time getting the low end Elo ratings for Dragon right because of this confusion.

I suspect that Dragon is less stable than the typical human with the same rating(meaning it has a bigger chance to win against stronger human players and bigger chance to lose against weaker human players).

[lkaufman]

I have a student who has a chess.com Rapid rating (based on hundreds of ten minute games) of 320. I always thought of a rating like 300 as being a virtual beginner; a USCF rating of 300 meant (at least about 15 years ago) a seven or eight year old kid who gave away pieces almost every other move. But this student was able to win (playing at a similar speed to chess.com Rapid games) against Dragon up thru Elo 1000 level, only losing at 1100. She does make many blunders, but maybe only every ten moves or so, not half the time. I'm sure that she would win almost every game against novices (meaning people who have played only a dozen games or something like that). While chess.com Rapid and FIDE (standard) ratings seem to be comparable above 2000 level, perhaps they are not at all close around the 1000 (FIDE floor) level? Does anyone have any comments regarding three digit ratings and whether a chess.com three digit rapid rating is a much stronger player than a USCF (or perhaps other national ratings if they go that low) comparable rating? Is a 1000 chess.com Rapid player much stronger than a FIDE 1000? I'm having a tough time getting the low end Elo ratings for Dragon right because of this confusion.

I suspect that Dragon is less stable than the typical human with the same rating(meaning it has a bigger chance to win against stronger human players and bigger chance to lose against weaker human players).

Yes, I think this is true of the low levels. Regarding your other comments, playing engine games at queen and two rooks handicap is not very informative, especially with NN engines, as they are basically trained in such a way that anything beyond a clear rook advantage (roughly speaking) is just a win, so they may play pretty randomly when up or down by more than that until the advantage drops below about a rook or so. Regarding your student, what time limit (or if none what typical time per game) do you play when giving queen and two rooks? I can only give queen and one rook to my 320 rated student, but those are fairly slow games (like 15' + 10" roughly, though I play about blitz speed myself); probably I could also give queen and two rooks if the games were strict game/10' as in the chess.com Rapid games upon which the 320 rating is based. I have increased the strength of the Elo settings below about 1500, especially below 1000, for Dragon 3, though not by enough to offset what you are describing.

I'm starting to understand the problem Based on what both you and I observe, it seems that although it is generally assumed/believed that a zero rating reflects either random play or perhaps the play of someone who just knows the rules but has only played one game to confirm knowledge of the rules. But this seems to be very far from the truth. If we assume that the chess.com Rapid ratings are still close to what FIDE ratings would be if they extended down to 1 or 0, which you seem to agree with, these players with ratings in the 300s are perhaps a thousand elo or so above the level of someone who has just played one game, and even that person is way stronger than random play. Random play should probably get a rating like negative 1000 or even worse. But on the CCRL lists, the engines that claim to be random have positive ratings like 100 or 200 or so, despite the fact that 1500 rated CCRL engines are much stronger than 1500 FIDE human players. So it seems that something causes severe compression of engine vs engine ratings (relative to human ratings) at these very low levels, and that is showing up in the testing of the low levels on Komodo. But I'm not sure exactly what is the cause of this compression or what to do about it. I suppose it is related to your "less stable" comment, though this still doesn't address the cause (both humans and weak engines make random blunders). This seems critical to getting the low elo ratings right.