Rubichess 1.9 64-bit Gauntlet for CCRL 40/15

[Graham Banks]

PGN - http://kirill-kryukov.com/chess/discuss ... p?id=47714

[Graham Banks]

Code: Select all

CCRL 40/15 Rating List - Custom engine selection
1206633 games played by 2770 programs, run by 23 testers
Ponder off, General books (up to 12 moves), 3-4-5 piece EGTB
Time control: Equivalent to 40 moves in 15 minutes on an Intel i7-4770k.
Computed on January 16, 2021 with Bayeselo based on 1'206'633 games
Tested by CCRL team, 2005-2021, http://ccrl.chessdom.com/ccrl/4040/

Rank                 Engine                   Elo   +    -   Score  AvOp  Games
1 RubiChess 1.9 64-bit                    3263  +16  -16  48.6%  +11.3  1234
  RubiChess 1.7.3 64-bit                  3190  +17  -17  50.7%   -3.0  1040
  RubiChess 1.8 64-bit                    3188  +20  -20  45.9%  +28.9   783
  RubiChess 1.7.1 64-bit                  3172  +22  -22  49.3%   +4.6   592
  RubiChess 1.6.1 64-bit                  3168  +23  -23  52.8%  -16.9   550
  RubiChess 1.6 64-bit                    3159  +20  -20  57.0%  -48.9   800
  RubiChess 1.5 64-bit                    3109  +20  -20  48.6%   +5.0   741
  RubiChess 1.4 64-bit                    3036  +17  -17  48.3%   +9.0   994
  RubiChess 1.3 64-bit                    2980  +21  -21  52.7%  -21.0   765
  RubiChess 1.2.1 64-bit                  2907  +19  -19  51.7%  -11.6   909
  RubiChess 1.1 64-bit                    2774  +20  -20  50.7%   -6.5   829
  RubiChess 1.0 64-bit                    2614  +21  -21  45.6%  +27.5   779
  RubiChess 0.9 64-bit                    2554  +22  -22  51.9%  -17.9   718
  RubiChess 0.8 64-bit                    2460  +26  -26  52.4%  -19.8   503
  RubiChess 0.7 64-bit                    2302  +23  -23  54.8%  -35.2   675

[RubiChess]

Thank you, Graham.

[Guenther]

Some food for thought for Andreas

Rubi played 56. Qd3, c7 should have drawn I guess
[d]2r5/5n1k/2PBpPp1/1P2P3/p4q2/5N1p/K3Q3/8 w - - 2 56

[pgn][Event "RubiChess Gauntlet"]
[Site "ChessGUI"]
[Date "2020.12.21"]
[Round "10.33"]
[White "RubiChess 1.9 64-bit"]
[Black "ShashChess 15.0 64-bit"]
[Result "0-1"]
[Time "4:54:11 AM"]
[ECO "B07"]
[Opening "Pirc; Ufimtsev; Antal Defense"]
[TimeControl "40/960:40/960:40/960"]
[PlyCount "126"]
[Number "411"]
[Termination "GUI adjudication"]
[BlackType "program"]
[WhiteType "program"]
[Variant "normal"]

1.e4 d6
2.d4 Nf6
3.Nc3 g6
4.Be3 Bg7
5.Qd2 Ng4
6.Bg5 h6
7.Bh4 O-O
8.O-O-O Nc6
9.h3 {1.18/30 0:23} Nf6 {-1.82/27 0:27}
10.f4 {1.22/32 0:32} Rb8 {-1.78/29 0:41}
11.Nf3 {1.17/28 0:52} d5 {-1.50/23 0:08}
12.e5 {1.19/32 0:41} Ne4 {-1.71/25 0:16}
13.Qe3 {1.37/30 0:25} Nxc3 {-2.01/31 1:19}
14.Qxc3 {1.32/30 0:25} f6 {-1.84/27 0:20}
15.g4 {1.27/30 0:27} fxe5 {-1.63/28 0:14}
16.fxe5 {1.28/27 0:24} Qe8 {-1.58/25 0:12}
17.Be2 {1.57/27 0:46} b5 {-1.64/25 0:15}
18.Rdf1 {1.56/29 0:52} b4 {-1.49/25 0:13}
19.Qe3 {1.65/30 0:42} Nd8 {-1.17/26 0:15}
20.Be1 {1.47/30 0:45} Bd7 {-1.22/32 1:05}
21.Nh4 {1.64/31 0:52} Rxf1 {-1.12/29 0:12}
22.Bxf1 {1.50/30 0:22} Bb5 {-1.09/32 0:29}
23.Bg2 {1.50/33 0:36} c5 {-0.95/28 0:44}
24.dxc5 {1.55/31 0:51} Bc6 {-0.98/24 0:11}
25.Qd3 {1.63/33 0:20} e6 {-1.12/25 0:15}
26.Kb1 {1.62/30 0:40} Kh7 {-0.93/26 0:20}
27.g5 {1.74/29 0:31} h5 {-0.43/26 0:18}
28.Qg3 {0.85/30 0:47} d4 {0.00/31 0:19}
29.Rf1 {1.66/31 0:25} Rb5 {0.00/32 0:19}
30.Rf6 {1.30/30 0:24} Bxg2 {0.00/35 0:23}
31.Qxg2 {2.16/30 0:16} Bxf6 {0.00/37 0:20}
32.gxf6 {1.08/34 0:22} Rxc5 {0.00/37 0:23}
33.Qe4 {0.95/34 0:22} a5 {0.00/37 0:23}
34.Bf2 {0.59/36 0:25} Rd5 {0.00/37 0:30}
35.Bxd4 {0.89/29 0:11} Qg8 {0.00/38 0:34}
36.b3 {0.44/33 0:24} Nf7 {0.00/42 0:42}
37.a4 {1.28/31 0:10} bxa3 {0.00/38 0:38}
38.Ka2 {0.84/29 0:08} Ng5 {0.00/36 0:37}
39.Qd3 {0.40/30 0:16} Nf7 {0.00/38 1:00}
40.Qe4 {1.50/29 0:05} Ng5 {0.00/38 0:40}
41.Qd3 {0.77/32 0:51} Nf7 {0.00/35 0:14}
42.Nf3 {0.40/34 0:18} Qc8 {0.00/37 0:43}
43.c4 {0.03/31 0:24} Rd7 {0.00/36 0:10}
44.Qe2 {0.00/35 0:24} Qc7 {0.00/38 0:19}
45.Be3 {0.82/35 0:41} Rd8 {0.00/32 0:31}
46.Bd4 {0.00/32 0:15} Qb7 {0.00/32 0:59}
47.Bc5 {0.00/34 0:16} Qc6 {0.00/30 0:10}
48.Bxa3 {0.00/33 0:28} a4 {0.00/31 0:23}
49.b4 {0.00/39 0:17} Rc8 {0.00/34 0:16}
50.Nd2 {0.00/41 0:21} Qh1 {0.00/36 0:22}
51.b5 {0.00/42 0:37} Qxh3 {0.00/34 0:23}
52.Nf3 {0.00/41 0:24} Qg4 {0.00/36 0:33}
53.c5 {0.00/39 0:25} h4 {0.00/31 0:16}
54.c6 {0.00/38 0:19} h3 {0.00/34 0:16}
55.Bd6 {0.00/39 0:24} Qf4 {0.00/36 0:16}
56.Qd3 {0.00/36 0:19} h2 {4.02/27 0:35}
57.Nxh2 {0.00/37 0:28} Qxh2 {4.44/26 0:22}
58.Ka3 {0.00/39 0:20} Qh1 {5.09/28 0:18}
59.Kxa4 {0.00/37 0:19} Qh4 {5.90/27 0:18}
60.Kb3 {-7.06/31 0:36} Ra8 {6.30/28 0:21}
61.Bb4 {-7.23/31 0:31} Nxe5 {6.38/30 0:19}
62.Qf1 {-8.91/31 0:39} Kg8 {6.53/30 0:19}
63.f7 {-9.53/28 0:28} Nxf7 {7.71/28 0:42}
0-1[/pgn]



Rubi played 59. e5, probably Rh4 would have won
[d]7R/1k6/2nK4/8/4P1b1/7p/7P/8 w - - 2 59

[pgn][Event "RubiChess Gauntlet"]
[Site "ChessGUI"]
[Date "2020.12.18"]
[Round "9.23"]
[White "RubiChess 1.9 64-bit"]
[Black "Laser 1.7 64-bit"]
[Result "1/2-1/2"]
[Time "11:17:20 AM"]
[ECO "D80"]
[Opening "Stockholm Variation/ Grunfeld"]
[TimeControl "40/960:40/960:40/960"]
[PlyCount "225"]
[Number "343"]
[Termination "50-move rule"]
[WhiteType "program"]
[BlackType "program"]
[Variant "normal"]

1.d4 Nf6
2.c4 g6
3.Nc3 d5
4.Bg5 Bg7
5.Bxf6 Bxf6
6.cxd5 c6
7.Rc1 O-O
8.dxc6 Bxd4
9.c7 {0.41/31 0:27} Bxc3 {-0.42/18 0:02}
10.Rxc3 {0.43/33 0:45} Qxd1 {-0.53/20 0:01}
11.Kxd1 {0.43/1 0:00} Nc6 {-0.31/21 0:03}
12.e4 {0.46/32 0:24} Bd7 {-0.37/25 0:22}
13.Nf3 {0.51/31 0:26} Rfc8 {-0.44/26 0:47}
14.Kd2 {0.50/33 0:28} Rxc7 {-0.52/26 1:02}
15.Bb5 {0.68/31 1:11} Re8 {-0.54/24 0:26}
16.Ke3 {0.97/28 0:32} e5 {-0.60/25 2:17}
17.Rhc1 {0.85/31 0:53} f6 {-0.74/23 0:42}
18.Ne1 {1.24/28 0:28} Rd8 {-0.75/25 0:41}
19.Nd3 {1.31/30 0:35} Be8 {-0.69/22 0:18}
20.f3 {1.26/29 0:50} Kf8 {-0.75/24 0:59}
21.Nb4 {1.35/28 0:24} Rd6 {-0.93/26 0:40}
22.b3 {1.27/30 0:48} h5 {-0.79/21 0:18}
23.g3 {1.46/30 0:36} Nxb4 {0.00/24 0:37}
24.Rxc7 {1.30/34 0:30} Bxb5 {-1.12/30 1:20}
25.Rxb7 {1.68/34 0:25} Rd3 {-1.16/29 0:42}
26.Kf2 {1.66/36 0:23} Rd2 {-1.17/26 0:15}
27.Ke1 {1.85/37 0:25} Rc2 {-0.86/27 0:13}
28.Rxc2 {1.81/36 0:25} Nxc2 {-0.95/22 0:01}
29.Kf2 {1.68/38 0:24} a6 {-1.35/30 0:32}
30.a4 {1.53/39 0:43} Bd3 {-1.35/29 0:30}
31.b4 {1.47/39 0:24} Nd4 {-1.15/30 0:20}
32.Rc7 {2.15/36 0:34} g5 {-1.28/29 0:18}
33.Ke3 {1.90/37 0:20} Bf1 {-1.42/32 0:27}
34.f4 {2.26/38 0:34} gxf4 {-1.14/30 0:15}
35.gxf4 {2.28/38 0:20} Ne6 {-1.44/32 0:16}
36.Rc8 {2.57/40 0:19} Ke7 {-1.17/33 0:22}
37.fxe5 {2.81/39 0:28} fxe5 {-1.21/31 0:14}
38.Rc1 {2.80/39 0:11} Bh3 {-1.35/31 0:12}
39.b5 {3.02/38 0:13} axb5 {-1.25/25 0:01}
40.axb5 {3.52/36 0:10} Nd8 {-1.24/34 0:32}
41.Rc5 {3.77/38 0:26} Ke6 {-1.13/35 0:23}
42.Kf2 {3.61/37 0:17} Bg4 {-1.75/35 1:07}
43.b6 {4.48/38 0:39} h4 {-1.82/30 0:16}
44.Rc8 {3.86/42 0:30} Kd7 {-1.82/34 0:29}
45.Rb8 {3.57/42 0:20} Nc6 {-2.33/35 0:21}
46.Rh8 {5.12/42 0:20} h3 {-2.42/34 0:27}
47.Rh7 {4.92/41 0:16} Kd6 {-1.75/40 0:25}
48.Rh6 {6.15/42 0:22} Be6 {-2.03/39 0:29}
49.b7 {6.56/42 0:26} Nb8 {-2.47/39 0:19}
50.Kg3 {6.85/42 0:28} Nc6 {-1.38/39 0:29}
51.Kh4 {8.16/43 0:34} Nb8 {-1.71/35 0:34}
52.Kh5 {8.71/41 0:28} Nc6 {-1.78/35 0:24}
53.Kg5 {9.11/39 0:18} Nb8 {-2.78/31 0:21}
54.Rh8 {9.08/39 0:26} Kc7 {-2.82/33 0:23}
55.Kf6 {9.65/41 0:28} Bd7 {-2.67/33 0:33}
56.Kxe5 {11.22/35 0:39} Nc6 {-2.85/32 0:21}
57.Kd5 {8.43/44 0:30} Kxb7 {-3.00/32 0:16}
58.Kd6 {11.17/45 0:25} Bg4 {-3.00/27 0:02}
59.e5 {10.59/46 0:19} Nd4 {-2.04/33 0:14}
60.e6 {9.90/48 0:23} Nxe6 {-2.11/38 0:18}
61.Rh4 {10.12/51 0:20} Bf3 {-2.13/40 0:14}
62.Rb4 {10.03/53 0:27} Kc8 {-2.13/43 0:13}
63.Kxe6 {9.86/59 0:18} Kc7 {-2.14/57 0:14}
64.Rh4 {9.86/72 0:26} Kc6 {-2.14/57 0:14}
65.Rc4 {9.86/64 0:19} Kb5 {-2.14/60 0:14}
66.Rc7 {9.86/73 0:32} Be4 {-2.14/63 0:15}
67.Kd6 {9.86/65 0:18} Bg2 {-2.14/61 0:14}
68.Rc8 {9.86/75 0:28} Bf3 {-2.15/57 0:16}
69.Ke5 {9.86/70 0:17} Bg2 {-2.15/63 0:16}
70.Kd4 {9.86/72 0:19} Bf3 {-2.15/63 0:18}
71.Rc5 {9.86/69 0:37} Kb6 {-2.15/62 0:17}
72.Re5 {9.86/75 0:18} Bg2 {-2.15/67 0:20}
73.Rc5 {9.86/69 0:33} Bf3 {-2.15/72 0:24}
74.Re5 {9.86/74 0:16} Kc6 {-2.15/65 0:22}
75.Re6 {9.86/64 0:17} Kb5 {-2.16/64 0:23}
76.Re3 {9.86/70 0:17} Bg2 {-2.16/67 0:21}
77.Rc3 {9.86/68 0:29} Kb6 {-2.16/68 0:26}
78.Rc8 {9.86/66 0:22} Bf3 {-2.13/66 0:43}
79.Rb8 {9.86/66 0:14} Kc6 {-2.00/65 0:50}
80.Rh8 {9.86/64 0:07} Bg2 {-2.00/66 0:55}
81.Rh7 {9.86/67 0:16} Bf1 {-1.72/65 0:40}
82.Ke5 {9.86/55 0:18} Bg2 {-1.94/64 0:38}
83.Re7 {9.86/58 0:16} Bd5 {-1.82/62 0:33}
84.Kd4 {9.86/54 0:18} Bg2 {-1.60/59 0:26}
85.Re6 {9.86/51 0:18} Kb5 {-1.58/59 0:30}
86.Re7 {9.86/52 0:25} Kb6 {-1.59/62 0:40}
87.Rf7 {9.86/48 0:20} Kc6 {-0.80/63 0:31}
88.Rh7 {9.90/44 0:21} Kd6 {-0.41/36 0:15}
89.Rh6 {9.86/46 0:20} Ke7 {-0.23/44 0:18}
90.Rb6 {9.99/47 0:21} Kd7 {-0.85/54 0:25}
91.Rg6 {9.94/49 0:46} Ke7 {-1.09/60 0:25}
92.Ke5 {9.71/38 0:29} Kd7 {-1.08/31 0:12}
93.Rd6 {9.71/39 0:21} Kc7 {0.00/45 0:21}
94.Ke6 {9.90/39 0:27} Be4 {0.00/56 0:13}
95.Ra6 {9.86/40 0:42} Bh1 {0.00/58 0:12}
96.Rd6 {8.72/39 0:57} Bg2 {0.00/63 0:13}
97.Ke5 {8.65/40 0:31} Bh1 {0.00/64 0:14}
98.Rd2 {7.66/39 0:38} Bg2 {0.00/61 0:13}
99.Rd4 {7.77/38 0:34} Kc6 {0.00/65 0:15}
100.Rd8 {6.50/37 0:53} Bf1 {0.00/64 0:13}
101.Rc8 {6.62/32 0:48} Kb5 {0.00/77 0:16}
102.Kd5 {6.31/31 0:24} Bg2 {0.00/66 0:15}
103.Kd4 {5.95/33 0:36} Bf1 {0.00/73 0:16}
104.Rc5 {5.27/27 0:28} Kb4 {0.00/86 0:19}
105.Rc2 {3.73/29 0:44} Bg2 {0.00/69 0:16}
106.Ra2 {0.00/26 0:38} Bf1 {0.00/73 0:17}
107.Rb2 {0.00/34 0:10} Ka3 {0.00/81 0:19}
108.Rb8 {0.00/63 0:15} Bg2 {0.00/73 0:18}
109.Rc8 {0.00/76 0:18} Bf1 {0.00/77 0:18}
110.Ke3 {0.00/65 0:16} Bg2 {0.00/78 0:19}
111.Rb8 {0.00/79 0:18} Bf1 {0.00/81 0:19}
112.Rc8 {0.00/72 0:17} Bg2 {0.00/89 0:22}
113.Rb8 {0.00/81 0:20}
1/2-1/2[/pgn]



Here is a nice overview with two filters for Rubi in this pgn file.
(WMaxEval >= 2 and no win, BMaxEval >= 2 and no win, extreme outliers are interesting)
(Note that the original pgn file was not completely sorted by game numbers and I have renumbered them
to fit graph number and sequence number in pgn)

[RubiChess]

Some food for thought for Andreas
[pgn][Event "RubiChess Gauntlet"]
[Site "ChessGUI"]
[Date "2020.12.21"]
[Round "10.33"]
[White "RubiChess 1.9 64-bit"]
[Black "ShashChess 15.0 64-bit"]
[Result "0-1"]
[Time "4:54:11 AM"]
[ECO "B07"]
[Opening "Pirc; Ufimtsev; Antal Defense"]
[TimeControl "40/960:40/960:40/960"]
[PlyCount "126"]
[Number "411"]
[Termination "GUI adjudication"]
[BlackType "program"]
[WhiteType "program"]
[Variant "normal"]

1.e4 d6
2.d4 Nf6
3.Nc3 g6
4.Be3 Bg7
5.Qd2 Ng4
6.Bg5 h6
7.Bh4 O-O
8.O-O-O Nc6
9.h3 {1.18/30 0:23} Nf6 {-1.82/27 0:27}
10.f4 {1.22/32 0:32} Rb8 {-1.78/29 0:41}
11.Nf3 {1.17/28 0:52} d5 {-1.50/23 0:08}
12.e5 {1.19/32 0:41} Ne4 {-1.71/25 0:16}
13.Qe3 {1.37/30 0:25} Nxc3 {-2.01/31 1:19}
14.Qxc3 {1.32/30 0:25} f6 {-1.84/27 0:20}
15.g4 {1.27/30 0:27} fxe5 {-1.63/28 0:14}
16.fxe5 {1.28/27 0:24} Qe8 {-1.58/25 0:12}
17.Be2 {1.57/27 0:46} b5 {-1.64/25 0:15}
18.Rdf1 {1.56/29 0:52} b4 {-1.49/25 0:13}
19.Qe3 {1.65/30 0:42} Nd8 {-1.17/26 0:15}
20.Be1 {1.47/30 0:45} Bd7 {-1.22/32 1:05}
21.Nh4 {1.64/31 0:52} Rxf1 {-1.12/29 0:12}
22.Bxf1 {1.50/30 0:22} Bb5 {-1.09/32 0:29}
23.Bg2 {1.50/33 0:36} c5 {-0.95/28 0:44}
24.dxc5 {1.55/31 0:51} Bc6 {-0.98/24 0:11}
25.Qd3 {1.63/33 0:20} e6 {-1.12/25 0:15}
26.Kb1 {1.62/30 0:40} Kh7 {-0.93/26 0:20}
27.g5 {1.74/29 0:31} h5 {-0.43/26 0:18}
28.Qg3 {0.85/30 0:47} d4 {0.00/31 0:19}
29.Rf1 {1.66/31 0:25} Rb5 {0.00/32 0:19}
30.Rf6 {1.30/30 0:24} Bxg2 {0.00/35 0:23}
31.Qxg2 {2.16/30 0:16} Bxf6 {0.00/37 0:20}
32.gxf6 {1.08/34 0:22} Rxc5 {0.00/37 0:23}
33.Qe4 {0.95/34 0:22} a5 {0.00/37 0:23}
34.Bf2 {0.59/36 0:25} Rd5 {0.00/37 0:30}
35.Bxd4 {0.89/29 0:11} Qg8 {0.00/38 0:34}
36.b3 {0.44/33 0:24} Nf7 {0.00/42 0:42}
37.a4 {1.28/31 0:10} bxa3 {0.00/38 0:38}
38.Ka2 {0.84/29 0:08} Ng5 {0.00/36 0:37}
39.Qd3 {0.40/30 0:16} Nf7 {0.00/38 1:00}
40.Qe4 {1.50/29 0:05} Ng5 {0.00/38 0:40}
41.Qd3 {0.77/32 0:51} Nf7 {0.00/35 0:14}
42.Nf3 {0.40/34 0:18} Qc8 {0.00/37 0:43}
43.c4 {0.03/31 0:24} Rd7 {0.00/36 0:10}
44.Qe2 {0.00/35 0:24} Qc7 {0.00/38 0:19}
45.Be3 {0.82/35 0:41} Rd8 {0.00/32 0:31}
46.Bd4 {0.00/32 0:15} Qb7 {0.00/32 0:59}
47.Bc5 {0.00/34 0:16} Qc6 {0.00/30 0:10}
48.Bxa3 {0.00/33 0:28} a4 {0.00/31 0:23}
49.b4 {0.00/39 0:17} Rc8 {0.00/34 0:16}
50.Nd2 {0.00/41 0:21} Qh1 {0.00/36 0:22}
51.b5 {0.00/42 0:37} Qxh3 {0.00/34 0:23}
52.Nf3 {0.00/41 0:24} Qg4 {0.00/36 0:33}
53.c5 {0.00/39 0:25} h4 {0.00/31 0:16}
54.c6 {0.00/38 0:19} h3 {0.00/34 0:16}
55.Bd6 {0.00/39 0:24} Qf4 {0.00/36 0:16}
56.Qd3 {0.00/36 0:19} h2 {4.02/27 0:35}
57.Nxh2 {0.00/37 0:28} Qxh2 {4.44/26 0:22}
58.Ka3 {0.00/39 0:20} Qh1 {5.09/28 0:18}
59.Kxa4 {0.00/37 0:19} Qh4 {5.90/27 0:18}
60.Kb3 {-7.06/31 0:36} Ra8 {6.30/28 0:21}
61.Bb4 {-7.23/31 0:31} Nxe5 {6.38/30 0:19}
62.Qf1 {-8.91/31 0:39} Kg8 {6.53/30 0:19}
63.f7 {-9.53/28 0:28} Nxf7 {7.71/28 0:42}
0-1[/pgn]

I remember watching this game live in Grahams broadcast and it was a pain. My first reaction when Rubi showed +1.2 and ShashChess showed -1.8 after book moves was "Hey, that's a biased opening but who cares when winning against Stockfish (clone)". But then Rubi played every bad move it could find until the very bad end. I already analyzed some of the positions and I remember that the final blunder is found by Rubi but takes far too long for this time control to avoid.

[pgn][Event "RubiChess Gauntlet"]
[Site "ChessGUI"]
[Date "2020.12.18"]
[Round "9.23"]
[White "RubiChess 1.9 64-bit"]
[Black "Laser 1.7 64-bit"]
[Result "1/2-1/2"]
[Time "11:17:20 AM"]
[ECO "D80"]
[Opening "Stockholm Variation/ Grunfeld"]
[TimeControl "40/960:40/960:40/960"]
[PlyCount "225"]
[Number "343"]
[Termination "50-move rule"]
[WhiteType "program"]
[BlackType "program"]
[Variant "normal"]

1.d4 Nf6
2.c4 g6
3.Nc3 d5
4.Bg5 Bg7
5.Bxf6 Bxf6
6.cxd5 c6
7.Rc1 O-O
8.dxc6 Bxd4
9.c7 {0.41/31 0:27} Bxc3 {-0.42/18 0:02}
10.Rxc3 {0.43/33 0:45} Qxd1 {-0.53/20 0:01}
11.Kxd1 {0.43/1 0:00} Nc6 {-0.31/21 0:03}
12.e4 {0.46/32 0:24} Bd7 {-0.37/25 0:22}
13.Nf3 {0.51/31 0:26} Rfc8 {-0.44/26 0:47}
14.Kd2 {0.50/33 0:28} Rxc7 {-0.52/26 1:02}
15.Bb5 {0.68/31 1:11} Re8 {-0.54/24 0:26}
16.Ke3 {0.97/28 0:32} e5 {-0.60/25 2:17}
17.Rhc1 {0.85/31 0:53} f6 {-0.74/23 0:42}
18.Ne1 {1.24/28 0:28} Rd8 {-0.75/25 0:41}
19.Nd3 {1.31/30 0:35} Be8 {-0.69/22 0:18}
20.f3 {1.26/29 0:50} Kf8 {-0.75/24 0:59}
21.Nb4 {1.35/28 0:24} Rd6 {-0.93/26 0:40}
22.b3 {1.27/30 0:48} h5 {-0.79/21 0:18}
23.g3 {1.46/30 0:36} Nxb4 {0.00/24 0:37}
24.Rxc7 {1.30/34 0:30} Bxb5 {-1.12/30 1:20}
25.Rxb7 {1.68/34 0:25} Rd3 {-1.16/29 0:42}
26.Kf2 {1.66/36 0:23} Rd2 {-1.17/26 0:15}
27.Ke1 {1.85/37 0:25} Rc2 {-0.86/27 0:13}
28.Rxc2 {1.81/36 0:25} Nxc2 {-0.95/22 0:01}
29.Kf2 {1.68/38 0:24} a6 {-1.35/30 0:32}
30.a4 {1.53/39 0:43} Bd3 {-1.35/29 0:30}
31.b4 {1.47/39 0:24} Nd4 {-1.15/30 0:20}
32.Rc7 {2.15/36 0:34} g5 {-1.28/29 0:18}
33.Ke3 {1.90/37 0:20} Bf1 {-1.42/32 0:27}
34.f4 {2.26/38 0:34} gxf4 {-1.14/30 0:15}
35.gxf4 {2.28/38 0:20} Ne6 {-1.44/32 0:16}
36.Rc8 {2.57/40 0:19} Ke7 {-1.17/33 0:22}
37.fxe5 {2.81/39 0:28} fxe5 {-1.21/31 0:14}
38.Rc1 {2.80/39 0:11} Bh3 {-1.35/31 0:12}
39.b5 {3.02/38 0:13} axb5 {-1.25/25 0:01}
40.axb5 {3.52/36 0:10} Nd8 {-1.24/34 0:32}
41.Rc5 {3.77/38 0:26} Ke6 {-1.13/35 0:23}
42.Kf2 {3.61/37 0:17} Bg4 {-1.75/35 1:07}
43.b6 {4.48/38 0:39} h4 {-1.82/30 0:16}
44.Rc8 {3.86/42 0:30} Kd7 {-1.82/34 0:29}
45.Rb8 {3.57/42 0:20} Nc6 {-2.33/35 0:21}
46.Rh8 {5.12/42 0:20} h3 {-2.42/34 0:27}
47.Rh7 {4.92/41 0:16} Kd6 {-1.75/40 0:25}
48.Rh6 {6.15/42 0:22} Be6 {-2.03/39 0:29}
49.b7 {6.56/42 0:26} Nb8 {-2.47/39 0:19}
50.Kg3 {6.85/42 0:28} Nc6 {-1.38/39 0:29}
51.Kh4 {8.16/43 0:34} Nb8 {-1.71/35 0:34}
52.Kh5 {8.71/41 0:28} Nc6 {-1.78/35 0:24}
53.Kg5 {9.11/39 0:18} Nb8 {-2.78/31 0:21}
54.Rh8 {9.08/39 0:26} Kc7 {-2.82/33 0:23}
55.Kf6 {9.65/41 0:28} Bd7 {-2.67/33 0:33}
56.Kxe5 {11.22/35 0:39} Nc6 {-2.85/32 0:21}
57.Kd5 {8.43/44 0:30} Kxb7 {-3.00/32 0:16}
58.Kd6 {11.17/45 0:25} Bg4 {-3.00/27 0:02}
59.e5 {10.59/46 0:19} Nd4 {-2.04/33 0:14}
60.e6 {9.90/48 0:23} Nxe6 {-2.11/38 0:18}
61.Rh4 {10.12/51 0:20} Bf3 {-2.13/40 0:14}
62.Rb4 {10.03/53 0:27} Kc8 {-2.13/43 0:13}
63.Kxe6 {9.86/59 0:18} Kc7 {-2.14/57 0:14}
64.Rh4 {9.86/72 0:26} Kc6 {-2.14/57 0:14}
65.Rc4 {9.86/64 0:19} Kb5 {-2.14/60 0:14}
66.Rc7 {9.86/73 0:32} Be4 {-2.14/63 0:15}
67.Kd6 {9.86/65 0:18} Bg2 {-2.14/61 0:14}
68.Rc8 {9.86/75 0:28} Bf3 {-2.15/57 0:16}
69.Ke5 {9.86/70 0:17} Bg2 {-2.15/63 0:16}
70.Kd4 {9.86/72 0:19} Bf3 {-2.15/63 0:18}
71.Rc5 {9.86/69 0:37} Kb6 {-2.15/62 0:17}
72.Re5 {9.86/75 0:18} Bg2 {-2.15/67 0:20}
73.Rc5 {9.86/69 0:33} Bf3 {-2.15/72 0:24}
74.Re5 {9.86/74 0:16} Kc6 {-2.15/65 0:22}
75.Re6 {9.86/64 0:17} Kb5 {-2.16/64 0:23}
76.Re3 {9.86/70 0:17} Bg2 {-2.16/67 0:21}
77.Rc3 {9.86/68 0:29} Kb6 {-2.16/68 0:26}
78.Rc8 {9.86/66 0:22} Bf3 {-2.13/66 0:43}
79.Rb8 {9.86/66 0:14} Kc6 {-2.00/65 0:50}
80.Rh8 {9.86/64 0:07} Bg2 {-2.00/66 0:55}
81.Rh7 {9.86/67 0:16} Bf1 {-1.72/65 0:40}
82.Ke5 {9.86/55 0:18} Bg2 {-1.94/64 0:38}
83.Re7 {9.86/58 0:16} Bd5 {-1.82/62 0:33}
84.Kd4 {9.86/54 0:18} Bg2 {-1.60/59 0:26}
85.Re6 {9.86/51 0:18} Kb5 {-1.58/59 0:30}
86.Re7 {9.86/52 0:25} Kb6 {-1.59/62 0:40}
87.Rf7 {9.86/48 0:20} Kc6 {-0.80/63 0:31}
88.Rh7 {9.90/44 0:21} Kd6 {-0.41/36 0:15}
89.Rh6 {9.86/46 0:20} Ke7 {-0.23/44 0:18}
90.Rb6 {9.99/47 0:21} Kd7 {-0.85/54 0:25}
91.Rg6 {9.94/49 0:46} Ke7 {-1.09/60 0:25}
92.Ke5 {9.71/38 0:29} Kd7 {-1.08/31 0:12}
93.Rd6 {9.71/39 0:21} Kc7 {0.00/45 0:21}
94.Ke6 {9.90/39 0:27} Be4 {0.00/56 0:13}
95.Ra6 {9.86/40 0:42} Bh1 {0.00/58 0:12}
96.Rd6 {8.72/39 0:57} Bg2 {0.00/63 0:13}
97.Ke5 {8.65/40 0:31} Bh1 {0.00/64 0:14}
98.Rd2 {7.66/39 0:38} Bg2 {0.00/61 0:13}
99.Rd4 {7.77/38 0:34} Kc6 {0.00/65 0:15}
100.Rd8 {6.50/37 0:53} Bf1 {0.00/64 0:13}
101.Rc8 {6.62/32 0:48} Kb5 {0.00/77 0:16}
102.Kd5 {6.31/31 0:24} Bg2 {0.00/66 0:15}
103.Kd4 {5.95/33 0:36} Bf1 {0.00/73 0:16}
104.Rc5 {5.27/27 0:28} Kb4 {0.00/86 0:19}
105.Rc2 {3.73/29 0:44} Bg2 {0.00/69 0:16}
106.Ra2 {0.00/26 0:38} Bf1 {0.00/73 0:17}
107.Rb2 {0.00/34 0:10} Ka3 {0.00/81 0:19}
108.Rb8 {0.00/63 0:15} Bg2 {0.00/73 0:18}
109.Rc8 {0.00/76 0:18} Bf1 {0.00/77 0:18}
110.Ke3 {0.00/65 0:16} Bg2 {0.00/78 0:19}
111.Rb8 {0.00/79 0:18} Bf1 {0.00/81 0:19}
112.Rc8 {0.00/72 0:17} Bg2 {0.00/89 0:22}
113.Rb8 {0.00/81 0:20}
1/2-1/2[/pgn]

This is probably an example of bad endgame evaluation where my NNUE is inferior to my handmade evaluation.
I saw several games with Rubi-NNUE evaluating a drawn endgame with score > 5. This one is an example.

Well, some homework... thanks for your analysis. Seems to be a nice tool that you are using for that.

Regards, Andreas