Marvin 5.00 - Ethereal 12.75, 1:0 mate after 33!! ... great game!

Frank Quisinsky · Post by **Frank Quisinsky** » Tue Mar 23, 2021 9:56 am

[pgn][Event "40 Moves in 12 min"]
[Site "fcp-tourney-2021, WASP-1"]
[Date "2021.03.21"]
[Round "26.15"]
[White "Marvin 5.0.0 NN AVX2 x64"]
[Black "Ethereal 12.75 PEXT x64"]
[ECO "A12"]
[Result "1-0"]

1. c4 {book 0s} c6 {book 0s} 2. Nf3 {book 0s} d5 {book 0s}
3. b3 {book 0s} Bg4 {book 0s} 4. e3 {book 0s} e5 {book 0s}
5. h3 {book 0s} Bxf3 {book 0s} 6. Qxf3 {book 0s} Nf6 {book
0s} 7. Bb2 {+0.72/20 21s} Bd6 {-0.13/29 31s} 8. g4
{+0.56/19 29s (Nc3)} O-O {-0.12/27 27s} 9. g5 {+0.28/21 28s
(Nc3)} Ne4 {-0.13/31 1:34m} 10. h4 {+0.36/22 20s (cxd5)} f5
{-0.43/29 50s} 11. Nc3 {+0.30/19 21s} Na6 {-0.01/28 30s}
12. cxd5 {+0.21/19 26s} cxd5 {+0.01/30 14s} 13. Nxd5
{+0.23/20 20s} Nb4 {-0.01/28 13s} 14. Bc4 {+0.34/21 20s
(Nc3)} b5 {-1.02/29 25s} 15. Ne7+ {+14.45/21 30s (Nxb4+)}
Kh8 {+17.46/30 1:02m} 16. Ng6+ {+15.96/23 1:15m} hxg6
{+17.20/29 37s} 17. h5 {+15.80/24 17s} Qxg5 {+18.65/25 18s}
18. hxg6+ {+17.35/22 32s} Qh6 {+19.26/28 43s} 19. Bf7
{+18.02/21 20s} Ng5 {+19.17/27 41s} 20. Rxh6+ {+22.32/21
16s} gxh6 {+23.03/26 52s} 21. Qxf5 {+23.90/21 22s} Nc2+
{+24.35/26 59s} 22. Ke2 {+28.66/20 17s} Rxf7 {+27.00/24
15s} 23. gxf7 {+30.05/20 24s} Nxa1 {+27.42/23 13s} 24. Qg6
{+31.04/21 16s (Qf6+)} Bc7 {+25.34/24 19s} 25. Qxh6+
{+199.83/17 1s (Qf6+)} Nh7 {+31.46/25 12s} 26. f4
{+199.85/16 0s (Qc6)} Nxb3 {+M6/59 10s} 27. axb3
{+199.87/14 0s} Bd6 {+M5/60 1s} 28. fxe5 {+199.89/1 0s} Ba3
{+M4/63 2s} 29. e6+ {+199.91/5 0s (Ba1)} Bxb2 {+M4/65 2s}
30. e7 {+199.93/1 0s} Bg7 {+M3/72 2s} 31. e8=Q+ {+199.95/1
0s} Bf8 {+M2/82 1s} 32. Qe5+ {+199.97/1 0s} Bg7 {+M1/127
0s} 33. Qexg7# {+199.99/1 0s} 1-0[/pgn]

15. Ne7+
Not checked with other engines!

Never saw that a newer Ethereal version lost a game after 33 moves!!

Best
Frank

Raphexon · Post by **Raphexon** » Tue Mar 23, 2021 10:49 am

Looks like Stockfish 13 would have also played b7-b5 until depth 19. (2 million nodes)

But even weaker engine seem to find it after a few seconds.
So not a very hard position, but one that will induce blunders at low tc.

Guenther · Post by **Guenther** » Tue Mar 23, 2021 11:14 am

Frank Quisinsky wrote: ↑Tue Mar 23, 2021 9:56 am [pgn][Event "40 Moves in 12 min"]
[Site "fcp-tourney-2021, WASP-1"]
[Date "2021.03.21"]
[Round "26.15"]
[White "Marvin 5.0.0 NN AVX2 x64"]
[Black "Ethereal 12.75 PEXT x64"]
[ECO "A12"]
[Result "1-0"]

1. c4 {book 0s} c6 {book 0s} 2. Nf3 {book 0s} d5 {book 0s}
3. b3 {book 0s} Bg4 {book 0s} 4. e3 {book 0s} e5 {book 0s}
5. h3 {book 0s} Bxf3 {book 0s} 6. Qxf3 {book 0s} Nf6 {book
0s} 7. Bb2 {+0.72/20 21s} Bd6 {-0.13/29 31s} 8. g4
{+0.56/19 29s (Nc3)} O-O {-0.12/27 27s} 9. g5 {+0.28/21 28s
(Nc3)} Ne4 {-0.13/31 1:34m} 10. h4 {+0.36/22 20s (cxd5)} f5
{-0.43/29 50s} 11. Nc3 {+0.30/19 21s} Na6 {-0.01/28 30s}
12. cxd5 {+0.21/19 26s} cxd5 {+0.01/30 14s} 13. Nxd5
{+0.23/20 20s} Nb4 {-0.01/28 13s} 14. Bc4 {+0.34/21 20s
(Nc3)} b5 {-1.02/29 25s} 15. Ne7+ {+14.45/21 30s (Nxb4+)}
Kh8 {+17.46/30 1:02m} 16. Ng6+ {+15.96/23 1:15m} hxg6
{+17.20/29 37s} 17. h5 {+15.80/24 17s} Qxg5 {+18.65/25 18s}
18. hxg6+ {+17.35/22 32s} Qh6 {+19.26/28 43s} 19. Bf7
{+18.02/21 20s} Ng5 {+19.17/27 41s} 20. Rxh6+ {+22.32/21
16s} gxh6 {+23.03/26 52s} 21. Qxf5 {+23.90/21 22s} Nc2+
{+24.35/26 59s} 22. Ke2 {+28.66/20 17s} Rxf7 {+27.00/24
15s} 23. gxf7 {+30.05/20 24s} Nxa1 {+27.42/23 13s} 24. Qg6
{+31.04/21 16s (Qf6+)} Bc7 {+25.34/24 19s} 25. Qxh6+
{+199.83/17 1s (Qf6+)} Nh7 {+31.46/25 12s} 26. f4
{+199.85/16 0s (Qc6)} Nxb3 {+M6/59 10s} 27. axb3
{+199.87/14 0s} Bd6 {+M5/60 1s} 28. fxe5 {+199.89/1 0s} Ba3
{+M4/63 2s} 29. e6+ {+199.91/5 0s (Ba1)} Bxb2 {+M4/65 2s}
30. e7 {+199.93/1 0s} Bg7 {+M3/72 2s} 31. e8=Q+ {+199.95/1
0s} Bf8 {+M2/82 1s} 32. Qe5+ {+199.97/1 0s} Bg7 {+M1/127
0s} 33. Qexg7# {+199.99/1 0s} 1-0[/pgn]

15. Ne7+
Not checked with other engines!

Never saw that a newer Ethereal version lost a game after 33 moves!!

Best
Frank

It needs depth 26 and 34s on my slow machine + slow compilation, from FEN (1 thread, 512MB)
I will try with feeding pgn moves too simulate some hash state, probably this will make it realize slower.

[d]r2q1rk1/pp4pp/3b4/3NppP1/1nB1n2P/1P2PQ2/PB1P1P2/R3K2R b KQ - 2 14

Code: Select all

Version: WinBoard 4.9.1 + Ethereal_1275-64
	mover viewpoint		fewer / Multi-PV setting = 1 / more
exclude: none best +tail                                          
dep	score	nodes	time	(not shown:  tbhits	knps	seldep)
 31	+0.01 	66.3M  	1:26.00	g8h8 d5b4 d6b4 e1g1 b7b5 c4e2 e4d2 f3b7 a8b8 b7a7 b8a8 a7b7 a8b8 b7a7 
 30	 -0.01 	57.0M  	1:13.83	g8h8 d5b4 d6b4 e1g1 b7b5 c4e2 e4d2 f3b7 a8b8 b7a7 b8a8 a7b7 a8b8 b7a7 
 29	 -0.01 	47.0M  	1:00.34	g8h8 d5b4 d6b4 e1g1 b7b5 c4e2 e4d2 f3b7 a8b8 b7a7 b8a8 a7b7 a8b8 b7a7 
 28	+0.01 	39.1M  	0:50.04	g8h8 d5b4 d6b4 e1g1 b7b5 c4e2 e4d2 f3b7 a8b8 b7a7 b8a8 a7b7 a8b8 b7a7 
 27	 -0.01 	31.2M  	0:39.95	g8h8 d5b4 d6b4 e1g1 b7b5 c4e2 e4d2 f3b7 a8b8 b7a7 b8a8 a7b7 a8b8 b7a7 
 26	 -0.01 	27.0M  	0:34.66	g8h8 d5b4 d6b4 e1g1 b7b5 c4e2 e4d2 f3b7 a8b8 b7a7 b8a8 a7b7 a8b8 b7a7 
 25	+0.75 	19.1M  	0:24.43	b7b5 
 25	+0.80 	16.4M  	0:20.79	b7b5 
 25	+0.01 	23.4M  	0:29.92	g8h8 d5b4 d6b4 e1g1 b7b5 c4e2 e4d2 f3b7 a8b8 b7a7 b8a8 a7b7 a8b8 b7a7 
 24	+0.70 	15.8M  	0:20.06	b7b5 d5b4 b5c4 b4d5 c4b3 f3d1 d8d7 d5c3 b3a2 d1b3 f8f7 b3d5 a8b8 a1a2 e4c3 b2c3 b8b1 e1e2 b1h1 d5a8 d6f8 a8h1 d7b5 d2d3 e5e4 h1d1 f5f4 a2b2 b5c6 c3d4 f4e3 d4e3 e4d3 d1d3 c6h1 d3d4 h7h6 e2d2 h6g5 h4g5 h1h5 d4d5 
 24	+0.98 	15.4M  	0:19.56	b7b5 
 24	+1.10 	14.7M  	0:18.64	b7b5 
 24	+0.95 	14.2M  	0:18.08	b7b5 
 24	+0.85 	13.9M  	0:17.64	b7b5 
 23	+0.75 	13.3M  	0:16.83	b7b5 d5b4 b5c4 b4d5 c4b3 f3d1 d8d7 d2d3 e4c5 d3d4 e5d4 d1d4 f5f4 d5f4 a8d8 b2a3 b3b2 a1b1 d7b7 e1g1 d6f4 d4c5 b7f3 c5c4 f8f7 e3f4 f3a3 f4f5 a3h3 b1b2 h3f5 c4c3 
 23	+0.59 	12.8M  	0:16.13	b7b5 
 22	+1.24 	8.35M  	0:10.45	b7b5 d5b4 b5c4 b4d5 c4b3 f3d1 d8d7 d5c3 b3a2 d1b3 g8h8 a1a2 e4c3 b2c3 f5f4 a2a5 f4e3 d2e3 a8b8 b3c2 d7b7 e1g1 b7f3 c3e5 
 22	+1.23 	7.68M  	0:09.61	b7b5 
 22	+0.90 	7.28M  	0:09.09	b7b5 
 21	+1.13 	5.73M  	0:07.14	b7b5 d5b4 b5c4 b4d5 c4b3 f3e2 d8d7 e1g1 a8b8 d2d4 e5d4 a2b3 d7f7 a1a5 d4e3 d5e3 d6c5 a5b5 c5e3 b5b8 f8b8 e2e3 b8b3 
 21	+0.97 	5.35M  	0:06.67	b7b5 
 21	+0.75 	5.19M  	0:06.49	b7b5 
 21	+0.60 	5.05M  	0:06.31	b7b5 
 20	+0.60 	3.97M  	0:04.97	b7b5 d5b4 b5c4 b4d5 c4b3 f3e2 b3a2 a1a2 d8b8 d5c3 b8b3 e2d1 b3d1 c3d1 a7a5 d2d3 e4c5 e1e2 f5f4 b2a3 f4e3 f2e3 a8c8 d1c3 c5b7 a3d6 b7d6 
 20	+0.84 	3.77M  	0:04.74	b7b5 
 20	+0.85 	3.50M  	0:04.41	b7b5 
 19	+1.08 	1.93M  	0:02.40	b7b5 d5b4 b5c4 b4d5 c4b3 f3e2 b3a2 e2c4 f8f7 c4a2 a8b8 b2c3 d8d7 a1b1 b8b1 a2b1 e4c3 d2c3 a7a5 b1d3 a5a4 e1g1 
 18	+1.02 	751372	0:00.89	b7b5 d5b4 b5c4 b4d5 c4b3 d5c3 d6a3 f3d1 a3b2 d1b3 g8h8 b3b2 d8d3 c3e4 d3e4 h1g1 a8d8 a1d1 d8d3 b2b5 f8d8 b5a6 d3d7 
 17	+0.74 	677843	0:00.81	b7b5 d5b4 b5c4 b4d5 c4b3 d5c3 d6a3 f3d1 a3b2 d1b3 g8h8 b3b2 d8d3 c3e4 d3e4 h1g1 f8d8 b2b1 e4f3 b1b5 f3e4 
 16	+0.41 	640139	0:00.78	b7b5 d5b4 b5c4 b4d5 c4b3 d5c3 d6a3 f3d1 a3b2 d1b3 g8h8 b3b2 d8d3 c3e4 d3e4 h1g1 f8b8 b2c3 b8c8 c3a3 c8d8 a3c5 d8d3 
...
  0	set Threads to 1

Guenther · Post by **Guenther** » Tue Mar 23, 2021 11:28 am

Guenther wrote: ↑Tue Mar 23, 2021 11:14 am
It needs depth 26 and 34s on my slow machine + slow compilation, from FEN (1 thread, 512MB)
I will try with feeding pgn moves too simulate some hash state, probably this will make it realize slower.

[d]r2q1rk1/pp4pp/3b4/3NppP1/1nB1n2P/1P2PQ2/PB1P1P2/R3K2R b KQ - 2 14

Code: Select all

Version: WinBoard 4.9.1 + Ethereal_1275-64
	mover viewpoint		fewer / Multi-PV setting = 1 / more
exclude: none best +tail                                          
dep	score	nodes	time	(not shown:  tbhits	knps	seldep)
 31	+0.01 	66.3M  	1:26.00	g8h8 d5b4 d6b4 e1g1 b7b5 c4e2 e4d2 f3b7 a8b8 b7a7 b8a8 a7b7 a8b8 b7a7 
 30	 -0.01 	57.0M  	1:13.83	g8h8 d5b4 d6b4 e1g1 b7b5 c4e2 e4d2 f3b7 a8b8 b7a7 b8a8 a7b7 a8b8 b7a7 
 29	 -0.01 	47.0M  	1:00.34	g8h8 d5b4 d6b4 e1g1 b7b5 c4e2 e4d2 f3b7 a8b8 b7a7 b8a8 a7b7 a8b8 b7a7 
 28	+0.01 	39.1M  	0:50.04	g8h8 d5b4 d6b4 e1g1 b7b5 c4e2 e4d2 f3b7 a8b8 b7a7 b8a8 a7b7 a8b8 b7a7 
 27	 -0.01 	31.2M  	0:39.95	g8h8 d5b4 d6b4 e1g1 b7b5 c4e2 e4d2 f3b7 a8b8 b7a7 b8a8 a7b7 a8b8 b7a7 
 26	 -0.01 	27.0M  	0:34.66	g8h8 d5b4 d6b4 e1g1 b7b5 c4e2 e4d2 f3b7 a8b8 b7a7 b8a8 a7b7 a8b8 b7a7 
 25	+0.75 	19.1M  	0:24.43	b7b5 
 25	+0.80 	16.4M  	0:20.79	b7b5 
 ...

Well, with feeding the pgn and letting each move analyse at least to depth 20 until the position, it even finds b5?? refuted much faster.
But this can happen. It's ofc impossible to have exact hash entries, still surprised Etherals King remained in the Bishop diagonal, though.

Code: Select all

	mover viewpoint		fewer / Multi-PV setting = 1 / more
exclude: none best +tail                                          
dep	score	nodes	time	(not shown:  tbhits	knps	seldep)
 25	 -0.26 	41.4M  	0:48.43	g8h8 f3d1 d8d7 d5c3 e4c5 d2d4 e5d4 e3d4 a8e8 e1f1 c5e4 c4b5 e4c3 b5d7 c3d1 a1d1 e8e7 d7b5 a7a6 b5c4 b7b5 c4d3 b4a2 f1g2 a2b4 d3b1 b4d5 h1e1 h7h6 e1e7 d6e7 g5h6 g7h6 d1e1 e7d6 e1h1 d5f4 g2f3 h8h7 
 24	 -0.09 	38.4M  	0:44.95	g8h8 f3d1 d8d7 d5c3 e4c5 d2d4 e5d4 e3d4 a8e8 e1f1 c5e4 c4b5 e4c3 b5d7 c3d1 a1d1 e8e7 d7b5 a7a6 b5c4 b7b5 c4d3 b4a2 f1g2 a2b4 b2a3 e7d7 d4d5 b4d3 a3d6 d7d6 d1d3 f8d8 h1d1 h8g8 g2f3 g8f7 d3d4 f7g6 
 24	 -0.16 	36.6M  	0:42.83	g8h8 
 24	 -0.14 	29.3M  	0:34.19	g8h8 
 23	  0.00 	22.7M  	0:26.47	g8h8 f3d1 b4d5 c4d5 d8d7 d1c2 d6b4 d5e4 f5e4 b2e5 a8d8 e5d4 d7g4 c2c4 f8f2 e1f2 g4f3 f2g1 f3g4 g1f1 d8f8 d4f6 g4f3 f1g1 f3g4 g1f1 
 23	 -0.13 	21.1M  	0:24.53	g8h8 
 23	 -0.27 	20.5M  	0:23.88	g8h8 
 22	 -0.41 	17.0M  	0:19.78	g8h8 f3d1 d6c5 d5c3 a8c8 a1c1 c5b6 a2a3 b4d3 c4d3 d8d3 d1e2 d3d7 c3e4 c8c1 b2c1 f5e4 c1b2 d7f7 b3b4 f7d5 b2c3 f8f3 h4h5 b6d8 g5g6 d8g5 a3a4 a7a6 e1g1 h7g6 h5g6 g5h4 
 22	 -0.08 	14.6M  	0:16.98	g8h8 
 22	 -0.01 	13.6M  	0:15.83	g8h8 
 21	+0.93 	4.83M  	0:05.56	b7b5 
 21	 -0.15 	12.4M  	0:14.52	g8h8 f3d1 d6c5 d5c3 a8c8 a2a3 b4d3 c4d3 d8d3 d1e2 d3d6 e1g1 e4c3 b2c3 f5f4 e3f4 e5f4 e2f3 c8c7 g1h2 c5a3 b3b4 a3b4 a1a7 b4c3 d2c3 
 20	+0.79 	4.63M  	0:05.35	b7b5 d5b4 b5c4 b4d5 c4b3 f3e2 b3a2 e2c4 f8f7 f2f3 a8b8 f3e4 b8b2 e4f5 e5e4 f5f6 d8d7 c4e4 d7b5 d5f4 d6f4 e4f4 b5d3 a1d1 
 20	+1.22 	3.95M  	0:04.55	b7b5 
 20	+0.91 	3.46M  	0:03.99	b7b5 
 20	+0.70 	3.34M  	0:03.85	b7b5 
 20	+0.56 	3.23M  	0:03.74	b7b5 
 19	+0.42 	3.05M  	0:03.51	b7b5 d5b4 b5c4 b4d5 c4b3 f3e2 b3a2 a1a2 f8f7 d2d3 e4c5 g5g6 h7g6 h4h5 g6g5 h5h6 g7h6 h1h6 d6f8 h6g6 f7g7 g6g7 f8g7 e3e4 a7a5 b2a3 a8c8 a2b2 f5e4 d3e4 
 19	+0.62 	2.96M  	0:03.41	b7b5 
 19	+0.73 	2.68M  	0:03.08	b7b5 
 19	+0.59 	2.58M  	0:02.97	b7b5

Frank Quisinsky · Post by **Frank Quisinsky** » Wed Mar 24, 2021 8:31 am

Hi Günther,

the next good position to test it with all the TOP-Engines.

Thanks Günther, I will do later with with TOP-100 ...
At the moment I collect such positions!

Best
Frank

mar · Post by **mar** » Wed Mar 24, 2021 5:10 pm

well, positions happen.

and why top engines?
Cheng 4.40 sees that b5 is a blunder in 2.6 seconds, 4.4 seconds to see it's a HUGE blunder (-6.0) and in 7.37s switches to Kh8, not a big deal. and this is single-threaded with 32MB hash, from scratch
the same version/setup sees Ne7+ in about half a second, +4.0 score in 1 second, +7.5 in 6.
I'm pretty sure there are weaker engines out there that will find this even faster.
also - I'm quite sure Ethereal itself will find that move (after b5??) quickly

still - a nice miniature

that being said - kudos to Marvin, because you need to get this position against Ethereal first
I'm pretty sure Cheng wouldn't be able to win a single game against Ethereal in a 200 game blitz minimatch
that's what elo is about - to get more positions right than the opponent

Guenther · Post by **Guenther** » Wed Mar 24, 2021 5:40 pm

mar wrote: ↑Wed Mar 24, 2021 5:10 pm ...
I'm pretty sure Cheng wouldn't be able to win a single game against Ethereal in a 200 game blitz minimatch
...

I bet against it Martin ;-)

mar · Post by **mar** » Wed Mar 24, 2021 6:05 pm

Guenther wrote: ↑Wed Mar 24, 2021 5:40 pm I bet against it Martin

challenge accepted

4.40 against 12.75, CCRL blitz TC, 8movesGM.bin, 200 games (I don't have syzygy though, so probably a 20 elo handicap for Ethereal)

Guenther · Post by **Guenther** » Wed Mar 24, 2021 6:11 pm

mar wrote: ↑Wed Mar 24, 2021 6:05 pm
Guenther wrote: ↑Wed Mar 24, 2021 5:40 pm I bet against it Martin ;-)
challenge accepted :) 4.40 against 12.75, CCRL blitz TC, 8movesGM.bin, 200 games

I would like to use my usual 6 plies start positions (no really lopsided openings in - testing with those since a few years - it's also there for download, but right now I don't remember the link, need to check it first), but we can try both :)

May be my slow hardware even helps Cheng who knows, I will use 40/120s.

Edit
My 6 plies start pgn is just on my download page.
https://rwbc-chess.de/download.htm

JVMerlino · Post by **JVMerlino** » Wed Mar 24, 2021 7:17 pm

Even lowly Myrddin (8 cores) avoids b5 in less than a second.

Code: Select all

 1    199      0           49 b4d5 c4d5 g8h8 d5b7 f5e4 f3e4
 1    195      0          141 a8c8 d5b4 c8c4 b3c4 d6b4
 1    147      0          185 b4a6 d5b4
 1    104      0          187 b4c6 d5b4
 1   -121      1          212 b4c2 e1e2 c2a1 h1a1 (13 KNPS)
 2   -121      1          293 b4c2 e1e2 c2a1 h1a1 (18 KNPS)
 3   -105      1          733 b4c2? e1e2 (45 KNPS)
 3    -89      1        18893 b4c2? e1e2 (1180 KNPS)
 3    -57      1        45531 b4c2? e1e2 (2845 KNPS)
 3     -9      1        66773 b4c2? e1e2 (4173 KNPS)
 3     39      1        89617 b4c2 e1e2 c2a1 d5c7 g8h8 c7a8 (5601 KNPS)
 4     23      1       126087 b4c2! (7880 KNPS)
 4      7      1       144762 b4c2! (9047 KNPS)
 4     -8      3       167181 b4c2 e1d1 c2a1 (5392 KNPS)
 5    -21      3       195718 b4c2 e1e2 (6313 KNPS)
 6    -21      3       219167 b4c2 e1e2 (7069 KNPS)
 7    -24      4       392088 b4c2 e1e2 c2a1 (8342 KNPS)
 8    -13      9       995643 b4c2 e1e2 c2a1 d5c7 g8h8 c7a8 d6a3 b2a1 d8d2 e2f1 f8a8 a1e5 d2a2 f3f5 e4f2 (10591 KNPS)
 9    -13     14      1701447 b4c2 e1e2 c2a1 (12067 KNPS)
10      3     28      3726617 b4c2? e1e2 (13261 KNPS)
10    -29     62      8600539 b7b5! c4d3 (13760 KNPS)
10      3     78     10886268 b7b5? d5e7 b5c4 (13938 KNPS)
10     35     89     12507681 b7b5? d5e7 (14037 KNPS)
10      7    120     16916382 b4c2 e1e2 c2a1 d5c7 g8h8 c7a8 d8a5 h1d1 f8a8 f3f5 a5a2 b2a1 e4d2 (14061 KNPS)
11     23    132     18837865 b4c2? e1e2 (14185 KNPS)
11     39    184     25981881 b4c2? e1e2 (14089 KNPS)
11     25    189     26402109 g8h8 e1c1 (13961 KNPS)
12     41    248     34699805 g8h8? f3d1 b4d5 a3b4 (13969 KNPS)
12     57    367     50614377 g8h8? f3d1 (13783 KNPS)
12     44    485     66750056 g8h8 f3d1 a8c8 a2a3 b4d5 c4d5 d8b6 d5e6 c8c6 a1c1 d6b4 c1c6 b6c6 (13737 KNPS)
13     60    512     70370571 g8h8? f3d1 (13730 KNPS)
13     65    679     92826384 g8h8 f3d1 (13656 KNPS)
14     57    848    115878251 g8h8 f3d1 d8a5 d5c3 (13658 KNPS)
15     51   1273    173886405 g8h8 f3d1 d8a5 d5c3 e4c5 d2d4 e5d4 e3d4 f8e8 e1f1 c5e4 f1g1 a8c8 c3e4 e8e4 d1d2 (13655 KNPS)
16     35   2668    362590253 g8h8! (13586 KNPS)
16     19   3884    521405504 g8h8! (13423 KNPS)
16      4   5353    711799870 g8h8 f3d1 f5f4 h1h2 f4e3 d5e3 e4c5 d1b1 d6c7 e1e2 a8c8 g5g6 e5e4 h2g2 h7h6 h4h5 d8h4 (13296 KNPS)
17      9   8100   1066144657 g8h8 f3d1 f5f4 d5c7 c7c6 (13162 KNPS)

jm

Marvin 5.00 - Ethereal 12.75, 1:0 mate after 33!! ... great game!

Marvin 5.00 - Ethereal 12.75, 1:0 mate after 33!! ... great game!

Re: Marvin 5.00 - Ethereal 12.75, 1:0 mate after 33!! ... great game!

Re: Marvin 5.00 - Ethereal 12.75, 1:0 mate after 33!! ... great game!

Re: Marvin 5.00 - Ethereal 12.75, 1:0 mate after 33!! ... great game!

Re: Marvin 5.00 - Ethereal 12.75, 1:0 mate after 33!! ... great game!

Re: Marvin 5.00 - Ethereal 12.75, 1:0 mate after 33!! ... great game!

Re: Marvin 5.00 - Ethereal 12.75, 1:0 mate after 33!! ... great game!

Re: Marvin 5.00 - Ethereal 12.75, 1:0 mate after 33!! ... great game!

Re: Marvin 5.00 - Ethereal 12.75, 1:0 mate after 33!! ... great game!

Re: Marvin 5.00 - Ethereal 12.75, 1:0 mate after 33!! ... great game!