Interesting position from a Blunder 7.60 game CCRL 40/15

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

Moderators: hgm, Dann Corbit, Harvey Williamson

User avatar
Guenther
Posts: 4584
Joined: Wed Oct 01, 2008 6:33 am
Location: Regensburg, Germany
Full name: Guenther Simon

Interesting position from a Blunder 7.60 game CCRL 40/15

Post by Guenther »

Unstoppable passer problem with a piece up, but tied to it.

[pgn][Event "CCRL 40/15"]
[Game "98"]
[Date "2022.01.12"]
[Round "806.3.442"]
[White "Blunder 7.6.0 64-bit"]
[Black "KnightX 3.1 64-bit"]
[Result "0-1"]
[ECO "A29"]
[Opening "English"]
[Variation "four knight kingside fianchetto"]
[PlyCount "91"]
[WhiteElo "2636"]
[BlackElo "2683"]

1. c4 {+0.00/1 0} e5 {+0.00/1 0}
2. Nc3 {+0.00/1 0} Nf6 {+0.00/1 0}
3. Nf3 {+0.00/1 0} Nc6 {+0.00/1 0}
4. g3 {+0.00/1 0} Bb4 {+0.00/1 0}
5. Bg2 {+0.00/1 0} O-O {+0.00/1 0}
6. O-O {+0.00/1 0} Bxc3 {+0.00/1 0}
7. bxc3 {+0.00/1 0} d6 {+0.00/1 0}
8. d3 {+0.00/1 0} e4 {+0.00/1 0}
9. Nd4 {+0.44/19 30} Ne5 {-0.42/21 30}
10. dxe4 {+0.49/20 30} Nxc4 {-0.33/20 30}
11. Rb1 {+0.46/19 39} h6 {-0.33/21 30}
12. Qd3 {+0.65/16 29} Nb6 {-0.39/21 44}
13. Nf5 {+0.42/17 29} Be6 {-0.42/19 29}
14. Qc2 {+0.49/17 38} Re8 {-0.36/18 29}
15. c4 {+0.48/16 29} Rb8 {-0.27/20 29}
16. c5 {+0.41/18 29} dxc5 {-0.09/20 29}
17. Rd1 {+0.31/18 29} Nbd7 {-0.03/21 29}
18. Bf4 {+0.28/17 29} Nh5 {-0.06/21 29}
19. Be3 {+0.40/19 29} b6 {-0.06/21 29}
20. Bc1 {+0.32/19 38} Nhf6 {-0.06/20 29}
21. f4 {+0.51/17 28} Bxf5 {-0.30/19 29}
22. exf5 {+0.37/21 28} Qe7 {-0.12/22 29}
23. e4 {+0.32/20 28} c4 {-0.01/20 29}
24. Kh1 {+0.24/19 61} Nxe4 {+0.54/21 42}
25. Bxe4 {+0.50/23 26} Qxe4+ {-0.01/23 28}
26. Qxe4 {+0.20/25 35} Rxe4 {-0.01/26 28}
27. Rxd7 {+0.20/24 34} Re1+ {-0.01/23 28}
28. Kg2 {+0.57/24 33} c5 {-0.24/23 28}
29. Rd2 {+0.60/23 37} Rbe8 {-0.27/24 28}
30. Kf3 {+0.58/24 23} b5 {-0.15/25 28}
31. Rc2 {+0.45/21 23} Rf1+ {-0.15/26 28}
32. Kg2 {+0.45/23 24} Ree1 {-0.12/25 28}
33. a4 {+0.42/23 23} Rg1+ {-0.02/26 28}
34. Kf3 {+0.41/26 23} a6 {-0.02/24 28}
35. Kf2 {+0.41/23 23} Kh7 {-0.01/23 28}
36. Ra1 {+0.41/22 23} h5 {-0.01/23 28}
37. Rb1 {-0.19/20 17} Rgf1+ {-0.01/22 28}
38. Kg2 {+0.00/24 25} Rg1+ {+0.00/25 28}
39. Kh3 {-0.18/21 6} b4 {-0.01/26 28}
40. Kh4 {-0.25/23 5} c3 {+5.46/27 28}
41. Kxh5 {-0.25/22 32} Rxc1 {+6.33/26 24}
42. Rbxc1 {-4.72/29 92} Rxc1 {+6.75/27 23}
43. Rxc1 {-4.92/28 109} c4 {+6.84/25 24}
44. Kg4 {-4.94/27 76} b3 {+6.99/27 23}
45. Rxc3 {-5.22/25 86} b2 {+7.08/27 23}
46. Rxc4 {-5.15/24 22}
0-1[/pgn]

Image
full size graph => https://rwbc-chess.de/Graphs/Blunder_760_98.png

Current 8.21 dev (for my hardware compiled) gets 40. Rxc4 now in depth 27 at least.
(it's the only move to draw!)
I don't know if 41...Rxc1! is pruned away, even in the PVs for 8.21 dev it does not
appear and also no major fail low visible despite changing to 40. Rxc4 in depth 27.

In the original game it only used 5 seconds for move 40, but I don't know if it
was low on time. (could be calculated though by guessing the original tc)

This is when walking through the whole pgn while analysing to fill the hash,
not directly fed a FEN.

Code: Select all

exclude: none best +tail                                          
dep	score	nodes	time	(not shown:  tbhits	knps	seldep)
 31	  0.00 	14.0M  	2:06.51	c2c4 e1e2 b1b2 g1c1 c4c1 e2b2 c1c5 b4b3 c5c6 b2a2 c6a6 b3b2 a6b6 f7f6 b6b5 h7h6 b5b8 h6h7 b8b6 h7h6 a4a5 a2a5 b6b2
 30	 -0.02 	10.2M  	1:44.36	c2c4 e1e2 b1b2 g1c1 c4c1 e2b2 c1c5 b2d2 c5c6 a6a5 c6c5 d2d6 c5b5 d6a6 h3h4 h7h6 b5b7 a6c6 b7b5 c6c2 h4h3 h5h4 b5a5 h4g3 h2g3 b4b3 a5a6 f7f6 a6a8 g7g6 h3g4
 29	 -0.21 	8.40M  	1:28.07	c2c4 e1e2 b1b2 g1c1 c4c1 e2b2 c1c5 b2d2 c5c6 a6a5 c6c5 d2d6 c5b5 d6a6 h3h4 h7h6 b5b7 a6c6 b7b5 c6c2 h4h3 c2e2 b5a5 b4b3 a5b5 b3b2 a4a5 g7g6 f5f6
 28	 -0.22 	7.05M  	1:14.46	c2c4 e1e2 b1b2 g1c1 c4c1 e2b2 c1c5 b2d2 c5c6 d2d8 a4a5 b4b3 c6b6 d8d5 b6b3 d5f5 b3a3 h7g6 h3g2 f5b5 g2f3 g6f5 h2h4 b5c5
 27	 -0.21 	9.66M  	1:02.88	c2c4 e1e2 b1b2 g1c1 c4c1 e2b2 c1c5 b2d2 c5c6 d2d8 a4a5 b4b3 c6b6 d8d5 b6b3 d5f5 b3a3 h7g6 h3g2 f5d5 g2f3 g6f5 h2h4 d5c5 f3e3 f5g4 e3d4
 26	 -0.52 	7.49M  	0:46.65	h3h4 c4c3 h4h5 e1d1 a4a5 g1e1 g3g4 d1d5 h5h4 f7f6 b1b3 d5d3 b3b1 d3d2 c1d2 c3d2 c2d2 e1b1 d2d6
 25	 -0.57 	4.40M  	0:33.45	h3h4 c4c3 h4h5 e1d1 a4a5 g1e1 g3g4 d1d2 c1d2 c3d2 c2d2 e1b1 d2d6 b1h1 h2h4 b4b3 d6b6 c5c4 g4g5 g7g6 f5g6 f7g6 h5g4 h1a1 b6a6 b3b2 a6a7 h7g8 a7a8 g8f7
 24	 -0.62 	6.53M  	0:25.79	h3h4 c4c3 h4h5 e1d1 a4a5 g1e1 g3g4 d1d2 c1d2 c3d2 c2d2 e1b1 g4g5 g7g6 f5g6 f7g6 h5h4 b4b3 d2d7 h7h8 d7a7 b1f1 a7a8 h8g7 a8a7 g7f8 a7a8 f8e7 a8a7 e7d8 a7a6 f1f4 h4g3
 23	 -0.52 	1.67M  	0:14.09	h3h4 c4c3 h4h5 e1d1 f5f6 g7f6 a4a5 g1e1 h2h4 d1d3 h5g4 e1g1 g4f5 h7g7 g3g4 g1e1 g4g5 d3d5 f5g4 f6g5 h4g5 g7g6 b1b3 d5d3
 22	 -0.38 	1.13M  	0:11.19	h3h4 c4c3 h4h5 e1d1 f5f6 g7f6 a4a5 g1e1 h2h4 d1d3 h5g4 e1g1 g4f5 h7g7 g3g4 g1e1 g4g5 d3d5 f5g4 f6g5 f4g5 d5d4 g4g3 d4d3 g3f4
 21	 -0.35 	648067	0:09.24	h3h4 c4c3 h4h5 e1d1 f5f6 g7f6 h5g4 g1e1 h2h4 d1d2 c1d2 c3d2 c2d2 e1b1 d2d6 h7g7 d6a6 b4b3 a6c6 b1a1 h4h5
 20	 -0.40 	796491	0:08.10	h3h4 c4c3 h4h5 e1d1 f5f6 g7f6 a4a5 g1e1 f4f5 d1d2 c1d2 c3d2 c2d2 e1b1 d2d6 h7g7 d6a6 b1f1 h2h4 f1f5 h5g4
 19	 -0.33 	358182	0:06.67	h3h4 c4c3 h4h5 e1d1 f5f6 g7f6 a4a5 g1e1 h5g4 h7g7 g4f5 d1d2 c1d2 c3d2 c2d2 e1b1 d2d6 b4b3 d6a6
 18	 -0.33 	325144	0:06.02	h3h4 c4c3 h4h5 e1d1 f5f6 g7f6 a4a5 g1e1 h2h4 d1d3 h5g4 f6f5 g4f5 d3g3 f5f6 g3g7 h4h5 e1h1 c2e2 h1h5
 17	 -0.30 	219984	0:05.44	h3h4 c4c3 h4h5 e1d1 h5g5 f7f6 g5h5 g1e1 g3g4 d1d2 c1d2 c3d2 g4g5 e1b1 g5g6 h7g8 c2d2 g8f8 d2d7
 16	 -0.20 	244870	0:05.05	h3h4 c4c3 h4h5 e1d1 h5g5 g1e1 f5f6 d1d2 c1d2 c3d2 c2d2 e1b1 d2d7 h7g8 d7a7 g7f6 g5f6
 15	 -0.03 	2.38M  	0:04.60	h3h4 c4c3 h4h5 g1h1 h5g5 f7f6 g5g4 e1d1 g4f3 h1e1 f3f2 h7g8 h2h4 g8f7 h4h5 e1e8
 14	+0.61 	39588  	0:00.26	c2c4 g1h1 f5f6 e1e2 h3h4 h1h2 h4g5 e2g2 f6g7 h7g7 c4c5 g2g3 g5f5 b4b3 c1b2 g7h7 b2f6
 13	+0.63 	28807  	0:00.18	c2c4 g1h1 f5f6 e1e2 h3h4 h1h2 h4g5 e2g2 f6g7 h7g7 g5f5 g2c2 c4c2 h2c2
 12	+0.74 	31079  	0:00.13	c2c4 g1h1 f5f6 e1e2 h3h4 h1h2 h4g5 e2g2 f6g7 h7g7 g5f5 g2g3 c1b2 f7f6 c4c5
 11	+0.68 	25086  	0:00.08	c2c4 g1h1 f5f6 e1e2 h3h4 g7f6 h2h3 e2h2 h4h5 h2h3 h5g4 h3h5
 10	+0.80 	8626 	0:00.03	c2c4 b4b3 c4c5 b3b2 b1b2 e1c1 c5c1 g1c1 b2b7 h7g8
https://rwbc-chess.de

trollwatch:
Chessqueen + chessica
User avatar
algerbrex
Posts: 596
Joined: Sun May 30, 2021 5:03 am
Location: United States
Full name: Christian Dean

Re: Interesting position from a Blunder 7.60 game CCRL 40/15

Post by algerbrex »

Interesting, thanks for looking into this so deeply.

Indeed, 8.21 finds Rxc4 at depth 24 on my machine after 3 1/2 minutes, using the default, non-optimized compile. With all the pruning Blunder has no, likely quite badly tuned by me, I'm not surprised it misses Rxc1. You're likely correct in that Rxc1 is being pruned away, since of course at first, it looks like a losing capture, so Blunder likely dismisses it fairly quickly, only releasing much later the danger of the move.

I'm going to experiment with playing around with a couple of areas in the search that I think might be pruning away Rxc1. I'll probably also take a look at the eval as well, and report back if I'm able to make any progress.