Nice queen sacrifice from Lc0

[Eelco de Groot]

This is not very interestng but Crystal is much better suited at investigating because it immediately detects the queen sac. Now I just changed the position (from the game) with white King much closer to the pawns and not in check because Queen would be moved as well.
After Nf5, now the Queen sacrifice no longer works because White can take back with 2.Kxf5{!} Black has not much better (as an alternative to the sac) than going back to pinning the Knight (because of the King on g4), and it seemed for a while this was a fortress (again) but Crystal discovers that White seemingly still can make enough progress, I just judged from the score, that White is winning in this fictitious position (diagram)

[d]3b2k1/5qrn/3p3p/2pPp1pP/1pP1P1KN/pP6/P5Q1/6RR w - -
1.Nf5 Qxf5 {?!} 2.Kxf5

After 1... Qd7 {probably best}, the best five moves:


[d]3b2k1/3q2rn/3p3p/2pPpNpP/1pP1P1K1/pP6/P5Q1/6RR w - -

Engine: Crystal 6 PMT (1024 MB)
gemaakt door the Stockfish developers (see AUTHORS f

45 127:41 +2.52 2.De2 Tf7 3.Tf1 Lf6 4.Tf2 Lg7 5.Thf1 Kh8
6.Kg3 Tf8 7.Kg2 Tf6 8.Kh2 Tf7 9.Df3 Lf8
10.Dd3 Lg7 11.Dh3 Tf8 12.Kg2 Tf7
13.Tf3 Tf8 14.Tg3 Dd8 15.Kg1 (6.501.335.060) 848

44 127:41 +2.53 2.Tf1 Tf7 3.De2 Tf8 4.Tf2 Kh8 5.Thf1 Lf6
6.Kg3 Lg7 7.Kg2 Tf6 8.De3 Lf8 9.Kh2 Tf7
10.Dg3 Lg7 11.Dh3 Tf8 12.Kg2 Tf7
13.Tf3 Tf8 14.Tg3 Ta8 15.Kg1 (6.501.335.060) 848

44 127:41 +2.34 2.Dc2 Tf7 3.Tf1 Lf6 4.Tf2 Lh8 5.De2 Tf6
6.Thf1 Kf8 7.Kg3 Ke8 8.Dg4 Kd8 9.Kg2 Ke8
10.Dg3 Dd8 11.Tf3 Dd7 12.Kg1 Dc8
13.Dg4 Dd7 14.Kh2 Kd8 15.Kg2 (6.501.335.060) 848

44 127:41 +2.31 2.Tb1 Tf7 3.Tbf1 Lf6 4.De2 Lh8 5.Tf2 Tf6
6.Thf1 Kf8 7.Kg3 Ke8 8.Kh2 Dc8 9.Dg4 Dd7
10.Kg2 Kd8 11.Tf3 Ke8 12.Dg3 Dc7
13.Df2 Kd8 14.Kg1 Dd7 15.Pxh6 (6.501.335.060) 848

44 127:41 +2.23 2.Te1 Tf7 3.Tef1 Lf6 4.De2 Lh8 5.Tf2 Tf6
6.Thf1 Kf8 7.Kg3 Ke8 8.Kg2 Dc8 9.Dg4 Dd7
10.Dg3 Dd8 11.Tf3 Dd7 12.Dh3 Kd8
13.Dg4 Ke8 14.Dg3 Dc7 15.Df2 (6.501.335.060) 848

I let this one calculate a bit further just out of curiosity and because the score went up only very slowly. Now it is a bit clearer after a few moves further and much machinetime:
[pgn]
[Event "?"]
[Site "?"]
[Date "2025.04.11"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "*"]
[SetUp "1"]
[FEN "3b2k1/5qrn/3p3p/2pPp1pP/1pP1P1K1/pP4N1/P5Q1/6RR w - -"]

1. Nf5 Qd7 2. Rf1 Rf7 *
[/pgn]

[d]3b2k1/3q1r1n/3p3p/2pPpNpP/1pP1P1K1/pP6/P5Q1/5R1R w - -

Engine: Crystal 6 PMT (1024 MB)
gemaakt door the Stockfish developers (see AUTHORS f

54 1180:11 +3.34 3.De2 Pf6+ 4.Kg3 Kh7 5.Th2 g4 6.Th4 Le7
7.Kg2 Dd8 8.Th2 Df8 9.Kg1 Ld8
10.Thf2 Le7 11.Kh1 Kh8 12.Kg2 Kh7
13.Kg1 Kh8 14.Ph4 De8 15.De3 Kh7
16.Pf5 (58.402.185.470) 824

54 1180:11 +3.21 3.Th2 Lf6 4.De2 Lg7 5.Thf2 Kh8 6.Kh3 Tf8
7.Kh2 Tf7 8.Dd3 Lf8 9.Dh3 Tf6 10.Dg2 Lg7
11.Dg4 Tf7 12.Dh3 Tf8 13.Kg2 Tf7
14.Tf3 Tf8 15.Tg3 Ta8 16.Kg1 (58.402.185.470) 824

53 1180:11 +3.00 3.Th3 Kh8 4.De2 Pf6+ 5.Kg3 Ph7
6.Pxh6 Tf4 7.Kg2 Pf6 8.Thf3 Dh7
9.Txf4 exf4 10.Pf5 Le7 11.h6 Lf8
12.Kg1 Dg6 13.Pxd6 Lxd6 14.e5 Le7
15.d6 Pg8 16.dxe7 (58.402.185.470) 824

53 1180:11 +2.99 3.Tf2 Pf6+ 4.Kg3 Kh7 5.Tff1 g4 6.De2 Tf8
7.Kg2 Tg8 8.Tf2 Tg5 9.Pg3 De8 10.Df1 Df8
11.Dd1 Le7 12.Thf1 Dd8 13.Tf5 Tg7
14.De2 Tg8 15.Dd2 Tf8 16.T5f2 (58.402.185.470) 824

53 1180:11 +2.96 3.Dc2 Kh8 4.De2 Le7 5.Tf2 Pf6+ 6.Kh3 Pe8
7.Kg3 Lf8 8.Thf1 Kh7 9.Kh2 Tf6
10.Dg4 Df7 11.Kg2 Kg8 12.Ph4 Tf4
13.Txf4 exf4 14.Pg6 Pg7 15.Te1 Dc7
16.Dh3 (58.402.185.470) 824