Nezhmetdinov - Chernikov - 1962 - Queen Sac

[reflectionofpower]

Don't know about this one, but a lot of Tal's sacrifices were unsound. A move can give great practical chances because it gives the opponent a hard defensive task, and they can err, but in many cases there is a possible defense or even a refutation.

--Jon

Yeah, That's what he did verse Botvinnik in 60. He just played as his usual self and Botvinnik played logical so of course this messed him up. The rematch in the year after Botvinnik kept things simple and nullified this advantage Tal had and won the title back.

[yanquis1972]

Don't know about this one, but a lot of Tal's sacrifices were unsound. A move can give great practical chances because it gives the opponent a hard defensive task, and they can err, but in many cases there is a possible defense or even a refutation.

--Jon

i think thats what makes this sac so special in the computer age. engines have no problem walking the tight-rope tal laid out for his opponents, but here the move not only springs from such a seemingly mundane position, is in fact a technically sound positional sacrifice. which might surprise nez himself, were he around today -- i'm reminded of the tal quote "there are 2 kinds of sacrifices; sound ones, & mine". these guys just wanted chaos on the board! i only hope we see another like tal in my lifetime.

[reflectionofpower]

Don't know about this one, but a lot of Tal's sacrifices were unsound. A move can give great practical chances because it gives the opponent a hard defensive task, and they can err, but in many cases there is a possible defense or even a refutation.

--Jon

i think thats what makes this sac so special in the computer age. engines have no problem walking the tight-rope tal laid out for his opponents, but here the move not only springs from such a seemingly mundane position, is in fact a technically sound positional sacrifice. which might surprise nez himself, were he around today -- i'm reminded of the tal quote "there are 2 kinds of sacrifices; sound ones, & mine". these guys just wanted chaos on the board! i only hope we see another like tal in my lifetime.

Nowadays I personally feel that VLM,Nakamura,Mamedyarov fit this description but out of them all I think Rapport is the maddest of them all.

[yanquis1972]

thinking a bit about how computers interpret this position, i dont think its terribly easy to get a top engine to prefer Qxf6. ideally, simply turning on or up contempt would give a substantial bonus to such a move, but bc its not only trading down material, but taking the queen off the board, i imagine the penalties there are too great.

[reflectionofpower]

thinking a bit about how computers interpret this position, i dont think its terribly easy to get a top engine to prefer Qxf6. ideally, simply turning on or up contempt would give a substantial bonus to such a move, but bc its not only trading down material, but taking the queen off the board, i imagine the penalties there are too great.

I even tried Chess Tiger 2007 Gambit Suicidal mode and it still picks Qh6. Chess System Tal II picks this move too. I would be interested to see if any engine would pick Qxf6.

[yanquis1972]

stock settings, probably not, but eg a "stock" komodo personality i have in my engines list (K10. but with dynamism & king safety cranked up to max) is quick to favor Qxf6 before discarding it.

i'd very much like to see an 'official' dynamic komodo, tuned by LK himself -- rybka 3 dynamic blew my mind in this position (didnt prefer Qxf6 either, but it was instantly its 2nd favored move, which was purely astonishing at the time)

New game
r1b2rk1/pp1ppp1p/5bp1/q7/3nP2Q/1BN1B3/PPP2PPP/R4RK1 w - - 0 1

Analysis by Komodo 10.2 64-bit TAL:

12.Qh3 d5
White is clearly better: +/- (1.44 --) Depth: 7 00:00:00 23kN
White is better: +/- (0.97 --) Depth: 10 00:00:00 116kN
12.Qg4 d6 13.Qd1 Nc6 14.Nd5 Bxb2 15.Rb1 Bg7 16.Bd2 Qd8 17.Be3
White is better: +/- (0.98 --) Depth: 10 00:00:00 134kN
The position is equal: = (0.15) Depth: 10 00:00:00 229kN
12.Qh6 Bg7
The position is equal: = (0.17) Depth: 10 00:00:00 237kN
The position is equal: = (0.19 --) Depth: 16 00:00:00 1989kN
12.Qg4 d6 13.Qd1 Nc6 14.Qe1 Qh5 15.f4 Be6 16.Rd1 Bxb3 17.axb3 Nb4 18.Qd2 Rac8 19.h3 Rfe8 20.Ra1 Nc6
White has an edge: = (0.29) Depth: 16 00:00:01 2590kN
12.Qxf6 Ne2+
White is slightly better: +/= (0.31 ++) Depth: 16 00:00:01 2624kN
White has an edge: = (0.27 --) Depth: 17 00:00:01 3436kN
12.Qg4 d6
White is slightly better: +/= (0.47 ++) Depth: 17 00:00:01 4020kN
White is slightly better: +/= (0.58 --) Depth: 18 00:00:01 5117kN
12.Qxf6 Ne2+
White is slightly better: +/= (0.64) Depth: 18 00:00:01 6022kN
White is slightly better: +/= (0.43 --) Depth: 19 00:00:01 7169kN
12.Qg4 d6
White is slightly better: +/= (0.50) Depth: 19 00:00:01 7419kN
White is slightly better: +/= (0.44 --) Depth: 20 00:00:01 8453kN
12.Qxf6 Ne2+
White is slightly better: +/= (0.56 ++) Depth: 20 00:00:01 9147kN
White is slightly better: +/= (0.51 --) Depth: 21 00:00:02 11687kN
12.Qg4 d6
White is slightly better: +/= (0.63 ++) Depth: 21 00:00:02 17763kN
White is slightly better: +/= (0.48 --) Depth: 21 00:00:02 19180kN
12.Qxf6 Ne2+
White is slightly better: +/= (0.55) Depth: 21 00:00:02 19201kN
White is slightly better: +/= (0.48 --) Depth: 26 00:00:14 136MN, tb=1
12.Qg4 d6 13.Qd1 Nc6 14.Bh6 Re8 15.f4 Be6 16.f5 Qc5+ 17.Kh1 Bxb3 18.axb3 Bxc3 19.bxc3 Ne5 20.c4 a5 21.Ra4 f6 22.Qe2 Qc6 23.Bf4 b6 24.Bh6 Red8 25.Qd2 Kh8 26.fxg6 Nxg6 27.Qd5 Qxd5 28.exd5 Rdc8 29.g4 Ne5 30.h3 b5 31.cxb5 Rxc2 32.b6
White is better: +/- (0.77 ++) Depth: 26 00:00:19 181MN, tb=1
White is better: +/- (0.85) Depth: 29 00:01:03 560MN, tb=1

[peter]

Hi!

ProDeo Q3 Tactical Engine:

r1b2rk1/pp1ppp1p/5bp1/q7/3nP2Q/1BN1B3/PPP2PPP/R4RK1 w - - 0 1

Analysis by ProDeo 2.2 SMP:
...
12.Dh6 Lg7 13.Dg5 Dxg5 14.Lxg5 e6 15.Tfd1 f5 16.La4 fxe4 17.Le3 Sc6 18.Sxe4 d5 19.Sd6 Le5 20.Sxc8 Taxc8 21.c3 b6 22.Tac1 Sa5
+/= (0.52) Tiefe: 17 00:00:25 59344kN
12.Dxf6 Se2+ 13.Sxe2 exf6 14.Sf4 d6 15.Tad1 Td8 16.Sd5 Le6 17.Sxf6+ Kg7 18.Sd5 Tac8 19.Ld4+ Kh6 20.c3 Te8 21.f4 Lg4
+/= (0.57) Tiefe: 17 00:00:34 80866kN
...
12.Dxf6 Se2+ 13.Sxe2 exf6 14.Sf4 d6 15.Tad1 b5 16.Txd6 Lb7 17.Txf6 Lxe4 18.Ld5 Lxc2 19.Tc6 Lf5 20.Tcc1 Tfd8 21.Lxa8 Txa8 22.b4 Dxb4 23.Lxa7
+/= (0.52) Tiefe: 21 00:02:43 402MN
12.Dh6
+/= (0.52) Tiefe: 21 00:03:41 550MN

So even this legendary sac- monster switches to another move later on again.

Without any alternative the move so far as for my few trials was only for an old Junior- setting of mine:

r1b2rk1/pp1ppp1p/5bp1/q7/3nP2Q/1BN1B3/PPP2PPP/R4RK1 w - - 0 1

Analysis by Deep Junior 11.2 UCI1553015:

12.Dxf6
+/- (1.21) Tiefe: 9 00:00:00 74kN
...
12.Dxf6 Se2+ 13.Sxe2 exf6 14.Sc3 Te8 15.Sd5 Te6 16.Ld4 Kg7 17.Tae1 b5 18.Te3 Lb7 19.Sxf6 Txf6 20.Td1 Dd8 21.Tf3
+- (2.08) Tiefe: 22 00:06:58 12616MN

BTW in Hiarcs CE online GM++ the move is stored with a ?, Qg4 has !? and Qh6 has !.
To me the sac isn't unsound, but I'd not call it the best move neither. I'd not give a Queen for no proveable advantage, especially not, if I'd want to play aggressively.
And that might be the reason why even "aggressive" settings don't chose it, if you don't lower the material eval of the Queen e.g., like I did here with F15. Black's Queen- value lowered for the worth of 3 Pawns:

r1b2rk1/pp1ppp1p/5bp1/q7/3nP2Q/1BN1B3/PPP2PPP/R4RK1 w - - 0 1

Analysis by Fritz 15:

12.Dxf6 Se2+ 13.Sxe2 exf6 14.Sc3 Dd8 15.Tfd1 b6 16.f3 Lb7 17.La4
=/+ (-0.61) Tiefe: 7 00:00:00 8kN
...
12.Dxf6 Se2+ 13.Sxe2[] exf6[] 14.Sc3 d6 15.Sd5 Le6 16.Sxf6+[] Kg7 17.Ld4 Tfc8[] 18.Sd7+ Kg8 19.Sf6+[] Kg7 20.Sd7+ Kg8 21.Sf6+[] Kg7 22.Sd7+ Kg8 23.Sf6+[] Kg7 24.Sd7+ Kg8 25.Sf6+[] Kg7 26.Sd7+ Kg8 27.Sf6+[]
= (0.00) Tiefe: 27 00:02:04 1162MN

Finally I have here some more SF- output. After Forward- Backward of the main lines with 12.Qxf6, 12.Qg4 and 12.Qh6, the Queen- sac in 6MV is among the 4 best ones at once, rising for one place later on.
That's at depth 41, after soon reduction to 3MV:

r1b2rk1/pp1ppp1p/5bp1/q7/3nP2Q/1BN1B3/PPP2PPP/R4RK1 w - - 0 1

Analysis by Brainfish 031216 64 POPCNT:

1. = (0.11): 12.Dg4 d6 13.Dd1 Sc6 14.Lh6 Td8 15.Ld2 Dc5 16.Le3 Da5 17.Df3 Tf8 18.Lh6 Te8 19.Ld2 Le6 20.Tad1 Sd4 21.Dd3 Sxb3 22.axb3 Dc5 23.Le3 Dc6 24.Ld4 Lxd4 25.Dxd4 Tec8 26.f4 Dc5 27.f5 gxf5 28.exf5 Lxf5 29.Sd5 Lxc2 30.Sxe7+ Kf8 31.Sxc8 Lxd1 32.Txf7+ Kg8 33.Tg7+ Kh8 34.Dxc5 dxc5 35.Tc7 h6 36.Txc5 Lxb3 37.Tc7 b6 38.Kf2 Le6 39.Se7 Kh7 40.Sc6+ Kg6 41.Txa7 Txa7 42.Sxa7

2. = (0.08): 12.Dh6 Lg7 13.Dg5 Sxb3 14.axb3 Dxg5 15.Lxg5 Lxc3 16.bxc3 f6 17.Lh6 Te8 18.Le3 a6 19.c4 d6 20.f3 g5 21.Kf2 Kf7 22.Ta5 Le6 23.Lb6 Tg8 24.Ke3 Tac8 25.Kd3 h5 26.h3 Th8 27.Th1 Tcg8 28.Le3 Ld7 29.c5 g4 30.hxg4 hxg4 31.Txh8 Txh8 32.cxd6 gxf3 33.gxf3 Th3 34.dxe7 Txf3

3. = (0.00): 12.Dxf6 Se2+ 13.Sxe2 exf6 14.Sf4 Kg7 15.Sd5 d6 16.Ld4 h5 17.Sxf6 Kh6 18.Le3+ Kg7 19.Ld4

[yanquis1972]

I'd not give a Queen for no proveable advantage, especially not, if I'd want to play aggressively.

i trust you would not, yet this is exactly what great aggressive players do; the value in completely upending the position, in creating a dense web through which their opponent must struggle over the board to find a correct path, outweighs the objective evaluation. similarly, in a hypothetical situation where a much stronger player has the white pieces (contempt), Qf6 is the best move on the board, unless...

in my checking i found that the new engines really like Qg4...peculiar move on the surface, but it may be that that hiarcs book needs revision. (i certainly don't think Qf6 merits a ?)


just because, some mpv analysis from the komodo personality earlier (thru d35; contempt off) --

Analysis by Komodo 10.2 64-bit TAL:

1. +/- (0.95): 1.Qg4 d6 2.Qd1 Nc6 3.Bh6 Bg7 4.Bxg7 Kxg7 5.Kh1 Be6 6.f4 Kh8 7.Qd3 Rae8 8.Rad1 Bg4 9.Rde1 f6 10.Nd5 Bd7 11.c3 Qd8 12.Rd1 Na5 13.Bc2 Qc8 14.Qg3 e6 15.Ne3 Nc4 16.Nxc4 Qxc4 17.Qf2 Qa6 18.Bd3 Qb6 19.Qxb6 axb6 20.Bc4 Bc6 21.Rxd6 Bxe4 22.Bb5 Bc6 23.Bxc6 bxc6 24.Rxc6
2. +/= (0.41): 1.Qxf6 Ne2+ 2.Nxe2 exf6 3.Nc3 Kg7 4.Nd5 Re8 5.Bd4 Re6 6.Rad1 b6 7.Nc7 Bb7 8.Bxe6 dxe6 9.Nxa8 Bxa8 10.f3 e5 11.Bc3 Qc5+ 12.Kh1 Bc6 13.Rd8 Bb5 14.Rd5 Bxf1 15.Rxc5 bxc5 16.Kg1 Bb5 17.Bd2 a6 18.Kf2 f5 19.exf5 gxf5 20.b3 f4 21.c4 Bc6 22.g3 fxg3+ 23.hxg3 Kf6 24.Be3 Ke6 25.Bxc5 f6

[Lyudmil Tsvetkov]

Don't know about this one, but a lot of Tal's sacrifices were unsound. A move can give great practical chances because it gives the opponent a hard defensive task, and they can err, but in many cases there is a possible defense or even a refutation.

--Jon

Tal is another case, most of the games I have seen of Nezhmetdinov are fully correct.

[Lyudmil Tsvetkov]

sorry lyudmil, youre going to have to prove Qf6 is a !!; ive never seen any evidence that it wins against best play. i would give it !!? if such a notation existed.

my personal favorite move ever played, & the very first position i check with any new engine, but for it to be considered a test position, Qxf6 has to be proven demonstrably better than the alternatives. id love to see that done, but im extremely skeptical.

which line is not completely clear to you?

my static eval for the position arising after the sac: Qf6 Ne2 Ne2 gf6 is 60cps white advantage: d7 is isolated central, d7 is also backward central, due to the fact that pieces of lower power, minors, attack the d5 square, some 100cps white edge, so d7 pawn migth be written off as any material asset, I have special terms about pair of bishops with different configurations, which gives more white edge, special terms for imbalance in the position, which favours white by at least 120cps, white has more compact pawn structure, etc., etc., so, in actual fact, white is leading in terms of eval in the root by at least 60cps, which is a straightforward win in the mg, though top engines of course do not see this, as they have very very basic eval, and there is nothing decisive and short-term for the search to find.

so, I am pretty much certain white wins this, let's say, 97%.