How can you resist a note like that?

33 Bxg6 Nxg6.
34 h8=Q+ Nxh8.
35 Qxh8+ Qxh8.
36 Rxh8+ Kg7!

Lasker was far ahead of his time, and it is hard
to blame Steinitz for his mistake: he fought with all his strength while under relentless, powerful assault. It is precisely because this exceptionally tense game was so far ahead of its time that it went under a cloud, remaining unappreciated: its contemporaries were simply unable to fathom what was going on here" (Kasparov).http://www.chesscafe.com/text/dvore...

Lasker vs Steinitz, 1894.
YOU ARE PLAYING THE ROLE OF LASKER.
Your score: 115 (par = 79)

a)...Kd8 37.Rxf5 Qe6 38.Rxd5+ Kc7 39.Qh2+ Re5 40.Rc5+ Kd6 41.Rxe8 Qxe8 42.Rc4 and White cannot make progress, but black is completely tied down.

b)...Kd7 37.Rxf5 Qe6 38.Qg4 Re1+ 39.Ka2 Re4 40.Qg7+ Re7 41.Qf8 Re8 42.Qg7+ draw.

Later, 40...Re7 Loses, But now the position is probably completely lost anyway.

This is an amazing game. I looked at it recently with my Cunningham book of the 1894 match, Kasparov’s OMGP I, and Stockfish. Some interesting kibitzes on the page too. :-)

Going into this seventh game, the match was knotted 2:2, with two draws. This game was the first of five straight Lasker victories. Quotes from the annotators are in brackets. I put the moves from the 19th century analysts in algebraic notation. Comments from me or Stockfish are in plain text.

1.e4 e5 2.Nf3 Nc6 3.Bb5 d6 4.d4 Bd7 5.Nc3 Nge7 6.Be3

In his first three games with White in this match, Lasker had played 6.Bc4, taking immediate aim at f7. <White varies his attack [from 6.Bc4]; but he does not improve it. If this piece is to move, then 6.Bg5, as played by Tschigorin v. Steinitz, and by Showalter v. Hodges, is probably stronger. The procedure, hereabouts, in the first game of the recent match between the two American players was as follows: 6.Bg5 f6 7.Be3 Ng6 8.Qd2 Be7 9.0-0-0 0-0 10.Bc4+, etc. — Mason. (Showalter vs A Hodges, 1894)> In a later game, Lasker tried 6.dxe5 dxe5 7.Bg5 and got a big advantage after 7….h6? 8.Bxc6 bxc6 9.Be3 (Lasker vs Steinitz, 1896).

6….Ng6. I’ve already quoted Pollock’s qualified endorsement of 6….f5. <Black’s difficulty is how to dispose of the Kt. In the former games the K’s Kt travelled all over the board. Now Kt3 in this instance, although perhaps preferable, is not a good place either subject as it must be to an early attack from the KRP. — Hoffer.>

7.Qd2 Be7 8.0-0-0 a6 9.Be2 exd4 10.Nxd4 Nxd4 11.Qxd4 Bf6 12.Qd2 Bc6 13.Nd5 0-0 14.g4?!

<Played with Morphy’s dash and inspiration. — Gunsberg. This premature advance is admirably taken advantage of by Steinitz. — Leeds Mercury.> Lasker and Mason wrote that 14.f3 was better; Kasparov added that 14.g3 Re8 15.Bf3 followed by h2-h4-h5 or 14.Nxf6+ Qxf6 15.f3 Bb5 16.c4 Ba4 17.Rde1 were alternative paths to a small advantage.

14….Re8

<Although this looks like a defensive move (to make room for the Kt) it is a subtle design which was entirely overlooked by Lasker. — Hoffer.>

15.g5

15….Bxd5

<Superbull> asked why Black couldn’t play 15….Be7, but …Bxd5 is a much better move. After 15.Be7, White’s attack gets back on track after 16.h4.

16.Qxd5?

Mason called this a <great mistake>, while if 16.exd5 Kasparov thinks Black has <stable compensation> for the exchange after 16….Rxe3 17.fxe3 Bxg5. I wonder if Steinitz would have played …Rxe3; his position is bad otherwise.

16….Re5

<A deep and defensive scheme, which is as beautiful as it is successful. — Gunsberg.> 17.Qd2?

<Much better was 17.Qxb7 Bxg5 18.Bxg5 Rxg5 19.Rhg1, although after 19….Rc5! Black has an easy game; …a6-a5, …Rb8, and then an attack on the kingside pawns by …Qh4, or an attack on the king along the b- and c-files. — Kasparov.>

17….Bxg5 18.f4 Rxe4 19.fxg5 Qe7 20.Rdf1?

The question mark is Kasparov’s. He recommends 20.Bf3 Rxe3 21.Bxb7 Rb8 22.Rhe1 Rxe1 23.Rxe1 Qd7 24.Bd5 (24.Bxa6 Qa4) <with only some advantage to Black.> The 19th century annotators are more positive, e.g. <One of the pawns might be recovered by 20.Bf3, but that would not be enough. His best chance is going on with attack on the King, at all hazards. — Mason. With 20.Rfd1 he commenced one of the most ingenious attacks, so that we cannot but feel grateful for the blunder he made, if such a fine piece of strategy was the result, even though Steinitz somewhat met him half way. — Hoffer.>

20….Rxe3 21.Bc4 Nh8!?

<21….Nh8 on the part of Steinitz is worse than a blunder — it is want of judgment. —Hoffer.> Among the kibitzers here, Gouki wrote <given the complications on the board, indeed 21….Nh8?? immediately removes the knight out of play and gives White a piece up advantage in position. surely, the worst place a knight could be is at the edge of the board, and the fact that a strong master like Steinitz to play this horrible move, shows that his positional understanding is poor compared to Lasker’s!> Llawdogg, on the other hand, calls it a <good defensive move.>

Obviously Steinitz didn’t intend for the knight to stay at h8 forever. Rather, the idea was to get the knight out of its exposed position on g6, then advance the queenside pawns to blot out the white bishop. Meanwhile Black’s rook would come to e8 (instead of the defensive f8), helping Black dominate the e-file. Steinitz presumably expected to drive the bishop off the f1-a6 diagonal, after which a rook would penetrate to e2 or otherwise force simplification. But blotting out the White bishop proves more difficult than Steinitz anticipated.

22.h4 c6 23.g6!

<The only chance: hopeless is 23.Rfg1 d5 24.Bd3 Ng6 25.h5 Nf4 or 23.Bd3 Re8 24.Rhg1 Qe6 25.Kb1 b5 26.h5 c5. — Kasparov. Allowance must of course be made for the fact that, being two Pawns behind, White has nothing to lose and everything to gain by desperate tactics. — Gunsberg.>

click for larger view

23….d5?

This is Steinitz’s first big mistake. As kibitzwc/Fritz12 pointed out, the position is more or less balanced (though still very difficult) after this. Instead 23….hxg6! 24.h5 and now 24…gxh5 25.Rxh5 Re8 (threatening …Re1+) 26.Rhh1 Qe5 or 24….Re8 25.hxg6 Nxg6 26.Bd3 Rxd3! work, as does 23.…hxg6 24.h5 gxh5 25.Rxh5 Re5! (Honza Cervenka).

Kasparov and Mason think that stronger still is 23….hxg6 24.h5 g5 25.h6 gxh6(?), and now if 26.Rxh6 Re8 27.Kd1 (27.Rhh1 Qe5) Qe4 <and the game is decided: 28.Rhf1 Ng6 29.Bd3 Qg4+ 30.Kc1 Qg3 31.Kd1 Nh4, or 28.Bd3 Qg4+ 29.Kc1 Qg3 30.Kd1 R8e6 etc. — Kasparov.> But there is a big hole in this line: 26.Qh2! (instead of Rxh6) 26….Qf8 27.Rf6! Qe7! 28.Rff1 and draw by repetition! So, to keep the game going, Black would have to play 25….d5 26.h7+ (Stockfish coldly dismisses White’s chances after 26.hxg7 Ng6 27.Bd3 Nh4 28.Qd1! Kxg7, but that would be tough for human to face) 26….Kf8 27.Bd3 g4 or the hair-raising 25….g6 26.h7+ Kg7 (27….Kf8? 28.Bxf7) 27.Qh2 f5 28.Qh6+ Kf6.

So 23….hxg6 24.h5 gxh5 25.Rxh5 Re8 or 25….Re5 seem like the best practical defenses.

24.gxh7+ Kxh7 25.Bd3+ Kg8

<25….Kg8 is inferior to 25….g6, though it looks more dangerous. Presumably, White would continue with 26.h5, then 26….Kg7 with comparative security. — Hoffer.>

Hoffer’s recommendation loses instantly after 27.hxg6 hxg6 (27….Nxg6 28.Bxg6) 28.Qh2. A lot of 19th century annotations are that bad.

26.h5

click for larger view

26….Re8

Gunsberg, Pollock, and the Leeds Mercury all recommended 26….Rg3, but 27.Kb1 Qe3? 28.Qb4 is an effective response.

27.h6 g6

28.h7+ Kg7 29.Kb1

<In this game there is something of the “Tal" element: White’s attack is rather abstract, but it will not come to an end — all the time some threats arise! — Kasparov.> But 29.Qh2 Qg5 30.Kb1 is probably a little more accurate.

29…..Qe5 30.a3

<Lasker’s last two quiet moves were completely inexplicable to his contemporaries. How can you play this way when two pawns down?! — Kasparov.>

Here are some comments by actual contemporaries. They seem to understand 29.Kb1 and 30.a3 pretty well. <Exhibiting consummate coolness in a “do or die” predicament. — Pollock. To enable the queen to move away, which is at present impossible on account of mate in three by …Re1+. — Leeds Mercury. 29.Kb1 and 30.a3 are necessary, the former to avoid a subsequent check, and the latter, of course, as a loophole for the King in case of emergency. — Hoffer.>

30….c5

click for larger view

Continuing his plan to blot out the bishop. But f5, here or on the next move, appears to be better.

31.Qf2 (Here is where Kasparov gives his notorious note 31.c3? c4 32.Bc2 Re2!, overlooking that 33.Qh6 is mate. Instead 31.c3 is well answered by 31….d4 32.c4 f5.)

31….c4 32.Qh4 f6

And here is where Llawdogg thinks Steinitz finally went astray, pointing out that Black would win after 32….Kf8 (also recommended by Chigorin, Mason, Pollock, and the Leeds Mercury) 33.Bxg6 Nxg6 34.h8/Q+ Nxh8 35.Qxh8+ Qxh8 36.Rxh8+ Kg7. But as Lasker wrote, the best answer to 32….Kf8 is 33.Bf5! as in the game. If then 33…gxf5 (33….R8e7 loses on the spot after 34.Bxg6 Nxg6 35.Rhg1) 34.Rhg1 f6 35.Rg8+ Ke7 36.Rfg1 Re4 37.Rxe8+ Kxe8 38.Rg6+ Kd7 39.Qh6 and White wins. After 33.Bf5 Re2 34.Bg4 everything remains all topsy-turvy. But after 32….f6, you get the sense that the game is beginning to slip away from Steinitz.

33.Bf5

One hell of a move, though since it is basically forced you can’t really give it an exclamation point.

click for larger view

33….Kf7

34.Rhg1! gxf5

Instead Kasparov gives a monster line that begins 34….b5 35.Bxg6+! Nxg6 36.Qg4 Nh8 37.Qg7+ Ke6 38.Qb7! and then continues 38….f5 39.Rg8 Kd6 40.Rxe8 Qxe8 41.Rxf5 Re5 42.Qxa6+ Kc5 43.Qa7+ Kd6 and here Stockfish thinks 44.Rf6+ Re6 45.Qb6+ Ke5 46.Rxe6+ Qxe6 47.Qb8+ Ke4 48.Qxh8 is crushing.

35.Qh5+ Ke7 36.Rg8

click for larger view

And here is where NeoIndian, and quite a few annotators, thought Steinitz finally blew it with 36…Kd6. Could Black have held after 36….Kd8? 37.Rxf5 Qe6 38.Rxd5+ Kc7 39.Qh2+ Re5 40.Rc5+ Kd6 41.Rxe8 Qxe8 42.Rxc4. NeoIndian says <White cannot make progress, but black is completely tied down.> Stockfish says <+1.53> (or similar), but if you play out some of the lines you wind up seeing stuff like a 3:2 R+P ending with all the pawns on one side of the board. So, probably a draw.

36….Kd7 37.Rxf5 Qe6 38.Rxe8 (or NeoIndian’s line) Qxe8 39.Qg4 Re4 seems to wind up quite drawn.

So, this was probably Black’s last chance to hold the game. Instead, Steinitz sticks to his principles (the King is a strong piece!) and loses.

36….Kd6? 37.Rxf5 Qe6 38.Rxe8 Qxe8 39.Rxf6+!(the King gets away after 39.Rxd5+ Kc6 40.Rc5+ Kb6 — this is probably what Steinitz focused on) 39….Kc5.

Mason and <Inspired by Morphy> preferred 39….Kc7, but Stockfish and Gypsy conclude that 40.Qxd5 wins. Still, the Stockfish variation is very long, intricate and confusing, so 39….Kc7 would have been a better practical chance.

40.Qh6!

click for larger view

40….Re7? Played with the idea of meeting 41.Rf8 with …Rxh7, but as it turns out the King is left defenseless. Chigorin recommended 40….Qe7, but after 41.Rf8 White wins several long, forced engine lines. After 41….Re6 White has the truly diabolical 42.Qd2!! leaving his rook hanging. Now 42….Qxh7 43.Rc8+ Rc6 44.Rxh8 Qf7 (44…..Qxh8 loses to another skewer after 45.Qb4+ and 46.Qc3+) 45.Qb4+ Kd4 46.Rf8 Qd7 47.Rf1 Qd8 48.Rd1+ Ke5 49.Re1+ Kf6 50.Qc3+ Kf7 51.Qh3 Kg7 52.Rh1 — a variation only an engine could love, but it works. If instead of 41….Re6, 41….Re1+ 42.Ka2 Re6 43.Qd2! still wins: 43….Qxh7 44.Rc8+ Rc6 45.Qb4+ Kd4 46.Rxh8 Qxc2 47.Rh4+ Ke5 (47….Kd3 loses to 48.Qe1!) 48.Qe7+ Re6 49.Qc7+ Rd6 50.Rh3! (stopping …Qb3+ and threatening various checks) 50….Qg6 51.Qe7+ Re6 52.Re3+ Kf5 53.Qd7! (would any human find this?) Qg8 (or this?) 54.Qxd5+ Kf6 55.Rf3+ Ke7 56.Qxb7+ <and the curtain comes down (Kasparov) after further incomprehensible engine moves (keypusher)>.

Kasparov says 40….Re2 would have saved Black, but Stockfish disagrees: 41.Qg7 Re7 42.Qg1+ Re3 43.Qg8 Re7 44.Rf8 Qg6 45.Rc8+ Kd6 <with a shaky equilibrium (Kasparov)> <46.Qd8+ wins (Stockfish).> Earlier, Stockfish says, 44.Ka2 puts Black in a short of zugzwang. Damn engines.

41.Qh2? It’s a shame to query such a beautiful move, but 41.Qd2, pace Kasparov, is immediately decisive: 41….Qd8 42.Qb4+ Kd4 43.Rd6! and the threat of 44.Qd2+ is so strong that Black has nothing better than to give up his queen. Meanwhile 41.Qh2 allows 41….Re1+ 42.Ka2 Qe7 averting immediate disaster, though the engine concludes that White eventually wins after 43.Qf4. Other defenses lose more quickly: 41….Rd7 42.Qd2! (again), or 41….Qd8 42.Qf2+ Kb5 43.b3 (Kasparov).

Losing quickest of all is the move Steinitz chose. Lasker’s closeout Qg1 g5+ maneuver is very pretty.

He then goes on to blunder a pawn, and then going all in by sacrificing another. These are usually the situations in which Steinitz shines, and <20.Rdf1??> should have lost the game if Steinitz had played <21...Rf8> instead of <21...Nh8?>, or <23...hxg6> instead of <23...d5??>, but now the game is equal again. Steinitz is two pawns up, but his knight is pretty much out of the game, and Lasker's heavy pieces stare right at his king.

You have to admire both how Lasker manages to keep his attack going -- it looks like it should fall apart any time, like in some other games of this match --, but also how long Steinitz manages to defend against this savage attack. Reading annotations about this game and checking them with Stockfish shows how even the greatest chessplayers since (yep, lookin' at you, Garry) didn't fully understand what was going on here. (Please be assured that I'm not claiming I do either :D I just dumbly follow Stockfish and am amazed how it just is right every single time, even though many details in this game would have required more computation time that I was willing to spend, which is about six hours.)

However, from move 37 on, Steinitz plays one more or less subtle mistake after the other, culminating in blundering his queen with <41...Qd7??>. And once again, Lasker wins a game which he could have lost.

I do get why the chess world back then called him "lucky". He is a phenomenon that even today is hard to understand. Of course, you can always construct some psychological explanation, but that will always remain guesswork.

To me, the games of this match so far show that he had an incredible determination and boldness and fighting spirit. He also was a cunning swindler, playing dubious moves at the right moment to incite his opponent to blunder horribly (although this doesn't show in this game).

Plus, in this match, he was young and had the stamina to keep up the incredible concentration required for these kind of games for much longer than his opponent. Was this something he gambled on? Probably.

Did he go for "impossible" positions on purpose just to destroy Steinitz' philosophy and thus -- because this is Steinitz we speak of -- Steinitz himself? Maybe, although personally I don't think so, at least not on purpose. However, I think this is the effect this game actually had on Steinitz, who lost the following four games in a row.