<10...f6> is a favourite Steinitz move, <f5> is a very playable alternative, see Von Popiel vs Schlechter, 1898

Steinitz makes a bad mistake with <44...Nc7?> when <44...Nxc3> would have maintained a dynamic equality. <45.Nxc3> d5 46.Qa6 (46.exd5 Qa3+ 47.Kd2 Bxd5 48.Nxd5 Rxd5 49.Ke1 exd4 50.exd4=) 46...Rd6 47.Qa5 Qd7 48.Rdd2 exd4=

With <45.d5> Lasker puts Steinitz's white squared bishop out of play and badly cramps Black's game. Steinitz finds it difficult to manoeuvre to cover threats on the both sides of the board.

<49..Rdd8?> show that Steinitz did not appreciate the developing threats on the 6th rank against <c6> and <f6>. After <49...Rb7> he has a difficult game, but he could have put up better resistance.

<51...Bh7?> shows Steinitz was still oblivious to the sacrificial undermining of his position which Lasker unleashes with <52.Rxf6!!>. This move diverts the Black Queen from her role protecting the Queen-side. Lasker sees that after his Queen pins Steinitz's rook from <c6> there is no answer to his Knight heading for <c7>.

<54.Qc6+!!> is also deadly as Black runs painfully out of moves <54...Rb7> 55.Nc5 dxc5 56.Qxf6 Ne8 57.Qc6 Nc7 58.d6 Ne6 59.Nd5 a5 60.Nc7+ Nxc7 61.dxc7

If <57...Ka8> Lasker has a deadly dance of his Knights to win by <58.Qc6> Bg6 59.Nc7+ Kb8 60.Na6+ Ka8 61.N4c5

I realize STeinitz had his own way of conceiving of the game, but even for amateurs if you play over and over you will see that whatever compensation you have in the game, usually you must continue to pursue that strategy to its logical end. Like if I have material and you have an attack, I have to exchange to blunt your attack. I dont keep grabbing material! Or space vs tempo, or development vs material or whatever. You dont keep adding more material to a material advantage, you look to see if your opponent is attacking your K. Or you dont keep piling up positional advantages, you look to post Rooks on empty files, or to post pieces where the missing pawn cant hit them.

THe pt. is whenever the first weakness is created the other guy usually gets some compensation if he is halfway competent. For steinitz to trade pieces again to create another positional weakness just doesnt make sense to me. The game just doesnt work like that.

Some beginner books will tell you 6....Bxe3 is a bad move, because it gives White an open KB file. Steinitz deals with that problem by playing ...Ng4, ...f6, ...Nh6 and ...Nf7. But then he plays ...g6 to keep Lasker's knight out of f5, and so the KB file turns out to be a weakness after all, which Lasker finally exploits on move 52(!). It looks like <chessical>'s (and Schlechter's) 10...f5 would have been a better way of solving the KB file problem. But Steinitz probably didn't want to do that because it would have meant giving up one of his "small advantages" -- Lasker's doubled e-pawns.

It's a very interesting game. It's Steinitz rather than Lasker who makes the important decisions in the opening -- ...Bxe3 and ...Nxb3 -- that determine the positional parameters of the struggle. Neither side has much after the opening. It's my vague impression that Steinitz plays too passively between moves 20 and 40. They are both doing a lot of tacking, but Lasker seems to be trying to make things happen and Steinitz is just responding. After 40 moves Steinitz has a permanent weakness at f6 and Lasker has traded off one of his b-pawns for Steinitz's b-pawn, leaving Black with an isolated pawn at a7. Steinitz doesn't seem to have accomplished anything. But I don't see any holes in <chessical>'s suggested alternative for Black at move 44, so it isn't like Steinitz played really badly before the end. He just didn't seem to give himself any chance to win.

Hired killer. But sometimes I work <pro bono>, so watch yourself.

Lasker, however, did not claim the forfeit and waited for Steinitz to feel well enough to finish the game.

Source:

Muskovsky Vedomosty, Nov. 14 1896

I have argued that the difference between the 1894 and 1896 matches was not that Steinitz got weaker but that Lasker got stronger. This game and game 17 from 1894 (Lasker vs Steinitz, 1894) are an intriguing direct test, since in both of them Lasker plays a stodgy Giuoco Piano -- something he did just once in each match -- and Steinitz reacts similarly, trading bishops on e3, trading knight for bishop on b3, and reinforcing his center with ...f6. In 1894 Steinitz won one of his finest games of that match, while in 1896 Lasker breaks through and wins after a long struggle.

Unfortunately, I don't think these games help prove my case. Lasker plays the opening here more confidently and aggressively than he did in 1894, and emerges with a small advantage rather than a small disadvantage. But then follows a long and rather boring maneuvering phase, in which Lasker tacks back and forth and Steinitz stays behind his lines, attempting nothing. The crisis of the game comes around move 40, and both Lasker and Steinitz make one mistake after another. Finally Steinitz's 44th move (as Chessical pointed out years ago) all but forces his opponent to grab a decisive positional advantage with 45.d5. The play isn't stellar after that, either. Even Lasker's pretty winning combination (52.Rxf6) features a large oversight, though it doesn't affect the final result.

Steinitz is actually the first to vary from the 1894 game, playing 6...Bxe3 rather than 6...Bb6 as in 1894. In one of my first chess books ...Bb6 was recommended, on the ground that Black doesn¡¦t want to open the f-file for his opponent. I don¡¦t think that is much of a concern (even though Lasker does manage to exploit the f-file at long last) -- even at deep search depths SF prefers ...Bxe3. After 9...Ng4 Steinitz is aiming at the same formation in the previous game. Note, though, that White can just play 10.h3 and the knight has to retreat, since it would be lost after 10...Nxe3 11.Qe2. Anyway, Lasker plays 10.Qe2 f6 11.d4 c6 12.0-0-0, which is the first major divergence from the earlier game. Steinitz responds sensibly with 12...Qe7, then 13.h3 Nh6 14.g4. Lasker seems to be playing the opening more directly and confidently than in the earlier game. Here SF recommends 14...b5 15.g5!? fxg5 16.dxe5 dxe5 17.Nxb5 cxb5 18.Qxb5+ Bd7 19.Rxd7 Qxd5 20.Qxe5+ Qe7 (20¡KKf8 21.Nxg5 also ends in a draw) 21.Qb5+ with a perpetual.

Not much of a guide to what Lasker or Steinitz is likely to play, but good to know, I guess. Anyway, after the more Steinitzian 14...Bd7, SF thinks White is just a little better (+0.50). Certainly the opening has gone better for him than it did in the 1894 game. Lasker¡¦s 15.Nh4 g6 16.Nf3 echoed the last game of the ¡¥94 match and presumably thoroughly disgusted Isidor Gunsberg (see Lasker vs Steinitz, 1894)

16...Nf7 17.Rdg1

click for larger view

An interesting moment. SF thinks the aggressive 17...b5 18.d5 0-0 with attacks on opposite wings is equal, but Steinitz plays the more circumspect 17...0-0-0. Lasker pushes forward with 18.b4?! (18.d5 ⩲), allowing 18...exd4 19.exd4 d5! 20.exd5 Qxb4 and if White wins a pawn with 21.dxc6 Bxc6 22.Qe6+ Rd7 23.Qxf6, Black can equalize with 23...Rhd8 24.e5 Bxd5 25.Nxd5 Rxd5 26.Qxf7 Ra5 27.Qe6+ Kb8 28.Kb1 Qa4 29.c3 Qa1+ 30.Kc2 Qa4+, etc. Again, you really can¡¦t expect Lasker or Steinitz to see all that, but I thought it was interesting.

19.Qf2

Lasker plays d4-d5 on move 45; SF wants Lasker to play d4-d5 on seemingly every move until he finally does it. I¡¦m not going to post variations on each move, but here on an overnight search SF¡¦s main line runs 19.d5 c5 20.b5 Ng5 21.Nd2 h5 22.gxh5 Rxh5 23.h4 Rdh8 24.Qe1 Nh3 25.Rxg6 Be8 26.Rg4 Bd7 27.Rh3 Rxh4 28.Rg8+ Rxg8 29.Qxh4 Ng5

click for larger view

It feels like White doesn¡¦t have very much in the final position. Lasker presumably thinks that he can leave the center as it is, since Steinitz can¡¦t undertake anything there (...exd4 should be in White¡¦s favor, and at least for now ...d6-d5 is unplayable), meanwhile White can maneuver and wait for a more favorable opportunity.

19¡K.Rdf8 20.Qg3?! h6?!
(Instead Black can strike with 20...h5! and if 21.gxh5 g5 followed by ...Rxh5 or 22.h4 g4 23.Nd2 Rxh5 and he is slightly better. 21.Nd2 g5 22.Qf3 hxg4 23.hxg4 Rh4 is also slightly better for him.)

21.Rf1 Nd8 22.Rhg1 Ne6 23.Rf2 Nc7 (23¡Kg5 to be followed by ...h5 is good here, with a small advantage for Black) 24.Rgf1 (note that White is preventing ...d5) 24...Rfg8 25.Nh4 Ne8 26.b5!

click for larger view

26...Ka8?!

Instead 26¡Kh5 27.bxc6 bxc6 28.Nf3 g5 is more or less equal. Now White gets a small advantage (+0.55, 38 ply).

27.bxc6 bxc6 (as in the 1894 game, it takes a long time for a pair of pawns to leave the board) 28.Nf3 g5

Leaving himself with a permanent weakness -- a backward pawn on an open file. But he will follow up with ...h5 and ...hxg4, getting the h-file for his rooks, and making the white pawn on g4 a target. 28...g5 is also SF¡¦s choice. (And maybe Steinitz was just sick of White¡¦s knight coming to h4¡K)

click for larger view

33...Rb8
A more active defense is 33...hxg4 34.hxg4 Nb5 35.Qd2 Rh1 36.Rxh1 (or 36.Nxb5 Rxa1 37.Nc7+ Kb7 39.Kxa1 Kxc7 40.Qa5+ Kb8 41.Qb4+ Ka8 42.dxe5 c5 43.exd6 Qxd6 44.Qd2 Qxd2 45.Nxd2 Rh3 and Black should be able to hold the ending) 36...Rxh1 37.d5 c5 38.Nxb5 Bxb5 39.Qa5 Qb7 40.Qd8+ Qb8 41.Qxf6 c4! and according to the engine, Black¡Çs counterattack is strong enough that he will either get a perpetual or force White to seek one. Again, though, it¡Çs asking a lot of a human player to see that. After the text move White¡Çs advantage is still small.

34.Qe2 Rb7 35.Rgg1Rh8 36.Rad1

36.d5 Nb5 37.Nxb5 cxb5 38.Ne1 hxg4 39.hxg4 Rc8 is rated at dead equal by SF. I suspect Lasker thought 36.Nd2 d5 would equalize for Black. But after 37.Qf3! hxg4 38.hxg4 Be6 39.Rh1 Rbb8 40.Rxh8 Rxh8 41.dxe5 fxe5 42.exd5 cxd5 43.Qg3! (+1.16, 49 ply, 218 minutes) 43...Bd7 44.e4 d4 45.Nd5 Nxd5 46.exd5 White has the advantage. But I can¡Çt really query Lasker¡Çs decision, because I think these lines are practically incalculable.

36...hxg4 37.hxg4 Rc8

It looks a bit odd to move the rook off the newly opened h-file to this cul-de-sac, but presumably Steinitz plans to move his knight to b5.

38.Qd3

click for larger view

38...Be6?

The first really clear error by either side, and not a very big one. 38...Nb5 seems only slightly better for White, and Black can also try 38...exd4, since 39.exd4 can be answered with 39...Ne6!, and the knight lands on f4, and if 39.Nxd4 d5 40.exd5 Nxd5 41.Nxc6 Bxc6 42.Nxd5 Bxd5 43.Qxd5 Qxe3 44.Qd3 and Black will have decent chances to hold the double rook ending.

39.Nd2?!

White seems to get an advantage via 39.d5 cxd5 40.exd5 Bd7 41.Nd2 Rb4 42.e4

click for larger view

Not easy, but white can work on Black¡Çs slightly exposed king, the weakness on f6, and even the h-file. But White is still better after the text move.

Now both sides play quite badly over the next half-dozen moves, as Lasker repeatedly gives Steinitz chances for counterplay, and Steinitz repeatedly fails to take advantage.

40.Ndb1?

This odd retreat allows Black a couple of fairly straightforward paths to equality:

40...c5 41.dxc5 (41.d5 Bd7 42.Nxb5 Bxb5 43.c4 Bd7 with a rock solid position) 41….Nxc3 42.Nxc3 dxc5 43.Nd5 Bxd5 44.Qxd5 (44.exd5? c4 is worse) 44….c4 45.Rgf1 Qc7 46.Rxf6 cxb3 47.cxb3 Qc2+ 48.Ka3 Rb8 (threatening 49...Qxb3+) 49.Rd3 Qc2+ 50.Ka2 Qc2+ =

40...Nxc3 41.Qxc3 (41.Nxc3 c5 transposes to the previous variation) 41...c5 42.d5 Bd7=

Instead Na4, as in the game a move later, keeps the advantage.

40...Rbc7? 41.Na4 Rb7 42.Kc1?

Another misstep. 42.Nd2 keeps an edge.

42...Rcb8?

Instead 42...exd4 43.exd4 c5 44.d5 c4! 45.bxc4 Bd7 46.Nac3 (46.cxb5? Bxb5 47.Qa3 Qxe4 with a crushing counterattack) 46...Nxc3 47.Qxc3 (47.Nxc3? Rbc7 ∓) 47...Qxe4 and Black is slightly better.

43.Rg2? (43.Kb2) Rd7? (44...exd4 and 45...c5) 44.Nbc3?! (44.d5)

click for larger view

44...Nc7??

The move is that bad. It’s SF’s 10th choice, and the evaluation jumps from about 0.5 to over 2.0 when it is played. But forget silicon for the moment: this move practically begs White to push his d-pawn, after which:

-the bishop will be driven offside, so that c6 will be a gaping wound for the rest of the game, and g4 won’t be a target. -the rook is rendered useless, at least temporarily -- why move it to d7 at all if you’re not going to play ...d5? -the knight loses its outpost on b5 and is mostly just in the way. -the queen cuts a miserable figure.
-White even gets rid of his doubled pawn!

It’s funny, in the 1894 game, inducing White to play d4-d5 solidified Black’s advantage. Was Steinitz somehow hypnotized by that memory into thinking that forcing Lasker to play d4-d5 would be good for him here?

As Chessical pointed out long ago, 44...Nxc3 45.Nxc3 d5 is much better, but SF thinks that White still keeps some advantage after the very surprising 46.Kb2. Black can also try 44...d5 immediately, or 44...exd4 45.exd4 Rdb7. It goes without saying that any of those alternatives, and a number of other moves, are much better than the move Steinitz played.

45.d5 cxd5 46.exd5 Bg8 47.Qc4 Rc8 48.Kb2 (another powerful alternative here is 48.Na2!, aiming for c6) 48...Rb8

Here, as I said, according to a manuscript book of the match reprinted by Winter (the annotations were apparently taken from <Deutsches Wochenschach>, annotations by Tarrasch), Steinitz went over the time limit, but Lasker “renounced his rights” and won over the board. JFQ, citing a Russian newspaper, gives a slightly different story, that Steinitz sealed 47...Rc8 and then failed to show up for the adjournment, but Lasker did not claim the forfeit, and waited for Steinitz to feel well enough to finish. JFQ’s version seems more plausible.

49.e4 (still winning, but much stronger is 49.Qc6+ Rb8 50.Rh2 Qg7 51.Ne4 Ne8 52.Rh8!! Rd8 (52...Qxh8 53.Qc8+ Rb8 54.Nb6+ axb6 55.Qxd7 and mates) 53.Nac5! dxc5 54.Nxc5 Rd7 55.d6! Qxh8 56.Nxb7 Rxb7 57.d7)

49...Rdd8? (as Chessical points out, 49...Rb7 is better) 50.Rf2

50...Rf8 51.Rdf1 Bh7

Also a mistake, but it’s quite hopeless by now.

click for larger view

52.Rxf6 Rxf6 53.Rxf6 Qxf6 54.Qxc7

Chessical and <Deutsches Wochenschach>/Tarrasch both pointed out that 54.Qc6+ Rb7 55.Nc5 was better. The magazine/Tarrasch’s curious comment is “Instead of this, White could win the queen by 54.Qc6+ Rb7 55.Nc5, but he quite fittingly disdains this continuation.” What’s wrong with it?

54...Qh8

Instead of this Steinitz could have dragged the game out with 54...Qd8. After 55.Qxh7 White is winning, of course, but I suspect Lasker would have been left wishing he hadn’t disdained winning Steinitz’s queen.

55.Qc6+ Rb7 56.Nb5 Kb8 (or 56...Qd8 57.Nxd6 Qc7 58.Nxb7 Qxb7 59.Qe6 a5 60.Nc5 and wins) 57.Qxd6+ Kc8 (57...Ka8 58.Qc6 and Chessical pointed out the win) 58.Qc6+.