White can't deal with both the threats, 29...Qxh3+ and 29...g4.

Black wins with 30..Ng5, below.

click for larger view

The threat is 31...Nf3# so white canot take the f rook.

Assuming 31 Be2, black wins material after 31...gxf3.

click for larger view

29.exf5 Qxh3+ 30.Kxh3 Rh6+ 31.Kg4 Rh4#

I spent a lot of time trying to make 28...Qh3+ work and not succeeding.

I then gave up as I had previously dismissed 28...g5 thinking it just fails to 29.Kh1. I should have looked further, as 29...Qxh3+ 30.Rh2 Qxf3+ 31.Rgg2 leaves B a couple of pawns up with an attack (e.g. 31...g4 to come).

Durciansky came, left that one comment, then left chessgames forever. He probably had a stunning 93-move forcing mate worked out in years and checked with the latest engines and supercomputers, but unfortunately, it was too large for the margin. Now, we'll never know why move number 6 is first mistake. (Hell, we won't even know if it's White's or Black's move!)

Personally, I'd say move number 6 is second mistake (please pronounce in pseudo-Japanese: miss-tar-keh), because move number 3 is first mistake, specifically White's move, because I somehow find the Three Knights the most boring of all openings.

However, there's a *lot* of inaccuracies on both sides (according to Stockfish, and according to my rather strict definition of that rather evasive term "accuracy"), but as usual with Lasker, *his* inaccuracies turn out to give him an actually pretty pleasant game. After move 20, Stockfish says the game is completely equal, but to me, Lasker has a quite clear advantage of having the initiative and more space, while is opponent has barely enough room to maneuver his pieces and cover all his weaknesses.

His rooks have the two back ranks to themselves and can suppress any attempt of White counterplay on the queenside, or start some play there of his own (which he wisely doesn't do). His knight on h5 is anything but dim, because there's the constant threat of <...Ng3>.

Of course, being able to convert this advantage into is what separates people like Lasker from people like me. Like a vulture, Lasker just circles over his victim, waiting for it to make a slight misstep step, and then immediately goes for the kill. (Probably more like a spider, actually, Lasker doesn't wait for his prey to die first, as any half-decent chess player he does like to kill.)

So Alapin is completely right, Janowski should have stayed nicely passive and be content to maybe get away with a draw, especially considering he was in time trouble. (Only the likes of Ding Liren can pull that one off, and there's not many likes of Ding Liren.) As soon as he plays the unfortunate <24.c4>, Lasker's knight pricks up his ears, winnies a war cry, and gallopps to e6, and now threatening <...Ng5>, targeting both f3 and h3.

Then Lasker does a few more moves that confuse the audience (and Stockfish too), but Lasker clearly already had the queen sacrifice in mind and was sure enough that Janowski wouldn't find <28.c5! bxc5 29.e5! dxe5 30.dxe5 Rf8 31.Bc4>, finally managing to free his pieces a little; but even so, Lasker would still have been better, so he must have been overly worried.

However, poor Janowski didn't see the devious trap, and after <28.Bd3??> all is lost (<- a great movie with an elderly Robert Redford and not a single word of dialogue except one that I am not allowed to say here because it would corrupt your soul and that of your children) after the deadly advance of the g-pawn and the even more deadly threat of <...Qh3>. I can imagine Janowski's shock when he finally sees that. He plays on a bit, but Lasker soon blows his position to dust, and he resigns in time. A plausible continuation would be <33.Qxe2 fxg4 34.e5 Nxh3 35.Rxh3 Qxh3+ 36.Kg1 Qxc3!>, and Black is up a rook and a pawn (and counting).