Fifth Game

Played in Munich on September 1

<Translated from Tarrasch's book on the match (http://www.google.com/books?id=0CgC...) with additional commentary from Soltis' <Why Lasker Matters>, Kasparov's OMGP I, Shredder, and me. Tarrasch's comments are in plain text; all other comments are in brackets.

More than a week had passed since the last game, and the venue had changed from Dusseldorf to Munich. But judging from how he played here, defeat in Game 4 still weighed heavily on Tarrasch.>

1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. 0-0 Be7 6. Re1 b5 7. Bb3 d6 8. c3 Na5 9. Bc2 c5 10. d4 Qc7 11. Nbd2 Nc6 12. h3 0-0 13. Nf1(?!) cxd4 14. cxd4 Nxd4 15. Nxd4 exd4

<Soltis: As Keres later showed in similar positions, Black's piece play compensates for his slightly inferior center (e.g. 16. Bb3 Qb6 17. Bg5 Ra7 18. Qd3 Nd7).>

16. Bg5(?!)

Up to this point the game is identical to the third of the match <see Lasker vs Tarrasch, 1908, where the opening is discussed in greater detail>. Bg5 is much stronger than the move Lasker played there, Ng3. The important difference is that the black king's bishop cannot, as in the third game, reach f6 so as to at least temporarily support the pawn on d4.

<OMGP I: I should remind you that the challenger played the fifth game in a depressed state after his heavy defeat in the fourth. Otherwise would easily have found the strong response 16....Qc5!, found by analysts only after the match: 17. Bh4 Be6 17. Rc1 Qb4 18. b3 Rac8 20. Ng3 d5 with advantage to Black. Which, in my view, after 21. Bxf6 Bxf6 22. exd5 Bxd5 23. Bxh7+ Kxh7 24. Qh5+ Kg8 25. Qxd5 Rxc1 26. Rxc1 Qd2 is not really so great, though it is clear that that the limit of White's dreams is a draw.>

<In his post-game note Tarrasch says that he anticipated 16. Bg5 but he and an unidentified master concluded that 16....h6 was an adequate reply. >

16....h6

Better is 16....Nd5, when 17. Bxe7 Nxe7 follows. White then attacks the pawn on d4 with knight, queen and rook, while Black can defend it only twice; moreover, its fall opens the way to an attack on the isolated and weak pawn on d6. The acceptance of the pawn sacrifice is thus not recommended, so the entire defensive system fails in my view. <Die Annahme des Bauernopfers ist also fur Schwarz nicht empfehlenswert, damit aber fallt meines Erachtens das ganze Verteidigungssystem zusammen.>

<Shredder is much more optimistic, rating Black as slightly better after 18. Nd2 Be6 19. Nf3 Nc6 20. Bb3 Bxb3 21. Qxb3 Rac8. His position certainly doesn't seem bad.>

17. Bh4 Qb6

Better was ...Be6 and ...Rc8. But in any event White would regain the offered pawn with Nf1-d2-f3 (or Nb3) and stand better.

18. Qd3

click for larger view

18....g5?

Very risky and not even necessary; against the threat of e4-e5 followed by Bxf6 and Qh7#, there is a much simpler and safer defense, namely a rook move, making room for the king, i.e. 18....Re8 19. e5 dxe5 20. Bxf6 Bxf6 21. Qh7+ Kf8 and White has nothing more than a check, whereupon he must retreat. But after 18....Re8 White could have played the better 19. f4(?) Be6 20. Bf2 or else 19. f4 Be6 20. b3 and Bf2. <Shredder: 19. f4 is a mistake, allowing 19....Nh5! 20. Bxe7? (better 20. e5) 20....Nxf4 followed by ...d3+, but 19. Rad1 Be6 20. Bb1 Rac8 21. Qxd4 is about equal.>

19. Bg3 Be6 20. Rad1(?!)

<Soltis: 20. e5 dxe5 21. Bxe5 Rfd8 22. Bxf6 Bxf6 23. Qh7+ Kf8 24. Qxh6+ Ke7? 25. Ng3! and Nf5+ wins, but 20....Bc4 21. Qf5 Be6 22. Qd3 Bc4 repeats.>

20....Rc8

<Soltis: 20....Bxa2 is downright ugly -- yet it's strong. It can't be refuted by the tempting 21. b3 Rfc8 22. e5 because Black keeps his extra pawn after 22....dxe5 23. Bxe5 Rc3 24. Bxd4? Rxd3 25. Bxb6 Rxd1 26. Rxd1 Rc8.

Instead, White should maintain his attack with 24. Qf5! Then 24....Bxb3? loses to 25. Bxb3 Rxb3 26. Bxd4 because the rooks are forked after 26....Qd6 27. Bxf6 and 28. Qxd5!

White isn't losing after the best defense, 24....Rd8 27. Bxf6 Qxf6 28. Rxe7! (28....Qxe7?? 29. Qh7+ mates). But neither is he winning (28....Qxf5! 29. Bxf5 Bxb3).>

21. Bb1

Here Black can no longer stop the pawn, and he is also menaced with a king-side attack via 22. e5 dxe5 23. Bxe5 followed by Bxf6, Qh7+ and Qxh6+.

21....Nd7(?)

So as to continue 22. Qxd4 Qxd4 23. Rxd4 Ne5, but above all to defend against the mate threat on h7 from f8.

<Soltis: The position remains "somewhere on the cusp of equality" after 21....Rc6 and then 22. e5 dxe5 23. Bxe5 Bc4. There isn't much for White in 24. Qxd4 Bc5 25. Qd2 (Khalifman/Soloviov) or 24. Qf3 Bd5 25. Qd3 with another draw by repetition.>

22. e5 Nf8 23. Qf3(!)

Lasker brings his queen into the attack on the black kingside and rolls it up completely in a few powerful moves <rollt ihn in wenigen kraftvollen Zugen vollen auf>. Naturally, regaining the pawn is much weaker, since now a higher goal beckons.

click for larger view

23....d5

An immediate 23....Kg7 is not better, because after 24. Qh5 then Black cannot play ...Ng6 because of 25. exd6 Bxd6 26. Rxe6 fxe6 Qxg6+, and after other moves also, e.g. ...Re8 (to defend the bishop on e6 from this combination) White always breaks the black kingside with f2-f4, for example 23....Kg7 24. Qh5 Re8 25. f4 dxe5 26. fxg5 Bxg5 (26....hxg5 27. Be5+ Bf6 28. Qxg5+) 27. Bxe5+ Bf6 (27....Kg8 28. h4 and Qxh6) 28. Bxf6+ and Qxh6+, and White, after also winning the d-pawn, will have the far better position and an extra pawn; or 23....Kg7 24. Qh5 Re8 25. f4 f5 26. exd6 and White wins a pawn and has a decisive attack, e.g. 26....Bxd6 27. fxg5, or 26....Bf6 27. fxg5 Bxg5 (27.....hxg5 28. Be5) 28. Be5+ Kh7 29. h4 Bd8 30. Bxf5+ Bxf5 31. Qxf5+ and Rd3 wins.

24. Qh5 Kg7 25. f4(!?)

<Bold, seemingly risky and certainly "complexifying" the game -- now both players will have to examine the possibility of ...d3+ on almost every move. I think Lasker had sized up his opponent well: Tarrasch immediately falls to pieces. Shredder slightly prefers 25. Nh2, but appreciates 25. f4 as well.>

25....f5(?)

Now the Black king-formation will be robbed of all its pawn-guards in a few moves; better was 25....Ng6!. Apparently this move cannot be played because of the fork f4-f5, but then would come first ...d4-d3+!, followed by an involved combination, the outcome of which for White would be not at all certain, namely 25....Ng6! 26. f5 d3+ 27. Bf2 Bc5 28. Ne3! Nf4 29. Qg4(?) d4! 30. Nf1 Bd5, and the position is at least as critical for White as for Black.

<Instead of 29. Qg4, Soltis thinks White is "probably" winning after 29. f6+ Kh7 30. Bxd3+ Nxd3 31. Rxd3 with threats of Nf5 and Ng4. Shredder finds a Tal-like shot in Soltis' line: 31. Nf5!!

click for larger view

...and despite loose White pieces all over the board, Black is defenseless. Better for Black is 30....Ng6, but he's still lost after 31. b4! Bf8 (31....Bxb4 32. Ng4! Bxg4 33. Bxg6+ fxg6 34. Qxg4 wins) 32. h4.>

But White can get off the path to these doubtful complications if, after 26....Ng6, he simply defends his f-pawn with 26. Qf3 <Allen diesen zweifelhaften Verwicklungen konnte Weiss jedoch aud dem Wege gehen, wenn er auf 25....Sg6 einfach seinen angriffenen f-Bauer mit 26. Qf3 deckte>, when, whether Black continues with 26....gxf4 27. Bxf4 Nxf4 28. Qxf4 or not, his broken king's wing remains under the threat of unending attacks <immer bleib sein aufgerissener Konigsflugel von nachhaltigen Angriffen bedroht>.

<Soltis (and Shredder) think that Black can bring the attacks to an end in Tarrasch's line after 28....d3+ 29. Kh1 Bg5 30. Qg3 Qd4 31. Rxd3 Qh4.>

<Black's other alternative to 25....Ng6 is the immediate 25....d3+. Soltis says 25....d3+ succeeds after 26. Kh2 Ng6 27. fxg5 Bxg5 28. Bxd3 Nf4, but not after 26. Bf2!, for instance, 26....Bc5 27. Bxc5 Qxc5+ 28. Kh2 gives White an "overwhelming initiative" (28....Ng6 29. fxg5 or 28....gxf4 29. Qh4! Ng6 30. Qf6+ Kh7 31. Bxd3).>

26. exf6+ Bxf6 27. fxg5 hxg5 28. Be5!

Thus White wins the g-pawn and the game. Lasker has conducted the attack logically and powerfully.

click for larger view

28....d3+ 29. Kh1 Ng6 30. Qxg5

<Soltis: here 30. Bxf6+! Kxf6 31. Rxd3 and Rf3+ would win faster. For example, 31....Nh4 32. Qh6+ Kf7 33. Rxd5!, or 31....Nf4 32. Qh6+ Kf7 33. Rf3 Rg8 and the quiet 34. g3 or 34. Re5 wins.>

30....Bf7 31. Ng3 Bxe5 32. Rxe5 Rh8

Black strives in vain to defend his king position. In addition to the attack on the king, the two connected passed pawns secure victory for White.

33. Bxd3 Ra7 34. Rde1

Threatening Re6.

34....Kf8

Here the game was adjourned, with White sealing the following move.

35. Bxg6 Nxg6 36. Qe3 <36. Qf4 is even stronger> 36....Rc7 37. Nf5

click for larger view

37.....Qc6

38. Qa3+ Kg8 39. Ne7+ was threatened. Best was ...b4.

38. Qg5 Resigns.

This game was played flawlessly by my opponent. The move 16. Bg5, which brought White the advantage in all variations <der Weiss immer in Vorteil bringt>, I had already recognized as the strongest immediately after the third game, but I had concluded that the game was not a forced loss, and that (as a master had informed me) Black could always respond with 16....h6 <hatte aber doch geglaubt, die Partie damit noch nicht verlieren zu mussen, da mich ein Meister damals belehrte, dass man dagegen immer noch h7-h6(?) ziehen konne.> But the move was not good, and I immediately got into difficulties; to escape them, I entered into new difficulties with ...g5. After I failed to play 25....Ng6, my position became completely untenable. This is the first game of the match in which I was outplayed from the outset.

<I wonder who the master was who told him Black was OK after 16....h6. The book makes clear that both Tarrasch and Lasker had seconds who assisted in making match arrangements, but did Tarrasch have a chess second as well? If so, blaming one's seconds for defeat, as Petrosian and Korchnoi tended to do, clearly has a long history. One thing I like about Topalov is that he always credits Cheparinov for his own opening successes, and never (as far as I know) blames him for his failures.

Turning to the game, Soltis and Shredder remain happy with Black's position much longer than Tarrasch (or Lasker) does. Though Tarrasch thinks Black is worse off throughout, the modern GM and the engine believe he starts to go astray only with 21....Nd7. It's interesting that neither Tarrasch nor Lasker so much as mentioned the possibility of 20....Bxa2 in his notes. In any case, Tarrasch, Lasker, Soltis and Shredder agree on one thing: after 25....f5?, it's hopeless.

It's worth pausing over that classic Lasker move, 25. f4!?, though neither Tarrasch, Soltis nor Lasker himself say anything about it. You could call it psychological, I suppose, but even an engine likes it! I do think it had practical virtues beyond a computer's ken: it's a hellishly difficult move to answer in time pressure, which Tarrasch was probably experiencing (the time control was at move 30, remember). Also, by that point in the match I think Lasker might well have concluded that his nerves were stronger and his tactical eye sharper than his opponent's. To the extent Lasker thought 25. f4 and, say, 25. Nh2 were equally good, these sorts of considerations might have helped him make up his mind. >