As keypusher pointed out, the moves given in Tarrasch's match book are more likely correct.

Panigma, to answer your question, if Black defends properly, he can always force a draw. In this particular game, if Lasker had made any other move at certain points in the endgame, Tarrasch would have won. He had to tread very carefully.

Game 14

Played in Munich on September 24-26th

<Translated from Tarrasch's book on the match (http://www.google.com/books?id=0CgC...) with additional commentary from <OMGP I> , Leopold Hoffer's short 1908 book on the match, Shredder and me. Tarrasch's comments are in plain text; all other comments are in brackets. German that I was particularly unsure about is also in brackets, right after the translation. As always, corrections are welcome.>

1. e4 e5 2. Nf3 Nc6 3. Bb5 Nf6 4. 0-0 Nxe4 5. d4 Be7 6. Qe2 Nd6 7. Bxc6 bxc6 8. dxe5 Nb7 9. Nc3 0-0 10. Re1 Nc5 11. Nd4 Ne6 12. Be3 Nxd4 13. Bxd4 c5 14. Be3 d5 15. exd6 Bxd6 16. Rad1

So far as a repetition of the eighth and tenth games. <The opening is discussed in greater detail in Tarrasch vs Lasker, 1908; The rook move seemed stronger to me than Qh5 or Ne4.

16....Qh4

This move I believe to be not quite good. <Dieser Zug halte ich durchaus nicht fur sehr gut.> Black lacks an immediately satisfactory continuation. <Es fehlt dem Schwarzen sogleich an einer gunstigen Fortsetzung.> Better to me seems ...Re8 followed eventually by ...Qf6.

<OMGP I: Correct is 16....Bb7! (16....Qf6 17. Qh5 Rb8 18. Bc1 Qf5 equalises, Capablanca vs Reti, 1914) 17. Bf4 (17. Bxc5? Qg5) 17....Re8 with equality, for example: 18. Qxe8+ Qxe8 19. Rxe8+ Rxe8 20. Bxd6 cxd6 21. f3 Re6 22. Kf2 Bc6 23. Nd5 Kf8 24. Ne3 g6 25. c4 Ke7 26. Rd3 a5 27. Ra3 a4 28. Nd1 g5! 29. Nc3 Rh6 30. Nxa4 Rxh2 when the draw is not far off.>

17. h3 Qb4

Instead of this Lasker recommends ...Bb7 at once. But then Nd5 and c2-c4 would follow. It is precisely one of the main purposes of Rad1 to block the threatening bishop-diagonal with Nd5.

18. Bc1

click for larger view

18....Be6

Lasker writes that 18....Bb7 would have been better, so as to continue after 19. Nd5 with 19....Bxd5 20. Rxd5 Qa4 21. a3 Re8. But on 18....Bb7, there would follow 19. a3 Qb6 20. Nd5 with a good game for White, as he must always stand well if he blocks the bishop on b7 with a knight on d5. Either the bishop is rendered ineffective, or it must take the knight, whereupon Black's attack on the kingside is weakened, while White's attack on the queenside is made easier. The development of the bishop to b7 also makes the following attacking continuation impossible, which Lasker describes as very tempting: 18....Be6 19. Qh5 Rfd8 20. Re4 Qb8 21. Rh4 h6 22. Bxh6 gxh6 23. Qxh6 Be5 24. Re1 Bg7! (not ...c6 because of Rh5 and Rg5) 25. Qh7+ Kf8 26. Rxe6. Now it seems that White must win, for if 26....fxe6, there follows 27. Rf4+ Ke7 28. Qxg7+ Kd6 29. Ne4+ Kd6 30. Qe5+(?) computer etc., except that after 26. Rxe6, Black takes, not the rook, but the pawn on b2, so that if the knight moves, the bishop is protected by the queen. Hence there follows 27. Re3 Bxc3 28. Rg4, when Black saves his seemingly strongly threatened game (28....Bg7? 29. c3) with 28....Be5. Now Black threatens mate, f2-f4 costs a rook after ...Qc1+, and Rg8+ comes to nothing. The entire king attack launched by Qh5 and Re4, though it may well seem worth consideration, is probably not correct, and is rather precisely the sort of attack which, though it seems promising, is not worth investigating, because the resources to implement it are insufficient. <Der ganze mit Dh5 and Te4 eingeleitete Konigsangriff kam also sehr wohl in Betracht, war aber wohl nicht korrekt und vielmehr gerade der Typus eines Angriffs, den man, obwohl er aussichtsvoll erschient, nicht unternehmen soll, weil die vorhandenen Mittel zu einer erfolgreichen Durchfuhrung nicht ausreichen.>

19. a3

On Nd5, there follows ...Qa4 and ...Re8.

19....Qb7 20. Qe4

This move was not made to bring about an ending (because I thought that Black would not permit it, because an endgame seemed so unsatisfactory to him), but rather to displace the black queen and then move the knight to d5. I expected 20....Qc8 21. Nd5 Bf5, and now I could probably play 22. Qc4 without fear of the sacrifice on h3, since after 23. gxh3 Qxh3 24. Nf4 it would prove to be incorrect.

20....Qxe4 21. Nxe4 Rfd8 22. Be3

Now Black is embarrassed by the difficulty of defending his queenside pawns. <Hierauf kommt Schwarz wegen der Schwierigkeit, seinse Damenbauern zu decken, in Verlegenkeit.> The counterattack with 22....Rab8 is not quite sufficient, as White gets, with 23. Bxc5 Rxb2 24. Nxd6 cxd6 25. Bxd6 Rxc2 26. Bc5 and Bxa7, an extra (passed) pawn. Best was probably 22....c4.

22....Bf5 23. Nxc5 Bxc2

Better was first 23....Rab8, although White after 24. b4 Bxc2 25. Rc1 Bf5 26. Nb3 obtains a permanent attack on his opponent's isolated pawns <Weiss nach 24. b2-b4 Lxc2 25. Tc1 Bf5 26. Sb3 immer den Angriff auf die isolierten Bauern des Gegners behauptet>. But after the text move Black gets a material disadvantage.

24. Rd2 Bf5

<Shredder prefers 24....Bxc5 25. Rxc2 Bd6, and Black is still inferior because of his isolated pawns, but material is even.>

25. Nb7 Rd7

On ...Rdb8 the pawn would fall after Nxd6 and Bf4.

26. Red1

click for larger view

26....Be6

Into consideration came 26....a5, with the possibility of making the b-pawn backward after ...a4, except that 27. Bc5 Bxc5 28. Rxd7 Bxd7 29. Rxd7 Bb6 30. Nd8 follows and White wins a pawn and also stands considerably better than he does in the game, for on 30....f6 31. Ne6 follows and on 30....Kf8 White plays 31. Nxf7 Ke8 32. Ne5.

With the bishop move Black sets a trap, i.e. if White plays 27. Bc5, so as to avoid unlike bishops, there follows 27....Bb3 28. Rc1 Bh2+ and Black wins the exchange.

27. Nxd6

<Hoffer: The alternative would be to retain the knight, pacing it at a5, It could be supported with b4, threatening also, at an opportune moment, Nc6; Black thus keeping his two isolated pawns, whilst Bishops of different colour would be avoided.> 27....cxd6 28. Rxd6 Rxd6 29. Rxd6 a5 30. b4

While White has now won a pawn, the opposite-colored bishops give Black, as usual, many drawing chances. However, the presence of rooks makes it possible that the game will be brought to a decision. In addition, the passed pawn advances very rapidly and puts pressure on the black game.

30....axb4 31. axb4 Kf8

Here the game was adjourned for the first time. Black sealed his move.

32. b5

<Hoffer: The King could not be brought into play at once, because of 32. Kf1 Bc4+ 33. Ke1 Ra1+ 34. Kd2 Bf1 35. g3 Bxh3 36. Kc3 and it is worth considering whether the King in play would not be worth the Pawn given up.>

32....Ke7 33. Rd1 Rd8

Naturally Black is eager to exchange rooks.

34. Rb1 Bd5 35. Bg5+

To cause a slight weakness on the Black kingside, because only then can White win. That the move ...f6 entails such a weakness is shown very clearly in the further course of the game.

35....f6 36. Bf4 Bb7 37. Re1+

This and the move following are made so that the king cannot blockade the passed pawn.

37....Kd7 38. Rc1 Ke6

If ...Rc8, Rd1+ follows.

39. b6 Rd7 40. Re1+

Either to drive the king still further from the passed pawn (with ...Kf7) or else to penetrate with the rook. In the former case the passed pawn would have swiftly brought about a decision; there would follow Bc7, Ra1, Ra7 and b6-b7, and nothing could be done about it.

40....Kd5 41. Re8 Kc6 42. Be3 Ba6 43. Ra8 Bd3 44. Rb8 Ba6

click for larger view

45. Kh2

Here White could win the game with h3-h4. The pawn threatens to advance to h6, whereupon 1....gxh6 2. Rf8 Rd6 3. Rf7 would win. (Here one sees how advantageous it was for White to lure the pawn to f6.) If Black responds to h3-h4-h5 by advancing his pawn to h6, all his pawns will stand on black squares, making them objects of attack of the bishop, so that even a bishop sacrifice against the three pawns (via Rf8, Bxh6, Rxf6 etc.) will lead to a win. In addition, this [...h6] gives white the possibility of exchanging rooks (for example, with Ra7) and then playing the bishop to f8. White's plan cannot be thwarted by ...h7-h5, because after Rh8, ...Be2, f2-f3 the pawn cannot be defended. I had the right plan, advancing the h-pawn, in mind, but unfortunately hesitated for some time in implementing it. <Ich hatte auch en richtigen Plan, den h-Bauer vorzustossen, gefasst, glaubte jedoch unglucklicherweise, mit seiner Ausfuhrung noch einige Zeit zogern zu durfen.>

<As with other wins he claimed to have missed in this match (see Games 6 and 9 of the match), Tarrasch gives a rather vague sketch of what he would do, so it's hard to evaluate. It does seem that White gets good wining chances after 45. h4 Bb7 46. h5 h6?!. In addition to the bishop sacrifice Tarrasch mentions, the king also threatens to penetrate on the kingside. But if Black simply lets the white h-pawn advance, his position should remain defensible despite his two isolated pawns, e.g. 45. h4 Bb7 46. h5 Kb5 47. h6 Rf7 48.hxg7 Rxg7 49.g3 Re7 50.f4 f5 51.Rf8 Be4 and I can't see any way for White to make progress. >]

45....Bd3 46. Bf4

With the intention of playing Bc7 and then Rb8-a8, whereupon the advance of the b-pawn wins the exchange. But Black can parry this threat at the last moment, by abandoning the 7th rank. So, the bishop maneuver was useless, and White later retracts it. <Allein Schwarz wurde diese Drohung im letzen Moment parieren, indem er den Turm von der siebenten Reihe wegzoge. Daher ware das Laufermanover nutzlos, und Weiss nimmit seinen Zug wieder zuruck.>

With this move, which White sealed, the game was adjourned for the second time.

46....Bg6

46....Rb7 was a mistake, after which White would win immediately with Rd8, because after ...Bg6 the rook would be permanently locked out of play by Bc7, and Rd7 followed eventually by the exchange of rooks and b6-b7 could not be prevented.

47. Be3 Bd3 48. g4

The right moment for h3-h4 has passed; Black can defend with either ...h5 (if Rh8, then ...Bg6) or ...Bg6 (if g2-g4, then ....h7-h5).

48....Bg6 49. Kg3

Instead 49. Rh8 came into consideration, to prevent ...h7-h5; however, Black would play ...Re7 and force the exchange of rooks with ...Re8.

49....h5 50. f4 hxg4 51. hxg4 Re7 52. Rc8+ Kb7 53. Rc3 Be4 54. Ra3 Kc6 55. Rc3+ Kb7 56. Ra3

Both sides were running out of time here. <The third time control was at move 60.>

56....Kc6 57. Rc3+ Kb7 58. f5 g6?

A mistake that gives White new winning chances.

click for larger view

59. g5!

After this White will obtain a passed pawn on f6, whether Black takes with the f-pawn or the g-pawn.

59....Rf7?

This second error should have led to a quick defeat. The best defense for Black is 59....fxg5 60. f6 Rf7 61. Bd4 Rd7 62. Rc4 Bd5 63. Ra4 Be6, when the exchange of rooks achieves nothing, for the White king cannot enter the black camp via either g4 or c6. Instead of exchanging rooks, White must try to force ...g5-g4, followed by the entry of the king at f4.

60. gxf6?

With 60. Ra3 White would have won immediately, e.g. 60. Ra3 Kc6 61. Ra7 Rxa7 62. bxa7 Kb7 (62....Kd7 63. fxg6 Ke7 64. g7 etc.) 63. gxf6 Bd5 64. fxg6; or 60. Ra3 Kc6 61. Ra7 Rf8 62. Rc7+ Kb5 63. Bc5 R-any 64. gxf6 Bd5 65. fxg6; or 60. Ra3 Kb8 61. Ra7 Bb7 62. fxg6 Rg7 63. Bd4 Rxg6 64. Bxf6; or 60. Ra3 Kc8 61. fxg6 Bxg6 62. Ra8+ Kb7 63. Ra7+ Kc6 64. Rxf7 Bxf7 65. gxf6. It was the 60th move and I had no time for reflection! The continuation actually chosen leads to the win of the bishop, but not the win of the game.

60....Rxf6 61. Rc7+ Ka6

If the king goes to a8, White wins the rook with 62. Ra7+ Kb8 63. Bf4+ Kc8 64. Rc7+ Kd8 65. Bg5; if the king goes to b8, there follows first 62. Re7 and if 62....gxf5 then again 63. Bf4+ Kc8 (63....Ka8? 64. Ra7#) 65. Rc7+ etc.

62. Ra7+ Kb5 63. b7 Bxb7 64. Rxb7+ Kc4

In the following endgame of rook and bishop against rook, for White to win it is necessary to force the black king to the edge of the board, and also to gain the opposition. <In dem folgenden Endspiel von Turm und Laufer gegen Turm gelingt es Weiss zwar, den schwarzen Konig an den Rand zu drangen, aber nicht, die Opposition zu erlangen, die zur Erzwingung des Gewinnes notwendig ist.> Incidentally, in my experience these endings are almost always won by the stronger side.

click for larger view

65. Kf4

With the thought of leaving the black pawn on the board, so as to avoid stalemate. <Um dem Schwarzen zur Vermeidung von Pattgefahr womoglich den Bauern zu lassen.> However, the pawn threatens later to become dangerous, and so must be taken. <Er droht jedoch spater gefahrlich zu werden und muss daher bald geschlagen werden.>

65. gxf6 66. Ke5 Rf8 67. Rc7+ Kd3 68. Bc5 Rd8 69. Bb4 f4 70. Rc3+ Ke2 71. Ke4 Re8+ 72. Kxf4 Re6

Here Lasker invoked the 50-move rule. <Hier verlangte Lasker die Abzahlung von 50 Zugen.>

73. Bc5 Kd2 74. Bd4 Rh6 75. Ke4 Re6+ 76. Kd5 Re7 77. Ra3 Re8 78. Be5 Rg8 79. Ke4 Rg4+ 80. Bd4

Here the game was again adjourned.

80....Re8 81. Rf3 Ke2 82. Rg3 Kd2 84. Be5 Rd8+ 85. Ke4 Kc2 86. Rc3+ Kd2 87. Rb3

click for larger view

87....Rc8

Only after this error does White succeed in driving the king into the corner. Instead of the rook move, Black should have played ...Ke2 (88. Rh3 Rf8).

88. Rh3 Re8

On 88....Kc2, White drives the king to the edge of the board with 89. Rh2+ Kb3 90. Rb2+, because the c-file is off-limits in light of the threatened loss of the rook. But ...Rc2 was much better than the text move, since Black would then prevent the king from being driven to the side of the board, or at least, by the attack with the rook regain the black kings' freedom, e.g. 88....Rc2 89. Bf4+ Ke2 90. Rh2+ Kd1 91. Rh1+ Ke2 92. Be3 Rc4+ 93. Bd4 Rc2, and White can drive the king to the side of the board only temporarily.

89. Rh2+ Kc1 90. Kd4 Rd8+ 91. Kc3 Kd1 92. Bd4 Ke1 93. Kd3 Rf8 94. Rg2

Without the kings being in opposition, the game cannot be won, as noted above. If, as here, the opposing kings stand a knight's move apart, the weaker side can defend successfully.

Here the game was adjourned for the last time.

94....Rf7 95. Be3 Rd7+ 96. Bd4 Rf7 97. Rg5 Rf8 98. Rb5

To drive the king to another place, where perhaps he will stand badly. <Um den Konig anderswohin zu drangen, wo er vielleicht ungunstiger steht.>

98....Kf1 99. Rb1+ Kg2 100. Rg1+ Kh3 101. Be3 Kh4 102. Ke4 Rf7 103. Bd4 Rf8 104. Bg7 Ra8 105. Bf6+ Kh3

Of course not now ...Kh5, when White takes the opposition with Kf5 and gains the winning Philidor position.

106. Be5 Ra4+ 107. Kf5 Ra3 108. Bf4 Ra5+ 109. Ke4 Kh4

White again threatened to gain the opposition with Kf3.

110. Bg3+ Kh5 111. Be5 Kh4 112. Kf5 Ra3 113. Rg2 Rb3 114. Ra2 Rf3+ 115. Bf4 Rb3 116. Rg2 Rb5+ 117. Ke4 Kh5 118. Be5 Kh4 119. Kf5 Rb3

Drawn.

After I had played quite correctly for a long time and obtained a significant advantage, I failed at move 45 to find the winning course; again, I hesitated, as in the 26th and 32nd moves of the first game of the match. Later, after finding the difficult 59th move (g4-g5), I missed the obvious follow-up Ra3, because I found myself in great time pressure. That it was the 60th move <and thus the third time control> must be considered a great misfortune. <Dass ich aber, wahrend ich den verstecken 59. Zug (g4-g5) fand, den auf der Hand liegenden Folgezug Ra3 ausliess, da ich mich wieder in grosster Zeitnot befand und es eben gerade der 60. Zug war, muss doch als direktes Ungluck bezeichnet werden.>

<Hoffer: This game has nothing to recommend itself, except its inordinate length, and as evidence of Dr. Tarrasch's deteriorated form in the match.>

<To me this seems an excellent game by Tarrasch. He got an advantage in a difficult endgame, which he did not managed to convert, because of Lasker's dogged defense. Tarrasch missed a win in time pressure at move 60, but it seems the only reason he got the winning position was because of Lasker's own time-pressure errors.

I didn't check the R+B v. R ending with a tablebase, but my understanding was that it's been done and Tarrasch never had a win.>