Here are some paraphrases and quotes from Soltis' book. Quotes are within carets. This game features some interesting opening play and a very tricky ending, where even the great Emanuel went astray more than once.

After 12....Ba6, White decides to ignore the threat of ...Nxc2 with 13. Bd2.

After 14. Rad1

<Black grabbed the pawn because he had forseen that his knight is not trapped (15. Bc3 Nb4 16. Be2 Rb8) and there are no immediate threats. But White's conception was broader. He wants to force an exchange on d6 so that he can open up an attacking diagonal leading to g7 and h8. Even then his compensation is more visual than concrete -- in other words, it's a positional sack.>

After 15....Nd4.

<The diagonal is safer than it seems -- 15....Bxd6 16. exd6 Nd4 17. Qe3 Bxd3 18. Qxd4 Bxf1 19. Bc3 Qf6!.> But White can transpose into the game with 18. Qxd3.

After 17....Bxd6

<Alekhine said Black's move was forced in view of 17....c5 18. Bh6 Bxd6 19. exd6 R-moves 20. b4, which regains the pawn with advantage. But it's another theme of the Sicilian -- the positional exchange sack -- that Alekhine and Bogoljubov discounted. Black can play 19....Qb6 (20. Bxf8 Rxf8 and ...Qxd6) with at least fair chances.>

After 18. exd6

<White seems to have ample compensation in view of 18....c5 19. b4!, 18...e5 19. Bc3 or 18...Nb5? 19. a4.

But again Black should offer an exchange sac, with 18...Qb6! 19. Bh6 c5. White would have to decide whether to get the pawn back with 19. Be3 c5 20. Bxd4 or make the sack permanent with 19. Bc3 c5 20. b4!? Nb5 21. bxc5 Qxc5 22. Bb2.>

After 19. Rfe1! <[T]his shows that White has evaluated the situation better. His long-term compensation is excellent. First he stops ...Qb6-b5 counterplay by threatening the e-pawn (e.g. 19....Re8 20. Bc3 Qb6? 21. Rxe5!).>

After 20. Bc3

<[The black queen] is vulnerable to f2-f4xe5, winning a piece. White can also work with ideas of Qc4+ and Rxe5, e.g. 20....Qxd6 21. Qc4+ Kh8? 22. Rxe5; 21....Qe6 22. Rxd4; 21....Rf7 22. f4 and 21....Qxd5 22. Qxd5+ cxd5 23. Rxe5.>

After 23....Kf7?

<Black apparently feared 24. g3 followed by White putting his bishop on c3 and his queen on e5. The king-move avoids Qh8+ before it was a threat. But 23....Qd8 was superior.>

Soltis thought Black could have put up stiff resistance with 25....Qa8!.

After 27. Rd4!

<Black, no doubt, counted on 27. Qxc5 Nxc5 28. Rxa7 Ra8 when he has real drawing chances.> But now 27...Qxc4 28. Rxd7+ leads to a hopeless ending.

It is only in the R+P ending that Lasker begins to go wrong. After 36. Kf3? Soltis writes:

<This turns a three-hour game into a four-hour game. After 36. a4! Ra8 37. Rxa8 Rxa8 38. Rc4 and Kf3 Black can call it quits.>

After 40....Rxa8

<White's technical problem has grown (41. Rd2 Ra4!). He wants to give up a pawn for all of Black's pawns (42. Rd5+ Kc6 43. Rxf5 Kd6 44. Rh5 Rxa2 45. Rxh7) but 45...Ke5 is not easy.> Surely it's winning, though?

41...Rxa2? 42. Rg7 Ra4 43 Rg5! would have lost quickly, since both black pawns fall.

After 44....f4 things are very interesting. Black has a number of possible drawing plans, including leaving White with only a rook pawn (1...f3 2. gxf3 Rxf3) or reaching a drawn version of the BP+RP ending. But after 45. Ke4!, 45....f3? 46. gxf3 Rxh2 47. Rc7 is a straightforward win.

After 47. h5, Soltis writes, <Black can try to reach the notorious f- and h-pawn ending with 47....f3 48. gxf3 Rh2 49. Rg5 Ke6. However, White wins with 49. g4! because he can answer 48...Rf1 with 49. Ke3 f2 50. Ke2.> In this line I would expect Black to try 49....Ke6 but I think White still wins after something like 50. Rg6+ Kf7 51. g5 Rh1 52. Rf6+. Comments/improvements?

After 48. Rg4?, Soltis writes, <After 48. h6! White wins easily because he can avoid the notorious ending (48....f3) by means of 47. h7 Rh1 50. Kxf3 or allow a winning form of it (49. gxf3 Rh1 50. Kf5! Rxh6 51. Rg6+).> Continuing Soltis' first line above after 50. Kxf3, it does appear that White would win after 50....Ke6 51. Kf4 Kf6 52. Ra7 Kg6 53. g4 Rxh7 54. Rxh7 Kxh7 55. Kf5 Kg8 56. Kg6.

After 48....Rh1?

<Counter blunder. After 48....f3! 49. gxf3 Black achieves notoriety. Now the game lasts a mere seven hours.>

After 50. Kf5!

<White's king will reach g6, virtually sealing the point. Because the g-file remains closed Black has no vertical checks to stop the winning plan of h5-h6-h7 and Ra5-a8.>

It's worth noting that Black does reach a version of the BP+RP ending -- just not the right one from his perspective.

Soltis' final comment: <No better was 59...Rh2 60. f4 Rxh5 because 61. f5 leads to a won Lucena position after 61....Rh1 62. Ra8+ Kd7 63. Kf7 Rh7+ 64. Kg6 Rh1 65. f6.> Bogoljubov put up a great fight, but resigned a lot quicker than Topa did today. :-)