|KEG: I agree with Mateo that Marco's 34. Rg3? was a losing blunder. But the game--which was of considerable interest for the first 33 moves--was probably already lost for Marco by then.|
The opening revisited a line of the Berlin Defense that had been explored by Janowski in his game with Tchigorin at London 1899 and that was to be tried again (with greater success for White) in Janowski's loss to Schlechter later in the Paris 1900 tournament.
The line features (after 8...Nb7) 9. Nc3 0-0 10. Re1 Re8.
Tchigorin played 11. Qc4, a move recommended by Alapin. Tchigorin lost the game, but not because of 11. Qc4, which Schlechter was to demonstrate is a very viable line.
Marco's 11. Be3, though less forcing than 11. Qc4, gives White a small advantage, especially after Janowski's 12...Bd7 (12...Nc5 was better).
With 15. Qa6, Maroc signaled his willingness to trade Queens and reduce to an approximately even endgame. (15. c4 would have been better as was noted by Rosenthal in the Tournament Book).
After 15. Qa6 Nd8 16. QxQ Janowski pondered his recapture in the following position:
click for larger view
Marco recommended the "careful" 16...BxQ. Rosenthal in the Tournament Book says that 16...BxQ leads to an advantage for White and prefers the more active 16...RxQ even though this costs Black a pawn (the a7 pawn). Not surprisingly, Janowski opted for the more aggressive 16...RxQ, and Marco promptly snatched a pawn with 17. Bxa7.
The battle lines were now drawn. Marco had his pawn and Janwoski messed up Marco's King-side pawn structure with 19...BxN and then began his center pawn advance with 21...c5 and 22...d4. The game was very much in the balance.
Marco calls his 28. Kg4 "embarrassing," but the move--though inferior to 28. h4 as Marco notes, doesn't look like a fatal blunder to me. The problem with the move--especially after Marco's poor 29. h4 (29. Kf5 was better-- is that it allowed Janowski to get in a powerful shot with 29...h5+. Had Marco taken the pawn with 30. Kxh5? he would have been blown away with 30...Rh8+ 31. Kg4 Rxh4+ 32. Kg3 Nf4 33. Rh1 Rah6!.
After Janwoski's 29...h5+, Marco should have played 30. Kf5. His 31. Kxh4 could have been the losing move, except that Janowski here erred with 31...Kh6 (paradoxically, 31...Kg6 was best). But Marco faltered again with 32. Rg1 (32. Re4--and not Marco's 32. Rad1 was best), and now Janowski played the crushing 32...d3!!
After Marco's 33. Nc3 (33. Kg3 or 33. Rad1 were relatively better), Janowski could have played 33...d2!! His 33...Nd4 was not as good, and with 34. Kg3 Marco might have had a sliver of a chance.
In fact, Marco blundered with 34. Rg3 ? walking into a Knight fork with 34...Nf5+.
I agree with dernier loup de T that 37. Rd1 would have been better for Marco than his 37. Kg2, but having lost the exchange and with Janowski's powerful attack and control of the board, the game was gone by this stage anyway.
All in all, a fine effort by Janowski despite a few lapses.