Having looked at so many of Alekhine's annotated brilliancies, I am starting to see a pattern. Inaccurate opening play by Alekhine's opponent followed by a less than obvious move or two by Alekhine to hinder development, and suddenly the opponent is in real trouble.
The current game follows the pattern. Black's problem has been that, when the Knight was on f6, he was unable to play ...Nxd5 because White would recapture with gain of a tempo.
Bradley Beach 1929
[FEN "r1b2rk1/1pqn1ppp/p2b4/2pBn1B1/4P3/2N2N1P/PP2QPP1/R4RK1 w - - 0 14"]
Alekhine played 14.Nh4! and commented,
In view of Black's cramped position the right policy is to avoid exchanges. Besides, Black is now forced to prevent the move 15.Nf5 and consequently has even less choice than before.
The game continued 14...Nb6 15.f4 Nc6. Now Alekhine played the surprising 16.f5!. Alekhine:
A paradoxical, but most effective, continuation of the attack, by which White "sacrifices" the central square e5. The "natural" advance 16.e5 instead would have left White -- strange as it may seem -- after 16...Be7 with but an insignificant positional advantage.
16...Ne5 17.Qh5 Re8.
Parrying the threat 18.f6 which now would be met by 18...g6 19.Qh6 Bf8.
18.Rf4 Be7. Alekhine:
This will be refuted by a pretty combination, but, as Black still could not take the powerful Bishop -- after 18...Nxd5? follows 19.Nxd5 Qc6 20.Nf6+ gxf6 21.Bxf6 etc. -- there was no longer a sufficient defense.
Because of Black's last move White is enabled to effect this advance in spite of the possible defence 19...g6 20.Qh6 Bf8 -- and this because of the following combination: 19...g6 20.Nxg6!.
He then gave a series of variations showing that White wins the Queen or checkmates Black.
After the following retreat which permits the opening of the f-file, the game is also practically over.
19...Bf8 20.fxg7 Bxg7 21.Raf1 Be6 22.Nf5 Bxd5 23.Nxg7! and Black resigned a few moves later. To play through the complete game see...
Alexander Alekhine vs Herman Steiner, Bradley Beach 1929