The Time Alekhine Sac'd the Whole House

Alekhine was only 19 when this game was played in St. Petersburg in 1912, fifteen years before he became World Champion.
His opponent was no lightweight either. Grigory Levenfish would later become a two-time Soviet Champion and even hold Botvinnik to a drawn match.
Yet in this game, Alekhine made him look completely helpless. By move 15, he was ready to sacrifice basically everything he owned.

Quick Note

If you play the London System, I've also created London System: A Compact Repertoire for White, a downloadable PGN repertoire covering the main setups against ...d5 and ...Nf6 + ...e6.
You can check it out here: https://payhip.com/b/XBDVa or through the link in my profile if you're interested.
Now, onto Alekhine sacrificing virtually everything he owns.

The Game

Why Did This Actually Work?

Looking at the final material count, Alekhine's attack almost seems quite surreal.
Black had taken both of White's rooks. White's king had voluntarily walked to f2. Black's queen was sitting deep inside White's position on h1.
And despite all of that, Black was completely lost.
The key to understanding the combination is that Alekhine wasn't simply throwing pieces at the king and hoping something would work.
The pawn on e6 was an absolute monster.
It prevented Black's king from escaping to f7, interfered with the development of the light-squared bishop, restricted Black's pieces, and eventually advanced to d7, where it helped create the decisive threats.
Black's opening had also left them one crucial tempo short.
The bishop never reached g7, which meant that once Nc7+ arrived, the king couldn't retreat to f8. Instead, it was forced onto the d-file, exactly where Alekhine's queen and pieces wanted it.
That is what makes 15.Nb5!! so remarkable.
At first glance, Alekhine seems to be ignoring the fact that his rook is attacked while simultaneously putting the knight somewhere it can simply be captured.
In reality, the material no longer played a role.
The knight's route to c7, the e6-pawn's control of key squares, and Black's undeveloped king were worth far more than either rook.

The Safest King on the Board?

There is also something quite ironic about 16.Kf2.
White has just sacrificed a rook, deliberately allows Black to take another one, and then moves the king onto a square that looks may seem a bit exposed.
Yet by the end of the game, White's king is remarkably safe.
Black's queen on h1 is nothing more than a spectator. It doesn't have any useful checks, while White's f4-bishop quietly helps defend the king by controlling important squares around it, especially h2.
Meanwhile, Black's own king is the one trapped in the middle of the board with almost nowhere to breathe.

Final Thoughts

This is a great example of why material alone tells you very little about an attacking position.
Alekhine didn't sacrifice two rooks because he was playing recklessly. He perfectly felt the whole nuance of the position: in particular, Black's king had become so restricted that the rooks were almost irrelevant.
The e6-pawn deserves enormous credit. It looked like just an advanced pawn, but it effectively imprisoned Black's king, paralysed the light-squared bishop, supported White's attack and eventually helped deliver the final blow.
And after 15.Nb5!! Qxa1+ 16.Kf2 Qxh1 17.Nc7+, the position becomes almost comical.
Black has collected nearly everything Alekhine offered.
The only problem is that Black's king is about to get checkmated.
Sometimes, apparently, the correct way to defend a rook is to offer your opponent a knight on top of that.