His Finest Endgame: How Capablanca Beat Nimzowitsch on Two Wings

Good day, chess world.

Aron Nimzowitsch wrote the most famous line ever written about passed pawns. A passed pawn, he said, is like a thief in the night: mere observation will not do, you must keep it under lock and key. In Riga in 1913 he sat down across the board from a young Capablanca and watched, helpless, as that exact idea was turned against him. Capablanca later called the endgame that followed one of the finest he ever played, and even Nimzowitsch tipped his hat.
If you believe opposite-coloured bishop endgames are always dead draws, this is the game that will change your mind. A pawn up in one of these is usually nothing, because the defender simply walls off your passed pawn on the squares your bishop can never touch. Capablanca's answer is the one method that cracks the fortress: make passed pawns on both wings, too far apart for a single defender to stop both. Watch the full deep dive lesson below, then follow the three turning points here.
https://www.youtube.com/watch?v=Y2yuC2EoOdI
Nimzowitsch versus Capablanca, Riga 1913, from an Italian Game. The result is 0-1, and the whole lesson lives in the ending: rooks and bishops of opposite colour, a tiny structural edge, patiently grown into passed pawns on both sides of the board until Nimzowitsch's pieces could not cover them all.

Why This Game Matters

Here is what I want you to understand. Opposite-coloured bishop endings are the hardest of all to win with an extra pawn, harder even than rook endings, because each bishop rules one colour completely and the defender can build a fortress on the squares your bishop cannot reach. Most players see the word "drawish" and give up trying. Capablanca does not. He shows the single technique that beats the fortress: do not push one passed pawn, push two, one on each wing, and place them so far apart that the enemy king and bishop cannot guard both at once. This is the game that teaches you how to win the ending everyone else agrees to draw.

The Opening

I do not dwell on the early moves. Out of an Italian Game Capablanca played a bold, almost reckless-looking line, giving up the two bishops and advancing his kingside pawns while his king still sat in the centre. The spectators, he later wrote, looked at one another. But every move had a point, and when the queens came off Capablanca held the better structure: two pawn islands to Nimzowitsch's three, more active rooks, and a king ready to march into the fight first. If you want the move by move detail, the video covers it. From here it is a technical masterpiece.

Key Position 1: A Tiny Edge, and a Big Plan (21.g4)

At a glance this looks close to equal, but it is not, and the pawns tell you why. Nimzowitsch has three separate pawn groups: a lonely a-pawn, a pair of doubled and isolated c-pawns, and his kingside majority. Capablanca has just two tidy groups. Nimzowitsch has just played g4, trying to start a kingside attack, but from a position that is already worse, that push only fixes his own pawns as future targets. That is one of the soundest rules this game teaches: when you stand worse in an endgame, be very reluctant to push pawns, because each push tends to leave a fresh weakness behind. Capablanca's plan is already forming, and it is the plan for the whole game: work on both wings at once, turn those scattered white pawns into targets, and manufacture a passed pawn on each side of the board.

Do Not Push Pawns When You Are Worse

This is the idea I most want you to keep, because Nimzowitsch, of all people, keeps breaking it. A player who is worse wants to sit tight and defend, keeping his pawns flexible and giving the stronger side nothing to attack. Instead Nimzowitsch pushes g4, then later a4, planting pawns on light squares, exactly the colour Capablanca's bishop controls and the colour no white pawn can defend. Later he plays c3, a mistake of a different kind, because it only speeds Black's own passed pawn on the b-file. Every push either hands Capablanca a fresh, fixed target or shortens the road for a black passer. Good defence is patience. Capablanca simply lets his opponent weaken himself, fixes each new weakness in place, and only then moves in to collect.

Key Position 2: Fix the Target First (31...a5)

After a spell of rook trades down the open central files, the rooks are gone and we have reached the pure opposite-bishop ending, the graveyard of extra pawns. Watch how Capablanca prepares. Nimzowitsch has pushed his pawn to a4, a light square, and rather than rush to win it Capablanca plays a5, nailing it in place. Now the a4 pawn cannot move and cannot be defended by another pawn; it is a permanent target sitting on Capablanca's bishop colour, and it will fall in good time. This is strategy over calculation. You do not have to see every variation, you only have to recognise a fixed weakness and know it cannot run away. From here the black king and bishop close in, win the a-pawn, and the first passed pawn is born on the queenside.

Two Wings, One Bishop

Here is the heart of why opposite bishops can still be won. A single passed pawn, however strong, can be blockaded: put a piece on the square in front of it and it is frozen. But the defender has only one king and one bishop. Create a second passed pawn far away on the other wing, and now the same two pieces must somehow guard both, and they cannot. So once his queenside passer is fixed and White's pieces are tied to it, Capablanca opens a second front on the kingside, rolling his own g- and h-pawns forward. The further apart the two passed pawns sit, the more hopeless the defence. In these endings it is not the material that matters, it is the distance between your passed pawns.

Key Position 3: The Breakthrough (45...Bxc4+)

Now the payoff, and it is spectacular. Capablanca gives up his bishop with a check, Bxc4, and Nimzowitsch cannot take it: after Kxc4 the a-pawn runs with a2, and while White busies himself stopping it the kingside pawns crash through. So White declines the bishop and grabs the a-pawn instead, but that changes nothing, because Capablanca simply manufactures a new queenside passer, this time on the b-file, and then a second one on the kingside with another sacrifice, g3. The finish is a study in the theme: he ends up with a passer on the b-file and another on the h-file, an entire board apart. Nimzowitsch's bishop stops one; the other queens. He resigned, and paid the compliment Capablanca treasured for the rest of his life. See the finish in the video: https://www.youtube.com/watch?v=Y2yuC2EoOdI

The Full Game

Play through the whole game here:

The Modern Take

An engine confirms the verdict and sharpens the lessons. Capablanca's technique is close to flawless, and the machine's main comments fall on the loser. Time after time it flags Nimzowitsch's pawn pushes, g4, a4, c3, as the moves that turned a difficult defence into a lost one, each one weakening his structure a little further, most of them on the very colour Capablanca's bishop rules. And it points to one big decision: on move 27 Nimzowitsch had to recapture, and he chose Bxf6, stepping into the opposite-bishop ending, rather than Rxf3, which would have kept the rooks on the board. Modern analysis agrees the rook ending offered the better practical chance, because a rook defends by making active counterplay, which suits the side that is worse. It is a lovely irony that the great theorist of defence and prophylaxis was undone by neglecting his own rules. The lesson for your own games is twofold: when you are worse, stop pushing pawns and sit tight; and when you are better in an opposite-bishop ending, do not settle for the draw, build a second passed pawn far from the first and let the fortress fall.
Which idea will you use first: refuse to push pawns from a worse position, fix the enemy's weakness before you win it, or make two passed pawns on opposite wings? And for the analysts, could Nimzowitsch have held the rook ending after Rxf3? Tell me in the comments.

Key takeaways

When you are worse, do not push pawns. A defender should keep his pawns flexible and give no targets. Nimzowitsch pushed g4, a4, and c3, and each advance either fixed a new weakness on Capablanca's bishop colour or sped a black passed pawn, while costing time he could not spare.

Fix the target before you win it. Do not rush to grab a weak pawn; nail it down first. Capablanca's a5 froze White's a-pawn as a permanent target, then took it at leisure. Strategy over calculation.

Two passed pawns beat the opposite-bishop fortress. One passer can be blockaded; two on opposite wings cannot, because the defender has only one king and one bishop. The further apart they are, the more the defence collapses.

This is Game 18 in my series on Capablanca's best endings. I am building a clean, fully annotated database of every game in my own words and coaching notes, lesson by lesson. Follow the series for new breakdowns and updates.
What did you find most instructive? Let me know in the comments.
Watch the full breakdown on the ChessExcellence channel: https://www.youtube.com/watch?v=Y2yuC2EoOdI
Watch the whole series here: https://www.youtube.com/playlist?list=PLZkwv5s1SbCANPY49gUl4Ht92K5tD8KEq