A knight planted deep in the enemy position, on a square the defender can barely challenge, is one of the most powerful pieces in chess. This game is a clinic in creating one and cashing it in. Out of a quiet Ruy Lopez, Capablanca saw that the square f5 could become a near-permanent home for a knight, supported by his own e4 pawn. To reach it he played the bold g4, pushing a pawn in front of his own king to drive Black's bishop back, and then dropped a knight onto f5, where it dominated for fifteen moves and strangled Black's game. Black could only remove it by giving up his good light-squared bishop, and when Marshall finally did, Capablanca took the two bishops, a pawn wedge and the open h-file, and broke through. Take one idea from this game: find the strong square your knight can own, and build your whole plan around putting it there.
Look for an advanced square your knight can occupy where the enemy can barely challenge it, ideally defended by one of your own pawns. Here it was f5, guarded by the e4 pawn.
A knight on a dominant square is worth a long journey. Capablanca played Nf1, Ne3 and then Nf5, and pushed g4 first to drive the bishop back.
A dominant knight cramps the whole position. Keep improving while the opponent suffers, and when they trade it off, collect the bishop pair and the open lines it leaves behind.
Watch the lesson, then play the game through on the board below. The board stays with you as you scroll. Make each move yourself while you read. You'll remember a plan you played with your own hands far longer than one you only watched.
By move 10 the position is a standard Ruy Lopez with level material and no weaknesses on either side, but Capablanca has seen the one idea that matters: the square f5. It is defended by his e4 pawn, and the only pawn that could ever attack it, by ...g6, would wreck Black's kingside, so a knight there is as good as permanent. He had already begun the journey with Nf1 and Ne3, and now he commits with the bold 11.g4, pushing the pawn in front of his own king to drive the bishop back from h5, and 12.Nf5. The knight lands and never has to move; for the next fifteen moves it cramps Black so badly that Marshall's queen does not leave its home square until the game is almost over. Black can only be rid of it by trading his good light-squared bishop, and when he finally plays 27...Bxf5, Capablanca recaptures 28.gxf5, keeping the two bishops, a pawn wedge on f5 and a wide-open h-file. From there he doubles rooks on the h-file and breaks through. One square, found early, decided the whole game.
Capablanca met Marshall's Steinitz Defense with the modest 4.c3 and 5.d3 rather than the sharp 4.d4, admitting the slower setup was not the most theoretical try but choosing it on purpose: against a tactician like Marshall he wanted a quiet, maneuvering game he could win on understanding. Black developed naturally but without a clear plan, and the early Bg4 only invited White to gain time later. By move 10 White had a pleasant space edge and, more importantly, a target square. The lesson begins there, with Capablanca about to turn a small positional plus into a lasting one.
After 10...Bh5 material is level and nothing is forcing, but one side has a clear way to improve and the other does not. Look at the square f5: White's e4 pawn guards it, and the only pawn that could ever challenge it, ...g6, would badly weaken Black's king. That makes f5 a dream square for a knight. Ask yourself how White gets a knight there and what it would be worth, before you read on.
Material is level and the position is quiet. What is White's plan, and why is the square f5 the key to the whole game?
A knight on f5 would dominate, and Black could barely challenge it. Work out the route to f5 and what White plays to drive the bishop back first, and ask what Black can ever do about a knight that lands there. Think, then reveal.
Capablanca's plan is all about one square. He has already played Nf1 and Ne3 to reroute the knight, and now 11.g4 drives the bishop back and gains space, so that 12.Nf5 installs the knight on its dominant post. The point is not a quick attack but a lasting cramp: the knight on f5 hits g7 and e7, controls key squares, and can be challenged only by the weakening ...g6 or by trading Black's good bishop for it. For the next fifteen moves Capablanca improves slowly behind it, bringing his king to g2, lifting a rook to h4 and doubling on the h-file, while Marshall shuffles and his queen never moves. When Marshall finally plays 27...Bxf5 to trade the monster off, 28.gxf5 leaves White with the bishop pair, a pawn wedge jammed on f5 and a fully open h-file pointing at the black king. The knight did its job long before it left the board, and the pieces it left behind finished the game. The method is worth copying: find the square, route the knight, and let a dominant piece do the work that no single tactic could.
Capablanca has just played 11.g4, driving the bishop back to g6, and it is White to move. Everything in the plan has been building to this one move, the move that turns a good square into a dominating piece. The e4 pawn guards f5, and the only pawn that could challenge it, ...g6, would wreck Black's king. Find the move that makes the whole plan real.
White to play. How does Capablanca convert all his preparation into a lasting advantage?
12.Nf5, the knight reaches its dominant square. This is the payoff for everything that came before. The knight lands on f5 defended by the e4 pawn, and because the only pawn that could attack it, ...g6, would ruin Black's king, it will sit there for as long as Capablanca wants, hitting g7 and e7 and cramping Black's whole position. Marshall can realistically remove it only by giving up his light-squared bishop, which is exactly what eventually happens, and when it does White is left with the bishop pair and open lines. Notice that the knight is not threatening an immediate tactic; its value is positional and lasting, which is precisely why it is so strong. In your own games, look for these squares early, in the center and in the enemy camp, defended by your pawns and hard for enemy pawns to reach, and route a knight to them. A knight on a dominant square is often worth more than a bishop and does its damage move after move, not in a single blow.
After living with the knight for fifteen moves, Marshall finally plays 27...Bxf5 and Capablanca recaptures 28.gxf5. Look at what the knight has left behind: White has the two bishops against bishop and knight, a pawn wedge jammed on f5 that cramps Black further, and a completely open h-file with a rook already on h4 and another ready to double. Black's king is stuck on the kingside with no defenders nearby, and his queen and knight are tangled on the back rank. The material is still level, but the position is winning, and every feature of it traces back to the square f5.
Capablanca doubles on the h-file and brings the queen to the kingside, and now the attack breaks through: 30.Qg6 ties Black up, and 31.Rxh6 is a rook sacrifice that rips the king open, 31...gxh6 32.Bxh6+. Here is the point worth noticing: the whole winning position was built without a single sacrifice, purely by positional means, and only once White was clearly winning did Capablanca give up material to force the end. After a forced sequence the two bishops and rooks overwhelm the defense, 36.Bf8 pins the queen to the king, and after 37.Rxh7+ and 38.Rxa7 Marshall resigned. From a single quiet knight move to a mating attack, the plan was positional first and tactical only at the finish.
The rule: look for advanced squares your knight can occupy where enemy pawns can barely challenge it; a knight there is a dream piece. In this game: f5, guarded by e4, was Capablanca's target from move 10. In your games: scan for these squares before you look for tactics.
The rule: it is worth several moves to bring a knight to a strong square, because once there it dominates. In this game: Nf1, Ne3 and g4 all served the single idea of Nf5. In your games: give a knight a long journey to reach a great square.
The rule: a committal pawn push is justified when it secures a key square and the center is closed enough to keep your king safe. In this game: g4 looked risky but drove the bishop back and prepared the dominant knight. In your games: weigh the square you gain against the shelter you give up.
The rule: a dominating piece cramps the opponent for many moves, and even trading it off leaves you better. In this game: the knight was swapped on move 27, but it had already won the game. In your games: value the long squeeze, not just the immediate threat.
The rule: when the opponent trades off your dominant piece, take the lasting gains, the bishop pair, a pawn wedge, an open file. In this game: 28.gxf5 handed White all three. In your games: recapture toward the advantage, not just the material.
It is an advanced square, deep in Black's half, that is defended by White's e4 pawn and that Black can barely challenge. The only pawn that could attack a knight on f5 is the g-pawn by ...g6, but that move would tear open Black's king and was never playable here, so the knight sits unmolested. A knight on such a square controls key points like g7 and e7 and cramps the whole enemy position. The opponent usually has to give up a bishop just to remove it, which is exactly what Marshall eventually had to do.
To drive the bishop back from h5 and grab kingside space before planting the knight. The black bishop on h5 was eyeing the kingside, and g4 kicked it to g6, a worse diagonal, while gaining room. Pushing a pawn in front of your own king is normally dangerous, but here the center was closed and Black had no way to open lines quickly, so White's king stayed safe. The space and the dominant knight that followed were worth far more than the slight loosening.
Because its value is positional rather than tactical. A knight on f5 controls key squares like g7 and e7, cramps the opponent's pieces, and denies the defender any freedom, move after move, for as long as it stays. That constant pressure is worth more than a one-move threat. The opponent is slowly squeezed until something gives, which in this game meant Marshall's pieces becoming so tangled that his queen never moved until the position was already lost.
Because the knight left a lasting legacy. When Marshall played Bxf5, Capablanca recaptured with the g-pawn, and that single recapture gave him three advantages at once: the bishop pair, a pawn wedge on f5 that kept Black cramped, and a fully open h-file aimed at the black king. The knight had already done its work by dominating for fifteen moves, and the features it left behind were more than enough to win. The lesson is to recapture toward your advantage.
Not until the very end. The entire winning position was built by positional means, with no sacrifice at all, through the dominant knight, the bishop pair and the open h-file. Only once he was clearly winning did Capablanca play the rook sacrifice 31.Rxh6 to rip open the king and force the finish. That is the model to copy: earn a winning position quietly first, and spend material only when it cleanly converts. The real work of the game was done by the knight on f5, not by any combination.
Six kingside attacks, the kind that crashes through once your pieces dominate, rising in difficulty. In each one the attacking side has the superior force aimed at the king, just as Capablanca did once his knight had done its work; find the way through and play it out. Drag the pieces; the board answers for the defence.
Knowing how to spot a dominant square, route a knight to it, and convert the squeeze it creates is the kind of positional understanding that wins games long before any tactic appears. In a 2-hour Strategy Assessment I go through your games personally and build you a roadmap for your chess, not a textbook student's.
Book Your $130 Strategy Assessment 2 hours via Zoom · Worldwide · Limited spots