Most players know how to stop a threat. Far fewer know how to stop a plan with a move that also makes their own position better. That is provention, or proactive prevention, and this game shows it in one quiet move. Oscar Chajes had studied an opening analysis Capablanca had published, decided it was wrong, and played straight into it to prove the point. Capablanca did not know. After the queens came off, he asked two questions: which of my pieces is doing the least, and why did my opponent just move there? His answer, 18.Nd4, began a plan that freed his worst piece and made castling far too costly for Black. The game also shows us something else, because Capablanca's printed notes contain mistakes his play did not. Take one idea from this game: find your worst piece, ask what your opponent wants, and look for the move that deals with both.
When there are no tactics, ask which piece is doing the least and give it a job. Here the rook on a1 could not move until the bishop on c1 did.
Before you answer a move, work out what it wants. Then look for a reply that deals with that idea and also improves your position.
Good judgment is not the same as accurate calculation. In sharp positions, look for three candidates and test your conclusions, because your opponent may already have done so.
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.
After 17...Ne7 there is nothing tactical to find. All of the engine's top moves are within about a quarter of a pawn, so the number will not teach you anything. The thinking will. Capablanca's rook on a1 has no job, because his own bishop on c1 blocks it. That same bishop is also not doing much. So the problem to solve is the bishop. 18.Nd4 looks like a simple knight move, but it heads for b3, and from there it clears the way for 20.Ba3. Now the bishop is active on a long diagonal, the rook on a1 is free, and Black's king has a new problem. That is what separates provention from a passive preventive move. You are not only saying no to your opponent. You are getting something for yourself at the same time. Black was never lost in the opening phase. For about ten more moves the engine has the game close to level. But White's moves are easy to find and Black's are not, and that is how Capablanca wins games. The decisive error does not come until 30...Rh6.
Capablanca reached America on an Argentine transport during the war, landing in Philadelphia in January 1915, and in April he played the New York Masters, finishing with twelve wins and two draws. Some months earlier he had published an analysis of the Open Ruy Lopez, ending at the position after 15.bxc3 with the claim that White had the better game. Chajes and a few other local players studied it and concluded he was wrong. So Chajes played right into it, and Capablanca, in his own words, was "very much surprised at his falling" into the variation, until a spectator told him why. One side note on the road to that position. At move 12 he chose 12.Ne4, and 12.Nb3 was slightly better, though only by a whisker, and 12...d3 was a better try for Black than 12...dxc3. The lesson begins at the very position in dispute.
This is exactly where Capablanca's published analysis stopped, with the claim that White stands better. Chajes and his friends thought Black was at least equal, and in fact better. Material is level and both sides have the bishop pair. White has weak pawns on a2 and c3, while Black's king is still in the center and White's pawn on e5 takes squares away from his pieces. Look at the position yourself and make your own judgment before you read mine.
Capablanca published that White is better here. His opponent studied it and disagreed. Which side do you prefer, and what is the one feature that decides your answer?
Compare the pawns first, then ask what the pawn on e5 does to the black pieces. Think, then reveal.
After 15...Rd8 16.Qxd7+ Bxd7 17.Rd1 Ne7 comes the moment the lesson is built around, and we will look at it below. The plan that follows is worth playing through on the board. 18.Nd4 h6, which Capablanca calls "a bad move", because Black is worried about a pin from Bg5 and does not see that the bishop is heading for a3. To be fair to Black, the engine says this costs very little. It is a misjudgment of what matters more than an error. 19.Nb3 Bb6 20.Ba3, and here is a small point to get right. Castling is not illegal, because the knight on e7 blocks the diagonal. But 20...O-O simply loses the knight to 21.Bxe7, since the king was its only defender. So in practice the king has to stay in the center. Black now finds a clever idea, 20...h5, to bring the rook out by way of h6. 21.Rd3 prepares to double, and the rook on a1, the worst piece on the board a few moves earlier, is about to come to life. 22.Rxd8+ Kxd8 23.Nc5 lands on a square that Black's pawns can no longer guard, and 25.Rd1+ brings that rook in with check. Then 26.Kf1, because in an endgame the king joins the fight.
It is White to move. The queens are off and there is no tactic to find. Black has just played 17...Ne7. This is one of those positions where most players either make a natural developing move or start looking for something clever. Capablanca asks two questions instead. Which of my pieces is doing the least? And why did my opponent put the knight on e7? Answer both of those before you look at a single move.
White to play. What is your worst-placed piece, what does Black want with 17...Ne7, and which move deals with both?
18.Nd4, heading for b3 and freeing the bishop. The worst piece is the rook on a1, shut in by the bishop on c1. Black wants ...Bc6 to trade off White's strong bishop and ease the cramp. Bf4 and a4 are both sensible moves, and if you chose one of them you have not done anything wrong, since the engine has all of these very close. What makes 18.Nd4 special is the plan behind it: 19.Nb3 and 20.Ba3. I want to be accurate about one thing. On the video I said the knight stops ...Bc6, but checking it again, it does not stop it completely. Black can play 18...Bc6 19.Bxc6+ Nxc6, and the engine calls that roughly level. White cannot play 19.Nxc6 instead, because of 19...Rxd1 mate. Look at why that mate exists: the rook on a1 cannot recapture, because the bishop on c1 is in the way. So your worst piece is not only idle, it is a tactical liability, and that is another reason to solve it. Chajes did not play ...Bc6 anyway, and his 18...h6 let the plan go through. Now the thinking, which you can use in your games tomorrow. Your opponent moves. You ask why. You look for a move that takes the idea away, or makes it less attractive, and that also grows your position. That is provention. Most books call it prophylaxis, a word I have never liked, since it only means a preventive measure. Yours should be active.
Black's breakout came with 26...g5 and 27...f5, played, as Capablanca says, "in desperation" to get rid of the pawn on e5. Here is a good test of your candidate moves, because the right answer is easy to miss: 28.exf6, en passant. It is the only move that keeps the advantage. Without it, the bishop has to run all the way to a8 and Black is a little better. After 28...Rxf6 29.Re1 Be6 30.Bf3 Black played the error of the game, 30...Rh6. Now 31.Bxh5+ wins the pawn with check, because the rook is overworked. It guards the pawn on h5 and the bishop on e6 at the same time, so 31...Rxh5 fails to 32.Rxe6. Black should have tried 30...Ng6, which also loses the h-pawn but keeps the game going.
After 31...Kd7 32.Bg4 Nc6 comes a neat capture. 33.Rxe6 works because the king cannot take back: the bishop on g4 covers e6. The rook can, with 33...Rxe6, but then it stands pinned against the king, and Capablanca uses the time to play 34.Bc1, hitting the pawn on g5. After 35.Bxe6 Kxe6 36.Bxg5 White is two pawns up with three connected passed pawns on the kingside, and Capablanca wrote that Black "might have resigned". Watch how he wins from there, because this is where many of you go wrong. He does not send one pawn up the board on its own. The pawns walk together with no gaps, the king comes to e4 and d5, and his pieces support them. After 48.f5 Black resigned, with nothing left to stop the pawns. In the endgame you have two plans: attack the weak pawns, and push your passed pawns. Capablanca did both.
The principle: when there are no tactics, improve the piece that is doing the least. In this game: the rook on a1 was stuck behind the bishop on c1, and 18.Nd4 began to solve it. In your games: if you have no plan, start with your worst piece.
The principle: every move has an idea behind it, so find it before you answer. In this game: 17...Ne7 prepared ...Bc6, and Capablanca saw it. In your games: after each enemy move, ask what it wants before you look at your own moves.
The principle: the best way to stop an idea also makes your position better. In this game: the route Nd4, Nb3 and Ba3 freed the rook and made castling cost a knight. In your games: look for the move that does both jobs.
The principle: list all your captures as candidates, because the one you forget can be the only good move. In this game: 28.exf6 was the only move that kept the advantage. In your games: when a pawn jumps two squares beside yours, stop and look.
The principle: connected pawns are strongest side by side, with your king and pieces in support. In this game: the f-, g- and h-pawns advanced as a group while the king went to d5. In your games: do not run with one pawn alone. Bring the others and bring the king.
Provention is my word for proactive prevention. You work out what your opponent wants to do, and you take it away with a move that also improves your own position. It is close to what most books call prophylaxis, but prophylaxis only means a preventive measure, and it can sound passive. In this game Capablanca saw that 17...Ne7 prepared ...Bc6. He answered with 18.Nd4, a move that began the plan of Nb3 and Ba3, freed his rook, and made castling very costly for Black. The test is simple: after your preventive move, is your own position better than before?
Go through your pieces one at a time and ask what each one is doing right now. Which has no moves, no targets, or no job? Here the answer was the rook on a1. It could not move along the back rank because the bishop on c1 blocked it, and after 18.Nd4 Bc6 the same problem meant White could not safely win a piece, because of mate on d1. Once you know your worst piece, the plan often follows. Capablanca did not need to find a brilliant idea. He only needed to move the knight so the bishop could leave c1.
In a way, both. At the position where the analysis stopped, the engine has White ahead by about a third of a pawn, so Capablanca's claim holds, but only modestly. It does not support the idea that Black was better, as Chajes thought, but Black's position was fully playable and the preparation was sound. Capablanca wrote that he was never in danger of anything worse than a draw, and I agree, though I think that says more about his strength than his analysis. His win came from outplaying his opponent over the board, not from the opening being refuted.
A few. He says Black could not castle at move 17, because of 18.Be3, but the line continues 20...Rd3 21.fxe3 Rxc3 22.Bd5 c6 and runs to dead level. At move 18 he calls 19...Bg4 the only satisfactory answer to Nb3, but the engine prefers 19...Be6, which I also point out on the video, and 19...Bg4 fails to 20.Rxd8+ Rxd8 21.h3. In the 19...Bf5 line, the immediate 20.Bg5 is stronger than the exchange on d8 first. Most clearly, after 29...Re6 he recommends 30.Bg6+, and there the knight simply takes with 30...Nxg6. The other move he offers in that note, a pawn to the fourth rank, is excellent, and with a bishop on a3 it can only mean 30.h4. None of this hurts his play. At the board his calculation was far better than in his notes.
Stop trusting your first judgment without checking it. Capablanca had the best feel for a position of anyone, yet he did not look for his own flaws, and he assumed his opponent knew nothing about his analysis. My view, as I explain on the video, is that habits like this are what later cost him, because Alekhine studied his games and notes and looked for positions that demanded precision. You are an adult improver and you do not have Capablanca's gift, so you need the process: find three candidates, rank them good, better, best, and calculate. Hard work beats talent if talent refuses to work hard. Put the process into your games and your results will follow.
Six quiet moves, rising in difficulty. In each one the winning first move is not a check or a capture, the kind of move that is easiest to overlook. Ask what each piece could be doing, and find the move that changes the position. Drag the pieces; the board answers for the defense.
Asking why your opponent moved there, and finding the quiet move that answers it, is a habit that can be taught. In a 2-hour Strategy Assessment I go through your games personally, see how you really make decisions, and build you a roadmap for your chess.
Book Your $130 Strategy Assessment 2 hours via Zoom · Worldwide · Limited spots