Day 28of 28, Capablanca's Chess Career

How to Stop Blundering: Always Find the Third Candidate

Capablanca vs R. T. Black · Rice Memorial, New York 1916 · The Third Candidate

You have felt this. You find a move that looks great, you pick up the piece, you let go, and in that instant your brain says: I just blundered. Most of the time it happens because you stopped looking too early. The cure is the Goldilocks principle: three candidate moves, ranked good, better, best, before you commit. Here Capablanca builds a beautiful position against R. T. Black (yes, that was his name) and sets a long trap for the enemy queen. Then at move 26 he hesitates for a long time, weighs exactly two moves, and plays the one that wins the queen but gives away most of his advantage. The third candidate, 26.Rc4, was much stronger, and he never looked at it. He still won, because he was Capablanca. Take one idea from this game: when a decision matters, push yourself to find a third candidate, because your brain is exceptionally good at lying to itself.

Three ideas you'll take from this endgame
01

Push for three candidates

Find three moves and rank them good, better, best. The third candidate is often the one you did not want to look for, and sometimes it is the best move on the board.

02

Take their plan away

Ask why your opponent played their last move, then take the idea away with a move that helps you too. 19.b4 stops ...b4 and fixes b5 as a target.

03

Play where your pawns point

In a closed center the pawn chain points like an arrow to the wing where your pieces have room. That is where your play belongs.

Study Mode: Watch & Play Along

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.

▶ The Lesson
♞ Play Through the Game: The Board Stays With You
The Big Idea

Your brain will lie to you. Make it find a third move.

This game has two parts. In the first, Capablanca shows you how to build. He closes the center with 10.d5, sees that his pawns point at the queenside, and starts to work there with 16.a4. When Black's rook comes to b8 to prepare ...b4, he stops it with 19.b4, which is provention at its best, because it also fixes the pawn on b5 as a long-term target. Then 22.Ra5 begins what he calls "a very long combination" to trap the queen. In the second part, we learn from his one real flaw. At move 26 he suddenly realizes that winning the queen will give Black a rook, a bishop and a pawn, plus a solid position. He tells us he "hesitated a long time", and that he compared 26.Bc6 with 26.Bxd7. Two candidates. I would have made him find a third, as I do with all of my students, and the third is 26.Rc4, which keeps the extra pawn on b4 and keeps every threat. The engine has it ahead by more than a pawn and a half. His instinct told him something was wrong, and it was correct. His process just did not go far enough to find out what. Do not call 26.Bc6 a blunder, because it still won. Call it a process flaw, because that part you can fix in your games.

How the Endgame Arose

A closed Ruy Lopez, and pawns that point the way

This was round one of the Rice Memorial in New York, in 1916, and the opening is a Ruy Lopez. At move 10 we get the game's first candidate decision, with the bishop pinning the knight on f3. 10.d5 is the top choice, ahead of 10.a4, with 10.Be3 third, because 10...d5 then gives Black an easy, level game. Capablanca's own note offers Be3 as the alternative, so his judgment was better than his modesty. Now the center is closed, and that tells you where to play. Your central pawns point like an arrow to the wing where your pieces have room. For White that is the queenside, and for Black the kingside. 11.h3 puts the question to the bishop, and it is a useful tip when you are pinned: trade or retreat, what do you want to do? Then 13.Nf1 and 15.Ng3, the standard Ruy Lopez knight route, with 14.g4 to gain space and shut out the black bishop. As Capablanca says, as Black prepares on the kingside, White "begins operations on the other wing", with 16.a4. After the trade on b5 the lesson begins.

Stop and Assess

The rook came to b8. What does Black want?

Material is level and the center is closed. White has more space and has just opened the a-file. Black has played ...Rb8 and then recaptured on b5. It would be easy to think the rook is only there to defend the pawn. Before you make a move, ask the question that I keep coming back to in this series: why did my opponent play that, and what would they do next if it were their move? This is what I call a provention moment.

White to play. What does Black want to do on the queenside, and how do you take it away with one move that also helps your own plan?

Look at the pawn on b5 and the rook behind it. If Black could push, what would happen to that pawn? Think, then reveal.

Capablanca vs R. T. Black, after 18...axb5: White to play in a closed Ruy Lopez.
After 18...axb5 Black wants ...b4. Capablanca plays 19.b4 himself
Reveal the assessment

Capablanca's plan

  • 19.b4. It stops ...b4, fixes the pawn on b5 as a long-term target, and in Capablanca's words also secures c5 for the knight. The engine has it first.
  • Aim at b5. 20.Bd3 is technically a bad bishop, with its central pawns on the same color, but it is very useful on this wing.
  • Stay on the queenside. The earlier g4 and Ng3 hold the kingside, so he can play where his pawns point.

Black's problem

  • ...b4 is gone. That move would have made Black's queenside pawn safe from attack, and now it can never happen.
  • No room. Black's pieces have to squeeze through just a couple of squares to reach the queenside.
  • Only one break. He needs a pawn break, so ...c6 comes, and that opens the board for the white pieces.
The Plan

The bait on c3, and a queen with nowhere to go

After 19...Qc8 20.Bd3 c6 21.dxc6 Qxc6 Black's queen wants the pawn on c3, and here you need three candidates, because you do not always have to answer a threat directly. I would call 22.Bd2 good, 22.Ra7 better and 22.Ra5 best, and that is the engine's order too, though the first two are very close. Capablanca writes that 22.Ra5 "apparently has no other object" than to swap pawns, when really it is the start of the long combination. Chernev compares it with Capablanca's game against Vidmar at London in 1922, where a similar trap appears. After 22...Qxc3 23.Bxb5, the book says this is much better than 23.Bd2, and here I disagree with him. The engine reads it the other way around, because 23.Bd2 gains time on the queen and lets White win b5 later without having to look after it. He was a little in love with his idea, and I will not blame him. 23...Qc7 24.Be3 Red8 25.Rc1 Qb7. White is now about two and a half pawns better, and we have reached the decision of the game.

The Critical Moment: Think First

It is White to move. The queen on b7 is short of squares, and Capablanca can see how to win her. He has two moves in mind. 26.Bc6 hits the queen and starts the trade, but he sees that Black will get a rook, a bishop and a pawn for her, along with a solid position. 26.Bxd7 Qxd7 27.b5 keeps things simple, with what he calls an excellent game. He hesitated for a long time here, which was unusual for him. Now imagine I am your coach, and I am going to ask you in review what you considered. You cannot show me two. Find me a third.

White to play. You have 26.Bc6 and 26.Bxd7. Find a third candidate, then rank all three good, better, best.

Capablanca vs R. T. Black, after 25...Qb7: White to play, the turning point of the game.
After 25...Qb7 Capablanca weighed two moves. The third, 26.Rc4, was the best
Reveal the move and the why

26.Rc4, the move he never considered. Ask what you would like to keep. The pawn on b4 is an outside passed pawn, so why give it away? 26.Rc4 protects it along the rank, and every threat stays on: 27.Bc6 and the invasion on the c-file are still coming, and the queen is still in trouble. After 26...Nb6 27.Bc6, White keeps a large advantage. Here is how the three compare. The engine has 26.Rc4 at about two and a half pawns, 26.Bc6 at about three quarters of a pawn, and 26.Bxd7, the line he called excellent, at only about a quarter of a pawn. So his "spidey senses", as I call them on the video, were right. Bc6 was not as winning as it looked. But be fair to him. 26.Bc6 was not a blunder, and he went on to win. The trap did work, although it needed some help. After 26...Qxb4 27.Ra4 Qb3 28.Ra7, the move he calls the key, Black had 28...Nhf6, and the engine sees no forced win of the queen, with White only about a pawn better. Black chose 28...Nhf8 and after 29.Nd2 the queen was soon lost. The lesson is not that Capablanca was wrong. It is that even the greatest natural talent you will ever study could give away most of an advantage without a third candidate. I train my students on process so that you can make consistent decisions. You will not always find the best move. You will find it far more often than you do now.

♞ Follow the Finish: The Board Returns
The Queen Falls, and the Real Work Starts

Winning her was the easy part. Converting took 30 moves.

The queen hunt ends quickly. After 29...Qb2 30.Qd1 Bg5 31.Rc2 she is attacked and has no good square, so Black takes what he can: 31...Bxe3 32.Rxb2 Bxa7 33.Rxb8 Nxb8. And now look at what he has actually won: a queen for a rook, a bishop and a pawn, which in raw points is exactly even. White has an edge of under two pawns on the engine. So this is not simply a won game. Although he definitely has an advantage, it is not easy to convert. That is where this video stops. The finish is a lesson of its own, and I cover it move by move in Capablanca's Best Chess Endings, Episode 22: Capablanca Was Up a Queen and the Game Lasted 31 More Moves. This lesson is how the advantage was built. That one is how it was converted.

Capablanca vs R. T. Black, after 33...Nxb8: White has a queen against a rook, a bishop and a pawn.
After 33...Nxb8 a queen for rook, bishop and pawn, and the conversion is still to come

You can play the rest through on the board above. Capablanca himself was unsure about it. In his note to 38...Kf8 he says, "I am not sure that White could win at all", if the king had stayed back. But Black marched the king forward with 38...Kf8, 39...Ke7 and 40...Kf6, right into a net, and 43.f4 opened it up. If 43...Nxf4, then 44.Nxf4 exf4 45.Qb2 is mate. By the end, after 63.Ne7, the black king is in the corner, the rook has no good move, and Black resigned, with the engine reporting a forced mate. For how to finish games like this, with three pieces working against an exposed king, watch the endgame episode. Winning material is not winning the game. You still have to find the plan that wins.

Capablanca vs R. T. Black, final position after 63.Ne7: the black king is driven to the corner.
The finish, 63.Ne7: thirty moves after the queen fell, Black resigns
Principles to Take to Your Own Board

This game is over. Your games are where it matters.

1 · Three candidates, every time it matters.

The principle: find three moves and rank them good, better, best, the Goldilocks principle. In this game: Capablanca compared two moves at move 26 and missed 26.Rc4. In your games: if you have only two moves in mind, you have not finished looking. Nothing replaces calculation, so calculate all three.

2 · When you hesitate, listen.

The principle: a doubt during a long think is information, so use it to look wider, not to choose faster. In this game: Capablanca's instincts were right that 26.Bc6 was less than it seemed. In your games: when something feels wrong, search for another candidate.

3 · Take their plan away.

The principle: ask why your opponent moved there, then stop the idea with a move that also helps you. In this game: 19.b4 stopped ...b4 and fixed b5 as a target. In your games: ask this question after every move your opponent makes.

4 · In a closed center, play where your pawns point.

The principle: the central pawns point to the wing where your pieces have room. In this game: White's pawns pointed to the queenside, and that is where he went with 16.a4. In your games: when the center locks, look at your pawn chain before you choose a plan.

5 · Winning material is not winning the game.

The principle: after you win material, count what you really have and make a new plan. In this game: the queen came at the cost of a rook, a bishop and a pawn, and the win still took 30 moves. In your games: do not relax when you win something. That is when the next job starts.

Test Yourself

Five questions before you go

1 · How do I stop blundering in chess?

Many blunders come from one habit, which is playing the first move that looks good. The moment you let go of the piece your brain suddenly finds the problem, because it was no longer busy defending your choice. The fix is a process. In a long game, find three candidate moves, compare them, and rank them good, better, best. Check the forcing moves for both sides, then decide. It will not make you perfect, but it brings consistency to your decisions at every level. In a quick game you still look for options, just more briefly. Even Capablanca gave away most of an advantage in this game because he stopped at two.

2 · What is the Goldilocks principle in chess?

It is my name for finding three candidate moves and ranking them good, better and best. You do not need to be sure of the exact order, and you are not looking for the most forcing move. You are forcing yourself to compare. In this game I rank 22.Bd2 good, 22.Ra7 better and 22.Ra5 best, and at move 10 I rank 10.Be3 behind 10.a4 and 10.d5. The engine agrees with both orders, but the order is not the point. The point is the third move. The moment you have to find one more option, you start to see the position differently, and that is where the moves you were missing come from.

3 · Why was 26.Rc4 better than Capablanca's 26.Bc6?

Because it keeps everything. 26.Bc6 lets the queen take on b4 with 26...Qxb4, and White loses the outside passed pawn. 26.Rc4 protects that pawn along the fourth rank and keeps 27.Bc6 in reserve, so the queen is still in trouble and White has not given anything back. The engine rates it at about two and a half pawns against about three quarters for the move played. Capablanca already felt something was not right, and he says what decided him was that Black's extra pawn would be the backward pawn on d6. He was comparing two options and choosing the better. He never asked if there was another one.

4 · What does "the pawns point like an arrow" mean?

In a closed center the pawn chains tell you where to play. Look at your central pawns and see which way they lean. White's pawns on d5 and e4 point toward the queenside, and that is the wing where his pieces have room to move and where the pawn breaks are. Black's pawns on d6 and e5 point the other way, so Black's pieces fit best on the kingside. Black's ...h6 and ...Nh7 fit that picture, and Capablanca answered at once on the other wing with 16.a4, as his note says, to keep the initiative on the whole front. It is a principle, not a law, but it is the first thing to check when the center locks.

5 · What is the one habit this game should change?

Stopping at two candidates, or one. Think about the last game you lost through a single move. Were you really comparing options, or had you already decided? Next time you have a real decision in a long game, imagine that I am going to ask you in review what you considered. You will not want to tell me you looked at one move. So find a third, just for me, and rank them. That one extra step is often what finds the move you were going to miss. It is what I would have told Capablanca at move 26, and it is what I want you to take into your games.

Sharpen Your Eye: Six Endgame Exercises

Now you convert the edge

Six traps, each one catching a queen, and in several of them she has just grabbed a pawn she should have left alone. They rise in difficulty. Look for the quiet move that takes away her last safe squares. Drag the pieces; the board answers for the defense.

Capablanca Stopped at Two. How Many Do You Find?

Every one of my students learns to find three candidates, and in review I ask them what they considered. In a 2-hour Strategy Assessment I go through your games personally, find where your decisions are breaking down, and build you a roadmap for your chess.

Book Your $130 Strategy Assessment 2 hours via Zoom · Worldwide · Limited spots
The finish has its own full lesson: watch Capablanca's Best Chess Endings, Episode 22, to see how he converted a queen against rook, bishop and pawn.