Day 4of 14, Master the Fundamentals with Capablanca

The Rule of Three

Capablanca vs Znosko-Borovsky · St. Petersburg 1913 · The Rule of Three

Two games running, Capablanca loses to the same habit. In Game 2 he never listed the saving check; here, as White, he calculates every branch with exactly two candidate moves and misses the third every time. His king sits behind a shattered queen-side, a debt he keeps refusing to price, and when Black finally breaks with 28...e5 the bill arrives all at once. Take one idea into your games, name three candidate moves for yourself and three for your opponent before you commit, because two is the number at which your brain stops looking.

Three ideas you'll take from this endgame
01

Three, not two

Two candidates feels like a search and is really a story. Capablanca's own published notes examine two options at every branch, and the refutation is always the third.

02

A weak king is a debt

The shattered c3 and c4 pawns in front of the king were not an inconvenience. They were borrowed money, and 28...e5 was the day it came due.

03

Prevent, then push

28.Qf2 would have stopped the e5 break and held the position. Capablanca played 28.c4 instead, stripping another pawn from his own king and allowing e5 anyway.

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

Two candidates, and the third one loses

Capablanca chose the McCutcheon and accepted a sharp imbalance, his queen-side pawns shattered by 12.bxc3 in return for damage to Black's king cover. Then he castled long, into the damage. His reasoning was that with Black's dark-squared bishop gone the king could not be attacked, and that reasoning has the same shape as the error in Game 2, it examines one line and stops. The permanently weak king is a debt, and Black's whole game is simply the e5 break that collects it. When it lands the tactics run against White at every turn, and the reason is always the same: Capablanca had looked at two continuations and the strongest was the third. In your games, force yourself to write three candidate moves for you and three for your opponent before you calculate. The move that beats you is rarely the second one you thought of.

How the Endgame Arose

A McCutcheon, and a shattered king-side of pawns

The French Defence McCutcheon gives Black an early initiative rather than a quiet defence, and after 10.Bxf6 the position is imbalanced in a very concrete way. White's queen-side pawns are doubled and loose, Black's king-side is torn open. Capablanca then castled queen-side, straight into his own damaged structure. It was an original decision and, as the loss shows, a misjudgement, because the weakness in front of that king never goes away and every later tactic in the game flows through it.

Stop and Assess

White has just played 28.c4. Where is the break?

Material is level, both kings are on the queen-side and both structures are damaged. White has just pushed 28.c4, which loosens his own king further. Black has one central break that opens every line toward that king. Stop and assess before you read on.

Material is level. What is Black's central break here, and why is it stronger than it looks?

Look at the semi-open files and at White's king cover. Which pawn move opens a line toward the king and gains time? Think, then reveal.

Capablanca vs Znosko-Borovsky, after 28.c4, both kings on the queen-side and Black about to break in the centre.
After 28.c4, White has loosened his own king cover and Black's central break is ready
Reveal the assessment

Black's method (Znosko-Borovsky)

  • Break in the centre. 28...e5 opens lines toward the exposed white king.
  • Invade on the second. 31...Re2 puts the rook where it cannot be tolerated.
  • Check and switch. 32...Rc2+ and 33...Rb8+ build a mating net around the king.

White's problem (Capablanca)

  • A permanently weak king. Castling long into shattered pawns was a debt he never repaid.
  • Two candidates only. Every branch is analysed with two moves; the refutation is always the third.
  • 28.c4 instead of prevention. 28.Qf2 stops e5; the text removes another pawn from the king.
The Plan

Break, invade, and mate the exposed king

Black wins in three clean stages. First the break, 28...e5 opens the centre and after 29.Qg1 e4 30.cxd5 exd3 the lines toward White's king are open. Second the invasion, 31.d6 looks fast but 31...Re2 is faster, the rook lands on the second rank where it cannot be tolerated. Third the net, 32...Rc2+ 33.Kb1 Rb8+ 34.Nb3 and White's king is surrounded, after which 37...Qe4 and the exchange of queens leaves a rook ending that is lost precisely because the king is so exposed. Capablanca's notes propose improvements at several points, and at each one he examines two moves and misses a third that refutes him. In your games, list three candidates, and list three for your opponent too.

The Critical Moment: Think First

White has pushed 31.d6 and the pawn looks unstoppable, one square from queening with tempo. Capablanca calculated this far and liked it. Black to move, and there is a single move that changes the evaluation completely.

Black to play. White's pawn has reached d6 and looks winning. What move turns the game around?

Capablanca vs Znosko-Borovsky, after 31.d6, Black to find the rook invasion.
After 31.d6, Black plays 31...Re2, and the invasion is faster than the pawn
Reveal the move and the why

31...Re2, the invasion that outruns the pawn. The rook reaches the second rank, and because White's king lives behind broken pawns the threats arrive faster than the promotion. After 32.d7 Rc2+ 33.Kb1 Rb8+ 34.Nb3 the king is caught in a net and the d7-pawn, so proud a move ago, is simply collected later by 39...Rxd7. Capablanca saw the push and the promotion, the two candidates that told a pleasing story, and did not look for Black's third move. In your games, when a plan looks winning, spend the extra thirty seconds listing your opponent's three best replies. The refutation is usually the one you have not written down.

♞ Follow the Finish: The Board Returns
Three Candidates, Every Time

Build the net, then trade into a won ending

The checks come in the right order. 32...Rc2+ drives the king to b1, 33...Rb8+ brings the second rook with tempo, and 34.Nb3 is forced, blocking the file with a piece that is now pinned to the job. White's king has no shelter, and every white piece is tied to defending it rather than to stopping the passed pawn.

Capablanca vs Znosko-Borovsky, after 33...Rb8+, the black rooks build a net around the white king.
33...Rb8+: both rooks join and White's king is caught in the net

The finish is instructive because it is not a mating attack, it is a conversion. 37...Qe4 offers the trade of queens, and White cannot decline usefully, because the resulting rook ending is lost for exactly the reason the middlegame was lost: the king is too exposed to defend anything. Black rounds up the d7-pawn with 39...Rxd7 and finishes with 40...Re1+. Carry it into your games, three candidates for you, three for your opponent, and price your king weaknesses honestly.

Capablanca vs Znosko-Borovsky, the final position after 40...Re1+, White is lost.
40...Re1+: the exposed king decides, and Capablanca resigns
Principles to Take to Your Own Board

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

1 · List three candidates, not two.

The rule: two candidates is a story, three is a search. In this game: Capablanca examined two moves at every branch, in the game and in his notes. In your games: write three before you calculate anything deeply.

2 · Give your opponent three moves too.

The rule: calculate their best three replies, not just the one that suits your plan. In this game: 31...Re2 was never on White's list. In your games: ask what you would play in their seat.

3 · A weak king is a debt that comes due.

The rule: structural damage in front of your king is permanent, not temporary. In this game: castling long into shattered pawns lost the game forty moves later. In your games: price king weaknesses when you accept them, not when they are exploited.

4 · Prevent the break before you push.

The rule: stop the opponent's freeing move first, then continue your own plan. In this game: 28.Qf2 held; 28.c4 loosened the king and allowed e5 anyway. In your games: prevention first, expansion second.

5 · A passed pawn does not outrun an invasion.

The rule: count the race honestly, threats against a bare king are faster than a pawn. In this game: 31.d6 lost to 31...Re2. In your games: compare tempi before you trust a runner.

Test Yourself

Five questions before you go

1 · What is the Rule of Three?

Before you calculate, name three candidate moves for yourself and three for your opponent. One candidate is confirmation bias, two still lets your brain finish a tidy story, and four is more than you can calculate at the board. Three forces a genuine search.

2 · Why was castling queen-side wrong?

Because White's queen-side pawns were already shattered by 12.bxc3, so the king moved toward the damage. Capablanca reasoned that Black's dark-squared bishop was gone and so no attack was possible, which is a one-line search. The weakness was permanent and it decided the game.

3 · What should White have played instead of 28.c4?

28.Qf2, preventing the e5 break. If Black plays e5 anyway, White gets Nxe6 with real play against the black king. The text move strips another pawn from White's own king cover and does not stop e5.

4 · Why is 31...Re2 stronger than the d-pawn?

Because it is faster. The rook reaches the second rank with threats against a king that has no shelter, and the pawn on d7 never queens, it is collected by 39...Rxd7. Against a bare king, threats beat pawns in the race.

5 · Is this the same weakness as Game 2?

Yes, and that is the point of putting the two games together. In Game 2 he missed a forcing check; here he misses the third candidate in branch after branch. It is a systemic gap in method, not bad luck, and it is the gap the Rule of Three closes.

Sharpen Your Eye: Six Endgame Exercises

Now you convert the edge

Six queen and rook endings, rising in difficulty. Heavy pieces punish an exposed king, so find the invasion square and convert. Drag the pieces, the board plays the defence for you.

Three Candidates. Every Critical Move.

Most club games are lost the way this one was, by stopping the search at the move that tells the nicest story. In a 2-hour Strategy Assessment I go through your own games, find the exact positions where your candidate list ran out too early, and hand you a calculation method built for your chess, not a hypothetical student's.

Book Your $130 Strategy Assessment 2 hours via Zoom · Worldwide · Limited spots
Next, Game 5: Emanuel Lasker vs Capablanca, St. Petersburg 1914.