When players castle on opposite wings, the game turns into a race: each side throws pawns and pieces at the other's king, and the attack that arrives first wins, whatever the scoreboard says. In this King's Gambit a young Capablanca castled queenside, aimed everything at the white king on g1, and simply out-ran his opponent. Raubitschek spent his queen chasing material on the other side of the board, winning a knight and more, and it meant nothing, because a queen on h8 does not defend a king on g1. Capablanca doubled his rooks on the g-file, threw one in with a check, deflected the last defender with a bishop, and mated. Take one idea from this above all others: in an opposite-side castling race, count tempos, not material, and the player who opens a file at the enemy king first is usually the one who gives mate.
When the kings go to opposite wings, both sides attack and defence barely matters. The faster attack decides the game.
Capablanca piled his rooks on the g-file against the white king. An open file pointing at the king is the highway an attack travels down.
Raubitschek won a knight and kept two spare pieces, and his king was mated anyway. Material that cannot reach the fight does not count.
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.
Both sides castled, but on opposite wings: White short, Black long. That single fact sets the terms of the game. From here it is a race, and Capablanca runs it faster. He swings the rook to the h-file with 17...Rh4, trades into the white camp, and when Raubitschek sends his queen after material with 20.Qg7 and 21.Qh7 and then wins the knight with 23.Qxf6, Capablanca does not chase it back. He doubles his rooks on the g-file with 23...Rdg8, crashes in with 24...Rxg2+, and after 25.Kf1 plays the quiet clearance check 25...Bc4+, opening the e-file so his queen covers the king's escape at e2; after 26.Nxc4 the move 26...Rg1 is mate. On material White was two whole pieces ahead. On the board his king was dead. A whole game decided by one idea: in an opposite-wing race the first attack home wins, and everything else is noise.
Raubitschek chose the King's Gambit, a sharp opening that hands Black a pawn for open lines, and the players committed their kings to opposite sides: White castled kingside on move 8, Black queenside on move 16. That is the signal to attack. Capablanca, even as a teenager, understood that an opposite-side castling position is not a quiet manoeuvring game but a sprint, and that the correct currency is the tempo, not the pawn. The lesson begins on move 17, with both kings housed on opposite wings and the race about to start.
After 16...O-O-O the kings sit on opposite wings and White is a pawn to the good. The pawn is not the point. This is the moment to stop counting material and start counting tempos. White's king is on g1 with an open-ish kingside in front of it; Black's king is tucked on c8. The question is simple and decides the game: which attack gets there first? Stop and assess before you read on.
The kings are castled on opposite wings and White is a pawn up. Whose attack is faster, and what should each side be racing to do?
Look at the files pointing at each king. Picture Black's rooks doubling on the g-file against g1, and White's queen wandering after pawns on the queenside. Count how many moves each side needs to land a blow, then reveal.
Capablanca's plan is pure attacking technique. He brings the rook to the kingside with 17...Rh4, exchanges a pair of minor pieces to clear the way with 18...Bxe5, and builds pressure with 19...Rd5 and 20...Rg4. Raubitschek, meanwhile, chases material: 20.Qg7 and 21.Qh7 hunt around and 23.Qxf6 wins the knight, but each queen move is a tempo not spent defending. Capablanca doubles on the open g-file with 23...Rdg8, sacrifices with 24...Rxg2+ to blow open the king, and after 25.Kf1 finishes with the clearance check 25...Bc4+. The immediate 25...Rg1+ would fail to 26.Ke2, but Bc4+ opens the e-file so the queen on e7 covers e2, the king's only flight; after 26.Nxc4 there is no escape and 26...Rg1 is mate, the rook guarded by its partner on g8. The white queen and the extra knight are spectators on the far side of the board.
Here is the payoff. Raubitschek has just played 24.Rf2 to shore up the second rank, and on material he is miles ahead, two whole pieces to the good. But Capablanca's rooks are doubled on the g-file aimed straight at g1, and it is Black to move. There is a forcing sequence that mates. The whole race comes down to the next three moves.
Black to play. How does Capablanca break through on the g-file, and why is the huge material deficit meaningless?
24...Rxg2+, and the king's shelter is torn open. After 25.Kf1 the rush 25...Rg1+ does not work, because the king slips out to e2. So Capablanca plays the quiet clearance check 25...Bc4+ first. It does two jobs at once: it gives check, and it vacates e6 so that the queen on e7 now rakes down the open e-file and covers e2, the king's only escape. White deals with the check, 26.Nxc4, and now 26...Rg1 is mate, the rook protected by its partner on g8 and the king with nowhere to run. Count the white army at that moment. A queen on f6, a rook on f2 right beside its own king, a bishop on c1 and an extra knight, every one of them powerless to cover g1 or give the king a flight. In your games, when you have castled on opposite wings, do not get into a capturing contest on the far side. Open the file in front of the enemy king, double your heavy pieces on it, and before you force the mate, take away the king's escape squares. The race is won by tempos, and mate ends it no matter what the scoreboard reads.
Look at the position after 25.Kf1. Capablanca has just given up a rook on g2 to rip open the king. The straightforward 25...Rg1+ would only chase the king to e2, so he plays the clearance check 25...Bc4+ instead: it opens the e-file for his queen on e7, which now covers e2, and forces White to answer the check. Look at the white pieces. The queen on f6 and the rook on f2 sit right by the action, yet neither can cover g1 or shield the king, and the bishop on c1 and the spare knight are just as useless. This is the picture of an opposite-wing race that has already been decided.
26.Nxc4 Rg1 is mate. A boy had checkmated his opponent while two whole pieces down on the board, because material counts for nothing when none of it can defend the king. Carry the real lesson into your games: when the kings are castled on opposite wings, treat the game as a race, open the file in front of your opponent's king, and pour your pieces down it. Do not be lured into winning material far from the action, and do not fear giving it up to land the final blow. The fastest attack wins, every time.
The rule: when the kings go to opposite wings, both sides must attack, because defence cannot keep pace with two pawn storms. In this game: White castled short, Black long, and the race was on. In your games: when you castle opposite your opponent, switch from manoeuvring to attacking at once.
The rule: an attack needs an open line; prying open the file in front of the enemy king is the first job. In this game: Capablanca opened and seized the g-file against g1. In your games: aim your pawns and rooks at the file that leads to the king.
The rule: in a race, the move that brings the attack closer is worth more than the pawn it ignores. In this game: White won a knight and lost the game. In your games: judge each move by whether it speeds the attack, not by material.
The rule: a king with one flight square is not mated, so before you strike, remove that square. In this game: 25...Bc4+ opened the e-file so the queen covered e2, the king's only escape. In your games: count the king's flight squares and take the last one away before forcing mate.
The rule: extra material helps only if it can reach the defence; pieces that cannot cover the key squares are worthless. In this game: White was two pieces up and not one of them could guard g1. In your games: judge your material by whether it can defend the king, not by its count.
Because defence and attack no longer balance. When both kings sit behind their own pawns on the same wing, an attacker has to break through a shield while staying safe himself. When the kings are on opposite wings, each side can hurl its pawns forward without exposing its own king, so both attack at full speed and the game becomes a race. The correct mindset switches from slow improvement to fast, forcing play.
Because all of it was on the wrong side of the board. A queen on f6, a rook on f2 and two spare minor pieces do nothing if not one of them can cover g1 or the king's escape squares. Capablanca's attack arrived first, and once the king is mated the scoreboard is irrelevant. Material only matters when it can influence the part of the board where the game is decided.
It tore open the king's shelter. Being behind on material is exactly why Capablanca had to finish by force rather than slowly, and the rook sacrifice opened the g-file and drew the king to f1 where the final deflection worked. In an attack, a sacrifice that keeps the initiative and leads to mate is not a loss of material, it is the price of the win.
Because the immediate 25...Rg1+ lets the king escape to e2. The move 25...Bc4+ fixes that: it checks, and by leaving e6 it opens the e-file so the queen on e7 covers e2, the king's only flight square. After White answers with 26.Nxc4, the rook check 26...Rg1 is now mate, because the king has nowhere to go and the rook is defended by its partner on g8. Spotting the move that removes the last escape square is a core finishing skill.
Counting material in a race. When you have castled on opposite wings, stop asking who is ahead on pawns and start asking whose attack is faster. Open the file in front of the enemy king, bring your heavy pieces to it, and spend material to keep the initiative. The player who lands the first mating attack wins, however lopsided the scoreboard looks.
Six kingside attacks, rising in difficulty. In each one you have the attacking side; find the blow that breaks through and play it out. Drag the pieces; the board plays the defence for you.
Knowing when a position is a race, how to open a file at the enemy king, and when to spend material to keep the attack alive are the skills that decide sharp games. In a 2-hour Strategy Assessment I go through your games personally and hand you a roadmap built for your chess, not a hypothetical student''s.
Book Your $130 Strategy Assessment 2 hours via Zoom · Worldwide · Limited spots