Steal Magnus's Superpower: Train Your Endgame Conversion

The Skill That Wins Rapid Tournaments

Watch Magnus in this month's rapid events and you'll notice something almost cruel. He reaches an endgame that most of us would shake hands on — a rook ending up a single pawn, an opposite-colored bishop position, a queenless middlegame with a slightly better structure — and thirty moves later his opponent is resigning. Nothing dramatic happens. There's no brilliant sacrifice. He just keeps asking questions until someone answers wrong.

That's not talent you can't touch. It's a trainable skill, and it's arguably the highest-value thing a club player can practice. Most games between 1000 and 1800 players are decided by endgame errors, not opening theory. Let's build the muscle.

Why Conversion Is Different From Endgame Theory

Knowing that the Lucena position wins is theory. Grinding is something else: it's the practical art of maintaining pressure, avoiding simplification into a dead draw, and forcing your opponent to make one accurate move after another.

Consider a classic Carlsen-style setup — a rook endgame where White is up a clean pawn:

White: Kg1, Rd4, pawns a2, f2, g3, h4
Black: Kg7, Ra6, pawns a7, f7, g6, h5

Many players here would try to "do something clever." Magnus doesn't. The plan is boringly correct: 1. Kg2 improving the king, 2. Kf3, 3. Ke4, centralizing, then push the queenside majority with the king supporting. The key idea is that you never rush the pawn — you improve every piece to its maximum first, and only then advance. Against a human at rapid time controls, that patience is what wins.

The Conversion Drill

Here's a concrete routine. Do this three times a week for 20 minutes.

1. Set up a small advantage against the engine. Use any engine set to roughly 1600-1900 strength (stronger than you, but beatable). Start from a position where you're up one pawn or have a clear structural edge — a winnable-but-not-trivial endgame.
2. Play it out fully. No takebacks, no resigning for your opponent. You must convert to mate or a won pawn race.
3. Log the result. Win, draw, or loss — write it down with a one-line note on where it slipped.

The magic is in step 4:

4. Replay the same position immediately. If you drew or lost, you already know the terrain. Fix your mistake and win it. This spaced-in-the-moment repetition is where real learning happens.

Where to Find Positions

Three reliable sources:

• Your own games. Filter for endgames you drew or lost from a better position. These are your personal weaknesses on a plate.
• Endgame study books like Dvoretsky, but only the practical chapters — skip the theoretical fortresses at first.
• Lichess board editor — build the rook ending above and play it against the computer.

The Three Grinding Principles

While you drill, keep these Carlsen habits in mind:

1. Improve your worst piece

Before pushing pawns or forcing trades, ask: which of my pieces is doing the least? In the rook ending above, it was the king. Fix that first.

2. Don't trade into a draw

The most common conversion error at the club level is trading down carelessly. Up a pawn in a rook ending? Great — that's often a win. Up a pawn but you just traded into a king-and-pawn ending your opponent holds? Disaster. Calculate every trade to the finish line.

3. Make your opponent solve problems

Even in "equal" positions, keep tension. A move like ...h5-h4 creating a passed pawn threat, or maneuvering a knight to a square that pins your opponent down, forces continuous decisions. Every decision is a chance for them to err.

A Two-Week Challenge

Track your conversion percentage over ten drilled positions. Most club players start around 40-50%. After two weeks of the replay method, you should see it climb toward 70%. That improvement translates directly into tournament points — those "agreed draw" games you used to split now become wins.

The Takeaway

You don't need Magnus's opening prep or his calculation depth to steal his most practical weapon. Conversion is a skill, and skills respond to deliberate practice. Set up a slightly winning endgame tonight, refuse to let the engine off the hook, and replay it until you win. Do that consistently, and you'll become the player nobody wants to reach an endgame against.