Quick answer
Quick answer: Your pawn sits on the seventh, your king sits on the promotion square, and their king is cut off by one file. You win by building a bridge. Rook to the fourth rank, king out, and the rook blocks the check that follows.
It's four moves and it works every time the position is right.
I read through the standard treatment before writing this. The part people get wrong isn't the idea. It's which rank the rook goes to.
What the position actually is
Four things have to be true, and if any one is missing this isn't a Lucena.
You have a rook and a pawn. They have a rook. The pawn is on the seventh rank, one square from promoting.
Your king stands on the promotion square, in front of your own pawn.
And your rook cuts their king off along a file, so it can't come across and help.
Why the position looks stuck
Here's the problem you're solving.
Your king is standing on the square your pawn wants. Move the king out and their rook checks you forever, because it sits on the far side and has the whole board to check from.
Stay put and the pawn never promotes.
So the win isn't about pushing the pawn. It's about giving your king somewhere safe to stand while it does.
The bridge, in plain terms
You use your own rook as a shelter.
Put it four ranks back from the pawn, on a rank it can slide along. Then walk your king out towards the enemy rook.
They check. You walk further. They check again. Eventually you interpose your rook, blocking the check, and now the checks are over.
That interposition is the bridge, and it's the whole technique. Everything else is setup.
The move order with a pawn on b7
Take the standard version. White king on b8, pawn on b7. Black king on d7, black rook on the a file.
White swings the rook to d1, giving check along the d file, and the black king has to step away to e7.
Then Rd4. That's the move people miss, and it's the one that matters.
Then Kc7, and the checks start. Rc1 check, Kb6. Rb1 check, Kc6. Rc1 check, Kb5. Rb1 check, and now Rb4 blocks it. The pawn promotes next move, and there is nothing Black can do about it.
Why the fourth rank
The rook has to sit exactly where it can step in front of a check on the rank your king will occupy.
Too close and your king has nowhere to go before the shelter is needed. Too far and the rook can't reach in time.
I counted the squares in the main line and the rook lands on the fourth rank, four below the promotion square. The fifth rank works in some versions of the position.
Get this wrong and the whole thing collapses into a draw, which is exactly how most players lose it.
Cutting the king off is a precondition
Without the cut, none of this happens.
If their king can walk over to the pawn, it defends the promotion square and you have nothing. One file of separation is the minimum.
So if you reach an endgame heading for this, your first job is getting the rook onto the file beside the pawn.
That one move earlier in the game decides whether the ending is winning or drawn.
Two files of separation is more comfortable still, because it buys you a spare tempo if the move order goes wrong.
Where it fails: rook pawns
A pawn on the a file or the h file usually doesn't win this way.
The reason is geometry. Your rook needs room on both sides of the pawn to build the bridge, and an edge pawn only has one side.
There are exceptions with the defending king badly placed, but the working assumption should be that a rook pawn is a draw.
The full treatment of rook and pawn against rook covers the exceptions properly.
Where it fails: the short side
The second exception is the one that catches strong players.
If the defending king sits on the short side of the pawn, meaning the side with fewer files, and has three or more files between it and your king, the checks come from the long side and never stop.
That's a draw despite everything looking right.
So when you're defending, that's where you put your king. Short side for the king, long side for the rook.
The defender's version
If you're on the wrong end of this, you have two jobs.
First, stop the pawn reaching the seventh with your king cut off. Once that happens you're usually lost.
Second, if it's already happened, get your king to the short side and your rook to the long side and hope the distances work.
The other defensive method is the Philidor, which works earlier and more reliably, and is a separate technique worth learning first.
And keep your rook active. A passive rook stuck defending the pawn from in front is how most of these endings are actually lost.
How often this comes up
More often than beginners expect and less often than endgame books suggest.
Rook endings are the most common ending in practical chess, and this is the single most important winning position inside them.
So the return on twenty minutes here is high. It's not a rare curiosity.
If you're picking what to study, this and the Philidor are the two positions that repay learning fastest. Our endgame study order puts them near the front for that reason.
Drilling it properly
Set the position up and win it against an engine ten times. Not five.
The first three you'll follow the moves. By the tenth you'll understand why each one is there, which is the difference between remembering it and knowing it.
Then set it up with the king on the other side and win it from there too, because the mirrored version feels completely different.
You can do both against a bot on our play page in about fifteen minutes.
Do it once more a week later without looking anything up. That is the check that tells you whether it stuck.
The one mistake almost everyone makes
Pushing the pawn too early.
People get a rook and pawn ending, see a free tempo, and run the pawn to the seventh before the king is in front of it.
Now the pawn is on the seventh with nothing behind it, the defending king has time, and a winning position has become a draw.
King first. Pawn second. Always in that order, and it's worth repeating to yourself in the game.
Recognising it three moves early
The useful skill isn't executing the Lucena. It's seeing that you're heading for one.
When you have an extra pawn in a rook ending, ask whether you can get your king in front of it and their king cut off. If yes, trade into it. If no, keep pieces on.
That single question turns a vague endgame into a plan.
The same habit applies to the whole subject, which we covered in winning rook endings.
Why it is named after nobody in particular
The name comes from a fifteenth century Spanish book, and the position isn't actually in it.
It was attributed by later writers and the name stuck, which happens a lot in chess.
None of that matters at the board. It's mentioned because people search for the name and then wonder whether they've found the right position.
The position itself is what counts, and it's been the same for centuries.
What to learn next
The Philidor position, which is the drawing technique and the thing you'll need more often.
Then the basic king and pawn oppositions, because rook endings collapse into them constantly.
Then the fortress ideas, for positions where you're much worse and need a target to aim at.
Those three plus this one cover the endgame situations that decide most games under two thousand rating. We went through opposition separately.