Wall Economy: When a Wall Beats a Step

Every wall you own is a promise you get to keep exactly once. The standard game hands each player ten and provides no way to get an eleventh — walls are never moved, traded, or recovered once placed (The Complete Rules has the formal statement). Ten is the entire budget for a game usually decided in a few dozen moves. Wall economy is the study of what that budget actually buys, and the honest answer is: less than it looks like, and only in a narrow window.

What one wall returns, counted exactly

Start from the race. When it is your move and your shortest route home is no longer than your opponent's, you arrive first — you are a half-turn ahead by virtue of moving next, and Quoridor Endgames proves that from the parity of the arrival plies. Tempo and the Race covers how to produce those two numbers; here we only need the comparison.

So suppose both routes are nine squares long and it is your turn. Instead of stepping, you place a wall that adds two squares to their route and none to yours. They now need eleven moves, you still need nine, and it is their turn. You were winning by a half-turn; now you are winning by a turn and a half.

That is the whole return: one tempo. Not two. Half of the detour you bought went immediately back into paying for the turn you spent buying it, because the alternative — stepping — would have moved you a square closer for the same turn. The general form prices every wall you will ever consider:

Tempo gained = k − j − 1, where k is the number of squares the wall adds to your opponent's shortest route and j the number it adds to your own.

Run the ladder. At k − j = 0 the wall costs you a full tempo and a wall: you handed your opponent a free move and paid for the privilege. At 1 it is exactly neutral in the race and still costs a wall, which makes it a pure loss. At 2 it nets one tempo. Only at three or more does a single wall start returning real value. Two is the bar, and most candidate walls on most boards fail it.

Everything below is what that expression does not tell you: how large k can actually get, what a wall costs besides the turn, and what the walls you never place are worth.

Why k rises and then collapses

k is not a free parameter you can choose by placing the wall well enough. A wall cannot remove a route, only lengthen one, so k is fixed by the gap between your opponent's best route and their best route that avoids the two edges you just blocked. That gap is the ceiling, and no amount of placement skill raises it.

Two independent quantities drive that ceiling, and they move in opposite directions across a single game.

The first is how many alternative routes are still live. That number falls as walls accumulate, as a pawn commits to a file, and as it drifts toward a side of the board. Fewer alternatives means the second-best route is much worse than the best, which pushes k up.

The second is how much board still lies between that pawn and its goal edge. That number falls as the pawn advances, and it pushes k down, because a detour needs somewhere to happen. Against a pawn one row from home, a wall across its path forces a sidestep and nothing more: one square, below the bar — two if the pawn is hugging a side of the board, where there is less room to step around (Wall or Advance works out why edges do that). No wall can do better, because the wall that would actually stop it is the illegal one.

So wall value is a hump, not a slope. It rises through the opening and the midgame as the board narrows, peaks while the target is committed to a line and still several rows out, and collapses toward zero as it closes on its goal edge. Walls appreciate, then expire, and the whole skill of wall economy is selling near the peak.

Both ends of the curve are expensive to ignore. Spending early is spending at the bottom: on a nearly empty board almost every detour has a free way around it, which is why Quoridor Openings treats the opening wall as a donation unless it has a specific job, and works out which early walls pay anyway. Holding too long is worse, because expiry is total — a supply still in your hand when your opponent steps onto their goal edge was delay you owned and never sold.

The cost side has a second line item

Walls also spend the board. A wall occupies its groove permanently and forbids anything that would overlap or cross it, so every placement quietly deletes a set of placements you might have wanted later. That deletion is the part nobody prices. The follow-up wall that would have sealed the position is frequently the one that would have crossed the wall you played on impulse fifteen moves earlier. You do not merely lose the piece; you lose the shapes it made impossible.

This is the second reason two squares is break-even rather than profitable. One tempo, one wall, and a permanent narrowing of what you can still build is not obviously a trade worth making — particularly if the tempo is surplus, since margin above the winning threshold buys nothing and Quoridor Endgames has the exact number of free turns a lead can afford. It is also the economic case for spending in planned pairs rather than in reactive singles, because walls placed in relation to each other buy a far better rate than scattered ones (Traps and Cages covers the structures).

What a held wall is worth

Now flip the ledger and price your opponent's reserve. The convenient shorthand is that every wall they hold is a tempo they can spend whenever they like, so four walls is roughly four tempo. That is an upper bound, not an exchange rate.

A held wall is worth about a tempo only where a k − j ≥ 2 placement actually exists against you, and such placements are not free-floating: they need an edge to anchor against, a wall already standing, or a pawn committed to a file. On open board there is nothing for a wall to lean on, and the reserve value of holding one is close to zero.

The practical consequence is that you should count placements, not walls. Look at where the k − j ≥ 2 grooves against you actually are. If your route runs through open ground and there are none, an opponent holding five walls is holding five options, none of which currently beats walking. If your route funnels into one narrowing corridor with three such grooves, a single held wall is a genuine threat and the second one is worse. This is why "I am four ahead and they have four walls, so it is level" tends to be wrong in both directions at once: early they cannot cash four walls at a tempo each, and late one wall in front of a committed pawn can be worth more than four.

Running out first

An empty supply carries no penalty of its own — that is a rules fact, set out in The Complete Rules — and games are won every day by the player who emptied theirs first, because they sold at the peak of the curve above. What running out actually costs is optionality. Wall placement is one of the two things a turn can be, and once it is gone your distance becomes a number your opponent can edit and you cannot. Every subsequent turn, they choose between two kinds of move and you choose between one.

The mirror image is just as strong. Holding the last wall in a level position is frequently the whole game: you have a tempo in your pocket that cannot be answered, and you can wait for the moment it prices at three rather than selling it at two. Once both supplies are empty the position is a pure race — see Quoridor Endgames for how it resolves.

Every number on this page is a snapshot. Each new wall reprices both routes, so a detour you valued at two can be worth four after their next wall or nothing at all after your own; recount after every placement.

A note on the four-player game

The four-player game cuts each supply to five while keeping the same board, which changes the economics more than the halving suggests. The total wall that can ever reach the board is unchanged — four supplies of five is the same twenty as two supplies of ten — but no single player now controls more than a quarter of it. Your personal leverage halves while the ground you must cross does not.

Worse, a wall you spend to slow the leader is paid for entirely out of your budget and enjoyed by everyone at the table — and the players who did nothing do better out of it than you do. The leader's route grows by k for all three of you, but you paid a turn for it and they did not, so you gain k − j − 1 on the leader while each of them gains k less whatever the wall happens to cost their own route. Every wall you place against the front-runner hands your two rivals a tempo more than it hands you. Reactive walling is wasteful in the two-player game; in the four-player game it is close to charity.

For the turn-by-turn procedure that puts this pricing to work, see Wall or Advance.