The unexpectedly hard windmill question (2011 IMO, Q2)

A single, carefully chosen starting line can drive a “windmill” rotation that repeatedly uses every point of a finite planar set as the pivot—infinitely often. The key is to pick the line so that it begins in a perfectly balanced position (a “middle” line), then track a quantity that refuses to change as the pivot hops from point to point. That invariant balance forces the windmill to cycle through all points rather than getting trapped among only some of them.

The setup starts with a finite set S of at least two points in the plane, with the condition that no three points are collinear to avoid ambiguity about which pivot should take over. A windmill process begins with a line L passing through a chosen point P. The line rotates clockwise around P until it first hits another point Q of S; then Q becomes the new pivot, and the line continues rotating clockwise until it hits the next point, and so on forever. The challenge is to show that there exists some choice of P and L such that every point in S becomes a pivot infinitely many times.

The solution begins by experimenting with small cases: with two points the line alternates; with three points it cycles among them. The first genuinely tricky behavior appears with four points, where a naive starting line can seem to “miss” interior points. The fix is to choose a starting line that runs through a point in the middle of the configuration. To make “middle” precise, the line is given an orientation, splitting the plane into a left side and a right side. For an odd number of points, the starting pivot is chosen so that the number of points strictly on the left equals the number strictly on the right; the pivot itself is treated as neutral. The crucial invariant is then the left–right count: as the rotating line moves from one pivot to the next, points that leave one side are replaced by points that enter the same side, so the number of points on each side stays constant.

Once that invariant is secured, the cycle becomes inevitable. After the line rotates by 180 degrees, it becomes parallel to its starting position. Because the left–right counts must still match the invariant, the only way to maintain the balance is for the line to pass through the same starting pivot again. Meanwhile, a half-turn swaps which points lie on the left versus the right, so the line must have hit every other point during the process. Repeating the motion forever means each point is used as a pivot infinitely many times.

For even |S|, the argument is adjusted slightly by counting the pivot as belonging to one side so that the “equal halves” condition still holds. The same invariant logic forces a return to a corresponding starting configuration after a full 360-degree cycle, again ensuring that all points are visited endlessly.

Beyond the mechanics, the problem’s difficulty is framed as a lesson in mathematical thinking: the hard part isn’t the final invariant-based proof, but the leap to identify a quantity that stays constant amid a chaotic process. That theme—finding invariants—echoes across mathematics and physics, from counting holes in topology to conserving energy and momentum in physics. The windmill puzzle becomes a compact Aesop: when motion is complicated, look for what cannot change.