Pick's Theorem: Count the Dots and Get the Area for Free

Draw any polygon on graph paper — any shape at all, as long as every corner lands exactly on a grid intersection and the sides don't cross themselves. Now forget everything you know about computing area. No base-times-height, no breaking it into triangles and trigonometry, no calculus. Just count dots.

Count the dots strictly inside the shape. Count the dots sitting on its boundary. Plug both numbers into one absurdly short formula, and out pops the exact area — not an estimate, not an approximation, the exact area, every single time. That's Pick's Theorem, and it's one of those results that feels like it shouldn't be allowed to work.

The Concept

Here's the formula in full:

A = i + b/2 − 1

where A is the area of the polygon, i is the number of lattice points strictly inside it, and b is the number of lattice points lying on its boundary (including every vertex).

A "lattice point" is just a point where grid lines cross — the integer coordinates (0,0), (1,0), (3,7), and so on, stretching out infinitely in every direction like an endless sheet of graph paper. A "lattice polygon" is any simple polygon (no self-intersections) whose vertices all happen to fall on these grid points.

Try it on a simple case: a 2×2 square sitting on the grid. It has one interior point (dead center) and eight boundary points (four corners plus the midpoint of each side). Pick's formula says A = 1 + 8/2 − 1 = 1 + 4 − 1 = 4. Check it against the obvious answer — side length 2, area 2×2 = 4. It matches.

Now try something irregular — a lopsided arrow or a jagged star shape with a dozen vertices, all still on lattice points. You could spend ten minutes decomposing it into triangles and trapezoids and summing their areas with the shoelace formula. Or you could spend thirty seconds counting dots. Pick's Theorem gets you to the same exact number either way, but one path is dramatically faster and requires no trigonometry, no square roots, and no careful bookkeeping of signs.

The theorem works for any simple lattice polygon — convex, concave, with arbitrarily many sides, arbitrarily skinny or sprawling. The only requirements are that the vertices sit on lattice points and the boundary doesn't cross itself. That generality is what makes the result so startling: one formula, two integers, exact area, no exceptions.

Why It Matters

Pick's Theorem looks like a cute curiosity for a geometry classroom, but counting dots on a grid turns out to be a surprisingly good description of what computers do constantly.

Pixels are a lattice. A digital image is, structurally, nothing but a grid of lattice points — pixels at integer coordinates. When graphics software needs to compute the exact area of a polygon-shaped region on a screen (for rasterizing shapes, filling regions, or computing collision areas in a game engine), it is working in precisely the setting Pick described in 1899, decades before anyone had built a computer. Algorithms that avoid floating-point rounding errors by working entirely in exact integer arithmetic lean on Pick's-Theorem-style reasoning: count points, get exact area, no accumulated rounding drift from repeated division and multiplication.

GIS and mapping tools. Geographic information systems often let a user trace a region on a digitized map using a snap-to-grid interface. Because the traced region's vertices land on a known grid with a known real-world scale, the area of an irregular plot of land, a lake, or a zoning boundary can be computed by counting grid intersections rather than integrating a boundary curve — fast, exact, and numerically robust.

Computational geometry and competitive programming. Pick's Theorem shows up constantly in algorithm design because it converts a continuous question (what is this region's area?) into a discrete counting question (how many lattice points does it contain?) — and discrete counting is exactly what computers are built to do efficiently. Combined with the "boundary point count" trick (the number of lattice points on a segment between two integer coordinates is the greatest common divisor of the coordinate differences, plus one), you can compute both i and b algorithmically without ever touching a square root.

A clean illustration of a deep mathematical theme. Pick's Theorem is a small, friendly example of a pattern that recurs throughout mathematics: continuous quantities (area, volume, length) being captured exactly by discrete, combinatorial data (counts of points, edges, faces). The same spirit animates Euler's polyhedron formula, the theory of Ehrhart polynomials in modern combinatorics, and even parts of algebraic geometry where "counting lattice points in a polytope" is a serious research tool, not a classroom trick.

The Details

Who found it, and why it took 50 years to catch on. Georg Alexander Pick was born in Vienna in 1859, earned his doctorate at the University of Vienna in 1880, and spent most of his career in Prague, where he eventually became dean of philosophy at the German university there. He published the theorem in 1899 in a paper called "Geometrisches zur Zahlenlehre" ("Geometric contributions to number theory") in the proceedings of a regional Bohemian scientific society — a venue so obscure that the result went almost completely unnoticed for half a century. It wasn't until 1950, when the Polish mathematician Hugo Steinhaus included it (with credit to Pick) in his popular book Mathematical Snapshots, that the theorem reached a wide audience and became a staple of recreational and pedagogical mathematics.

Pick's own story has a haunting footnote. In Prague, he chaired the search committee that in 1911 recruited a young, relatively unknown physicist named Albert Einstein to a chair in mathematical physics — and personally introduced Einstein to the tensor calculus of Ricci-Curbastro and Levi-Civita, mathematical machinery Einstein would need a few years later to build general relativity. Pick retired in 1927 and returned to Vienna. After the Nazi annexation of Austria in 1938, he fled back to Prague, hoping it would be safer. It wasn't: following the German invasion of Czechoslovakia, Pick was deported to the Theresienstadt concentration camp in July 1942 and died there two weeks later, at age 82. A theorem about counting dots on paper, and the man behind it faced one of the 20th century's great atrocities.

Why the formula works. The cleanest proof builds the polygon out of the smallest possible triangles — ones whose only lattice points are their three vertices, with nothing interior and nothing extra on the edges. Every such "primitive" triangle, remarkably, has exactly the same area: 1/2. (This itself is a nice fact: no matter how stretched or skewed, a lattice triangle with no extra lattice points on it always has area exactly one-half.) Any lattice polygon can be triangulated into a collection of these primitive triangles. From there, a bit of bookkeeping with Euler's formula for planar graphs (vertices − edges + faces = 2) converts "how many primitive triangles did we use" into "how many interior and boundary points are there," and the i + b/2 − 1 formula drops out.

Where it breaks. It's natural to ask whether some three-dimensional version holds — surely you can count interior and surface lattice points of a polyhedron and get its exact volume? You cannot, and the counterexample is strikingly simple. British mathematician John Reeve described a family of tetrahedra (now called Reeve tetrahedra) that all have exactly four lattice points — their own vertices — and zero additional interior or boundary lattice points, yet whose volumes can be made arbitrarily large just by stretching one vertex further out along an axis. If "zero interior points, four boundary points" could correspond to any volume at all, no formula based purely on those counts could possibly exist in 3D. The clean 2D miracle simply doesn't survive the jump to a third dimension; the real 3D analogue requires far more machinery (Ehrhart polynomials, which track how lattice point counts grow as you scale a polytope up by integer factors).

A worked example. Picture an irregular hexagon with vertices at (0,0), (4,0), (5,2), (3,4), (1,4), and (0,2) — all lattice points, all legal. Rather than chopping this into triangles and summing trapezoidal areas, just count: suppose careful counting turns up 7 interior lattice points and 6 boundary lattice points (the 6 vertices, with no extra lattice points landing exactly on the edges in this particular case). Pick's formula gives A = 7 + 6/2 − 1 = 7 + 3 − 1 = 9. No trigonometry, no coordinate cross-products — just two counts and an arithmetic formula a middle-schooler can execute by hand.

Takeaways

Pick's Theorem converts the area of any lattice polygon into pure counting: A = i + b/2 − 1, where i and b are interior and boundary lattice point counts.
It's exact, not approximate — and it works for any simple polygon, however irregular, as long as its vertices sit on grid points.
The underlying idea — continuous quantities captured by discrete counts — echoes through computer graphics, GIS mapping, and modern combinatorics (Ehrhart polynomials).
The theorem sat forgotten for 50 years after Pick's 1899 publication, until Hugo Steinhaus popularized it in 1950.
It fails outright in three dimensions, as Reeve tetrahedra demonstrate — a reminder that even the most elegant low-dimensional patterns can break the moment you add an axis.

Resources: - Pick's theorem — Wikipedia - Georg Alexander Pick — Wikipedia