HomeArticlesEuler Characteristic

Euler's Formula: The Number That Survives Every Polyhedron

Why V minus E plus F always equals 2 for a sphere-like solid, how it proves there are only five Platonic solids, and what changes to 2 − 2g on a donut.

mysimulator teamUpdated June 2026≈ 8 min read▶ Open the simulation

A count that never changes

Take any convex polyhedron — a cube, a tetrahedron, a soccer-ball-shaped truncated icosahedron — and count its vertices V, edges E and faces F. Compute V − E + F. Leonhard Euler noticed in 1750 (with an earlier trace in Descartes' unpublished notes) that the answer is always 2, for every convex polyhedron, regardless of how irregular its faces are or how many of them there are. A cube gives 8 − 12 + 6 = 2. A tetrahedron gives 4 − 6 + 4 = 2. A soccer ball gives 60 − 90 + 32 = 2. This is not a coincidence of small numbers; it is a topological invariant, and it is the first and most famous example of a whole field — algebraic topology — that studies which numbers stay fixed no matter how you deform a shape.

live demo · rotate eight polyhedra while V, E, F and V−E+F update live● LIVE

Why it works: the planar-graph proof

The cleanest proof does not touch three dimensions at all. Puncture one face of the polyhedron and flatten the rest onto the plane like a net (this is always possible for a convex solid) — you get a connected planar graph with the same V and E, and F − 1 bounded regions plus the one unbounded outer region standing in for the punctured face. Now delete edges one at a time, each deletion either merging two faces (F drops by 1, E drops by 1, V−E+F unchanged) or, if the edge is a dead-end (a 'bridge'), removing a vertex instead (V drops by 1, E drops by 1, V−E+F unchanged). Keep deleting until only a single vertex remains: V=1, E=0, F=1, so V−E+F = 2. Since every single deletion step preserved the quantity, it must have equalled 2 at the very start too.

Platonic solids fall out of the same formula

Euler's formula is also the fastest way to prove there are exactly five Platonic solids — convex polyhedra where every face is the same regular p-gon and every vertex joins the same number q of faces. Each face has p edges shared between 2 faces (E = pF/2), and each vertex touches q edges shared between 2 vertices (E = qV/2). Substituting V = 2E/q and F = 2E/p into V − E + F = 2 gives:

2E/q  -  E  +  2E/p  =  2
=>   1/p + 1/q  =  1/2 + 1/E   >  1/2

only five integer pairs (p,q) with p,q >= 3 satisfy this:
(3,3) tetrahedron   (4,3) cube        (3,4) octahedron
(5,3) dodecahedron  (3,5) icosahedron

Any other combination — a hypothetical solid made of regular hexagons meeting three at a vertex, for instance — makes 1/p + 1/q equal to exactly 1/2, which forces E to be infinite: it is not a solid, it is the flat hexagonal tiling of the plane. Go past that and the sum drops below 1/2 entirely, which is exactly where hyperbolic tilings live instead.

What changes on a donut: genus and χ = 2 − 2g

The formula 2 is specific to shapes topologically equivalent to a sphere — anything you can inflate into a ball without cutting or gluing. A torus (a donut, a coffee mug, anything with exactly one hole) always gives V − E + F = 0, not 2, no matter how it is triangulated. A two-holed torus gives −2. In general, for a closed orientable surface of genus g (its hole count), the Euler characteristic χ = V − E + F = 2 − 2g. This single formula is why a coffee mug and a donut are considered 'the same shape' in topology — both have genus 1, hence χ = 0 — while a sphere (genus 0) and a pretzel (genus 3, χ = −4) are provably, permanently different, no matter how you stretch or squash them, as long as you never tear the surface or glue new parts of it together.

Star polyhedra break the rule — on purpose

Self-intersecting solids like the Kepler-Poinsot star polyhedra do not satisfy V − E + F = 2, and that is not an error, it is the whole point of studying them: their faces or vertex figures wind around the centre more than once, so the simple planar-graph argument above (which relies on the polyhedron being embeddable in the plane without crossings) no longer applies. Computing the correct generalised Euler relation for those solids requires tracking the density of the winding, a topic the Kepler-Poinsot article on this site picks up directly.

Frequently asked questions

Does V − E + F = 2 work for every polyhedron?

Only for polyhedra topologically equivalent to a sphere — solid, non-self-intersecting shapes with no holes through them. A torus-shaped polyhedron gives 0, a two-holed one gives −2, and self-intersecting star polyhedra like the Kepler-Poinsot solids need a more general, density-aware version of the formula.

Why are there only five Platonic solids?

Because the constraint 1/p + 1/q > 1/2 (from combining Euler's formula with the requirement that every face and every vertex look identical) has only five integer solutions with p, q ≥ 3: the tetrahedron, cube, octahedron, dodecahedron and icosahedron. Any other regular combination either tiles the flat plane forever or requires hyperbolic space.

What is the Euler characteristic of a sphere versus a donut?

A sphere (genus 0) has χ = 2. A donut, or anything topologically equivalent to it such as a coffee mug (genus 1), has χ = 0. The general formula is χ = 2 − 2g, where g counts the number of holes — it is one of the most basic invariants that tells two surfaces apart without ever measuring distance or angle.

Try it live

Everything above runs in your browser — open Euler's Formula and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.

▶ Open Euler's Formula simulation

What did you find?

Add reproduction steps (optional)