The paper gives a new proof that a fundamental operation on elliptic curves is associative. An elliptic curve can be thought of as a special kind of smooth curve defined by a polynomial equation. Mathematicians put a group structure on such curves using a geometric recipe called the chord-tangent law: to add two points, draw a line through them, find where it hits the curve again, then reflect that intersection point to get the sum. The tricky part is proving this addition is associative, meaning the order in which you group three additions does not matter. This associativity is what makes the construction a proper group law, and it is notoriously awkward to verify directly.
The authors prove associativity by tracking what happens when you add three points P, Q, and R in two different orders. Each addition requires drawing a line, so two lines appear in the calculation. The central insight is an algebraic identity that packages both lines together and replaces them with a single quadratic curve (a parabola-like object) passing through all four relevant points at once. Because both groupings of the addition lead to the same quadratic curve, they must produce the same final point, establishing associativity. The argument uses polynomial factorization, a formal notion of derivatives that works even when the underlying field has unusual arithmetic properties, and a type of polynomial interpolation that handles the special case where two of the points coincide.
A key strength of the proof is that it works over any field, including those with so-called positive characteristic, such as finite fields used in cryptography. In those settings, the usual geometric intuition can break down, and earlier proofs often required case-by-case arguments or heavy algebraic machinery. Here, tangency is handled through formal partial derivatives of the defining equation rather than through limits or calculus, so the proof remains valid everywhere. The paper also includes a shorter self-contained version for the simpler Weierstrass form commonly used in practice, connecting the approach to an earlier construction due to Zwegers.