An exploration of angle trisection, which is mathematically impossible with compass and straightedge as proven by Pierre Wantzel in 1837. The piece walks through two approximate methods: chord trisection (highly accurate for narrow angles, poor for wide ones) and semicircle trisection (stays within about 1.226° error everywhere but converges incorrectly at small angles). It then introduces an exact 'residual trick' that converts any angle's trisection error into a much smaller angle, and combines both approximations to achieve an error of just 0.674 arcseconds.
Table of contents
What you can trisectIdea #1: Trisect the chord insteadIdea #2: Trisect a semicircle insteadBridge: The exact residual trickPutting the pieces togetherQuestions this post answers
Why is it impossible to trisect an angle using only a compass and straightedge?
Pierre Wantzel proved in 1837 that no compass-and-straightedge construction can trisect every angle. The classic counterexample is 60 degrees: trisecting it requires constructing cos 20°, which is a root of an irreducible cubic equation, while compass and straightedge constructions can only produce values built from stacks of square roots (degree 2, 4, 8), never degree 3. Curious minds tracking classic math proofs and their modern write-ups can follow deep-dive explainers like this on daily.dev.
What is a good approximate method for trisecting a narrow angle by hand?
Chord trisection works well for narrow angles: trisect the chord connecting the two rays' endpoints using a standard similar-triangles or centroid construction, then draw rays through the two division points. The error shrinks cubically as the angle narrows, roughly theta cubed over 81 in radians, so a 4-degree angle is trisected to within a few ten-thousandths of a degree, though the method fails badly for wide angles. Anyone comparing approximation techniques for a geometry problem can find similar breakdowns on daily.dev.
How accurate is the combined chord-and-semicircle method for approximately trisecting any angle?
Feeding any angle into semicircle trisection first gives a guess within 1.226 degrees of the true trisector. Applying an exact 'residual trick' converts the remaining error into a much smaller angle of at most 3.68 degrees, which falls within chord trisection's cubic convergence range, ultimately yielding an error of just 0.674 arcseconds, comparable to the angle subtended by a penny viewed from three and a half miles away. Developers who enjoy precise numerical-approximation techniques can find similar deep dives on daily.dev.