Post #3179040

@cascheranno@hachyderm.io @gregeganSF@mathstodon.xyz Hmm.. but by the same token, if two points "near" 4 get picked together, wouldn't you get a line roughly parallel to 4, thus crossing 2 and 3? I have an odd feeling you end up getting a nasty cancellation that makes any pair equally likely 🤔 (edit: actually- I have an almost-proof using barycentric coordinates and the way they interact with ratios of areas and sampled points that shows all pairs of sides are equally likely, but don't trust me before I've written it out)