Bertrand's paradox is a famous puzzle in probability theory from the 1800s. The original problem asks: if you draw a random chord (a straight line connecting two points) across a circle, what is the probability that the chord is longer than the side of an equilateral triangle inscribed in that circle? The paradox arises because there are at least three natural ways to define "random chord," and each method gives a different answer: 1/2, 1/3, or 1/4. This seems troubling because there should only be one correct answer to a well-posed probability question. The paradox highlights that the word "random" is ambiguous unless you specify precisely how the randomness is generated.
This paper creates a discrete version of the same paradox. Instead of working with infinitely many possible chords, the authors build finite collections of chords and then pick one uniformly at random, meaning every chord in the collection has an equal chance of being selected. This is a cleaner mathematical setting because uniform probability over a finite set is completely unambiguous. The key question becomes: do the three classical methods, when translated into this finite discrete setting, still give different probabilities, or do they converge to the same answer?
The authors find that the discretization preserves the paradox: the three methods still yield different probabilities even in the finite setting, mirroring the original continuous case. As the number of chords in each collection grows larger, the probabilities from each method approach the same limiting values that Bertrand originally calculated. This work helps clarify that the paradox is not just an artifact of dealing with infinity or continuous probability, but reflects a genuine ambiguity in what it means to choose a geometric object at random, an ambiguity that persists even in simple, finite, discrete situations.