Maths/Stats challenge - 1 of 2 - polls

[robve]

(05-23-2022 11:01 AM)EdS2 Wrote:  Here's a challenge. I think it's a bit like finding fractional approximations to arbitrary numbers, but different.
Given a poll result, come up with a series of ever-better approximations of how many votes might have been cast.

My guess is that Euclidean method applied to 10^n upscaled percentages (to remove fractions) and sum over normalized is one possible step to get an answer, but for ever-better approximations?

- Rob