[EM] Naive random UWA, timeshare, etc.

[Forest Simmons]

Here's an idea for turning random ballot into a deterministic method that
is equivalent for all practical purposes, but allows checking of the
result by independent parties:

Suppose that there are three candidates and that their respective "random
shares" are proportional to  3:7:9, respectively, from smallest to
greatest.

Form the numbers n=3+7+9=19, and N=3^2+7^2+9^2=139, and then calculate the
number N mod n, which is the remainder associated with the quotient N/n .

In this case 139/19 = 7 with a remainder of 6.

Represent the proportion 3:7:9 by ...|.......|........., and count six
places from the left.

We end up in the second candidate's region, so the second candidate is the
winner.

Forest