Question about a simple game

[Albert Chan]

Hi, Namir

Your different results probably due to quality of the random generator.
This is especially true here since we are subtracting two random outputs.

Also, many random generator have low linearity complexity issues.
That meant it is ok to use the randoms as *values*, but not in pairs.
The pairs may not be independent enough to give good result.

This also meant jamming random bits to make bigger random numbers is a bad idea.
see http://www.azillionmonkeys.com/qed/random.html

My patched Lua were using xoroshiro128+ PRNG, with decent randomness properties.
Just to be sure, I redid the simulation using *two* random generators:

With 5000 runs, ME = xoroshiro128+, PC = Mersenne Twister, results are similar:

For T=40, N ≈ 6.12 T^2
For T=1, N ≈ 9.26