Does anyone know if these laws (formulas) work for VP? My impression
is they don't, but do apply to some gambling games, such as roulette.
This impression is based on the definition of the laws which state the
sample must be from "identically distributed random variables". Lots
of Google sites discuss these formulas.
If they do apply to VP, there are some easy formulas for computing
sample sizes for any desired degree of confidence and level of
accuracy ... eg. the session size required to have 95% confidence of
being within .1% of the computed return.