... Do I calculate return using an exponential curve or the S-shaped
logistics curve?
Since the number of respondents is not infinite at present, but is
theoretically unbounded over time, you have to use a sigmoid curve,
but not this logistical sigmoid:
Y = a + b / (1 + exp(-c*(X - d)))
That one is wrong because the number of respondents is constantly
growing, because all those people who have gotten rich can support
large families. Instead, you really want the asymmetrical Gompertz
sigmoid:
Y = a + b*exp(-exp(-c*(X-d)))
IN REALITY: LESS THAN A HUNDREDTH OF A PERCENT OF THOSE WHO SEND
PONZI SCHEME CHAIN LETTERS EVER MAKE A CENT ON THEM, AND MANY OF
THOSE PEOPLE END UP IN COURT FACING SERIOUS CHARGES. PLEASE IGNORE THEM.
Cheers,
James