for games i've looked at, covariance tends to be 10-12% of variance.
it's consistent enough that i generally don't bother computing it, i
just stick in some vaguely round figure in that range and worth with
that.
That's probably true for games that uses E. Moody's patent (e.g. N-
Play, spin-poker, etc); the 10-12% number isn't true for IGT's other
multi-line video poker patents.
i was referring to n-play, which was what i believed the original
poster to be asking about.
I know this to be the case for one
patent since each line has no covariance with the other lines.
are you talking about multistrike? if not, i'm curious what you're
referring to. a patent number would be fine, i'm not asking for any
sort of privileged information.
cheers,
five