Congrats Philadelphia

[Yes_Probably_Maybe_No]

It's actually not going to work for a 30 year sample set.

E(x) = (1/n)*sum(x)

E(x_i + x_i+1) = (1/(n-1))*sum(x_i + x_i+1)

If n=30, E(x) = 1/30*sum(x)

but E(x_i + x_i+1) = 1/29*((x1+x2) + (x2+x3) + (x3+x4) ... + (x28 + x29) + (x29 + x30))

=1/29*(30*2*xbar - x1- x30) *[note: sum(x) = n*xbar]*

So there is a bias because of the way we are sampling the data. But as n -> infinity, E(x_i + x_i+1) -> 2*E(x)

Wheeeeeeeee.

Q.E.D

I decided to do it the long way so I can see it and it ends up being pretty much the same over a two year period. Now that I look at it the longer way (with less "x"s and "xbar"s), you were right in that it should be the exact same thing.

[Rainshadow]

I've completely killed this thread by being a total nerd, haven't I?

I'm going to have to complain to a mod about this thread being hijacked.