Hi simplex,

Appreciate your formula, however I'd like to say this is not the general equation, because delta only equals to 19 in this specific case, and the 20 is not derived from delta + 1, its origin is from the same book page 13, that 20 is a constant in this equation, if the delta changes your formula could be wrong.

the correct equation should be K = 10^(-3/(20*(longP-shortP))

no offence, just want to help

I know this topic is old, but it seems worth mentionning...

The zorro examples in this thread all run the prices through a roofing filter AND THEN through an automatic gain control (AGC) before calculating the correlation coefficients.

Ehlers code on the other hand, doesn't use an AGC at that point.

Besides wouldn't the AGC mess up some of the correlations in the data?