Hi Spy128Bit -

Great to hear about your results. Attacking a problem with different
approaches may be a bit redundant, but the payoff is that we get to learn
something about how specific techniques contribute to performance.

On the subject of performance, I'm now curious about the difference in
performance between our PC's. I'd be very interested in testing the same
data on both. I'm guessing that our data values are extremely different, but
at the same time, the procedure as constructed should be somewhat immune to
that. It would be valuable to understand this by running a side-by-side
comparison.

If you'd feel comfortable emailing a data worksheet along with the
processing time, I'd be grateful. It could be dummied-up if you have
sensitive data. Alternatively, I could email mine. It's up to you if you
want to do this and I understand if you don't; no explanation necessary. My
adrs is jc .sresearch a t hotmail_com (modify to look like a standard email
address).

One additional note. Credit to sali for the overlap function. It distilled
the overlap calculations down to a few short lines which were easily
incorporated into the procedure.

----
Jay