Well, my example was just that...a very bad example. Your correct that trying to extrapolate from such a broad range given such a small range will result in very unrealistic data.

My real world situation is quite different, however. I'd post the data, but it violates an NDA agreement that I signed, so I'll refrain. Just note that I have plenty of data to generate a nice curve and mathmatically, I'd think predicting further data points would be pretty accurate...much more so than picking a repeating value and running on a linear scale.

Thanks!
J