Nice work last year, and a good prediction too:

I have bookmarked. It's a similar method to the ones mentioned by dalz and joseph I think.

Estimations are OK, is there really no accurate way to plot SP from day 1?