My alternative approach - log vs log(log) from 01.11.2011 to present. Conclusion: The log-log model seems to better fit the data than the linear-log model, and according to this model we are actually slightly under the trend line. Comments?
Wow ... that's some wild-ass result you got!! ... but hard to argue with the math, the fit looks good (just eyeballing but some best-fit stats would be nice) if we are indeed in the super-exponential phase (vertical) right now.