It is true that for observables only 50--60 e-foldings matter. (As pointed out also by Zag). The paper of Liddle & Leach is very good. I only like to comments that if BICEP2's results, of large r, is confirmed, then many inflationary models will be constrained. With large r, soome models give so...