User blog comment:Hyp cos/SCG(n) and some related/@comment-10429372-20140812170828/@comment-10429372-20140812181853

Yes, I know, there comes the 101.17*10 21135 from. I analysed it up to that point, but I lost my analysis. (see above) and I didn't want to do the whole thing again.

Up to reducing 5-paths, we have only small divisions and subtractions only count for the first few. So the bound still holds up to that point. But than we have a large value, around 2^^^2^^2^^2^^2^^2^^2^^63.

So I believe the bound holds with the normal FSes (never use 's as plural in English), beacause I rounded the 101.17*10 21135 down.

Anyway, note that I stated in the article just f_e_2, though f_(e_2+e_0*2) certainly isn't unreasonable.