• $f(v)=-dF(u)/du$
  • $f(1)=f(-1)=0$
  • $F(-1)=F(1)$
  • $g(v):=2(F(v)-F(-1))=-2\int_{-1}^v f(u)\, du$,
  • $\mu^2:=-f'(-1)=-f'(1)$

In particular, $g'(v)=-2f(v)$ and $g(1)=0$.

Now, it is claimed that by the definition of $g$ it is "quite standard" to get the estimate $$ f(v)g^{-1/2}(v)=\mu(1+O(1-v)), $$

And, to be honest, I have literally no idea how to see this!

  • That's not true as is, the leading order estimate of $g^{-1/2}$ will have a $\mu^{-1/2}$ in it that will spoil that result. Did you maybe want the expansion of $g$ to not involve $\mu$? – Ian Nov 8 at 18:30
  • Sorry, I forgot to mention extra information which I now added. – Salamo Nov 8 at 18:56
  • Still exactly the same issue as before. – Ian Nov 8 at 19:40
  • You are right, I now added the whole context. Sorry for the chaos. – Salamo Nov 8 at 19:40
  • It seems to be assuming $f'(1)$ and hence $g''(1)$ are nonzero, while $g'(1)=0$. Thus g scales like $(1-v)^2$. Multiply and divide by $(1+v)^{-1}$ to get to $(1-v)^{-1} f (g (1-v)^{-2})^{-1/2}$ and then estimate g. – Ian Nov 8 at 19:54

Your Answer


By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.