Tapir

Dear Maple Experts,

  I'm fairly new to Maple.  I've been trying to compute Gaussian tail probabilities accurately.  As I understand it, Maple has an "erf" function, but not an inverse erf.  Other posts have suggested doing something like:

  inv_erf := x->solve(erf(y/sqrt(2.0))=x,y);

At first ; for instance, about 68% of a Gaussian's probability is within 1 standard deviation, so:

> erf(1/sqrt(2))

    0.6826894920

> evalf(inv_erf(0.6826894920));