ERF(x)

Returns an approximation of the error function:

For abs(x) < 1, ERF uses the Mclaurin series:

and iterates until any additional term is less than 1.E-10 times the current solution. In general the series converges within several steps for small x and takes up to 15 terms before converging to the solution for x close to 1.0.

For abs(x) > 1, ERF uses an approximating 9th order polynomial that is accurate to +/- 1.E-7 for all values of x.

See also:

CDF
NORMDIST