pub fn erfc<F>(x: F) -> Fwhere F: Float,
1 - erf(x), without the loss of digits for large x that 1 - erf(x) has.
1 - erf(x)