> > N(x) := erf(x/sqrt(2))/2. > > that should be N(x) := (1 + erf (x / sqrt(2)))/2 $ if i'm not mistaken. > or N(x):= (1+erf( (x-m) / (s * sqrt(2) ) )/2 if you want to include the mean m and the standard deviation s>0. -- -------------------------- Mario Rodriguez Riotorto www.biomates.net