SMOOTH <data>,<series>,<smoothened_series>,<width> smooths <series> in <data> using a technique based on FFT (Fast Fourier Transformation) with a window whose full width is of order <width> neighboring points and saves the smoothened values as <smoothened_series>. SMOOTH can be used for smoothing any variable Y in relation to another variable X by first sorting the data by X and then applying SMOOTH for Y. Numerical method: SMOOTH removes any linear trend and uses then FFT to low-pass filter the data. The linear trend is reinserted at the end. <width> gives the "amount of smoothing", specified as the number of points over which the data should be smoothened (not necessarily an integer). Zero gives no smoothing at all. The results of SMOOTH are generally in accord with the notion "draw a smooth curve through these scattered points". Reference: "Numerical Recipes" (1987) by Press, Flannery, Teukolsky, and Vetterling V = More information on transformations S = More information on statistical operations