Extrapolation is easiest when you can cast a straight line through data points. To extrapolate from data that define a distribution, it would be nice to have a straight line or at least a line through the data points that smoothes out the noise and allows projecting past the data. There is special graph paper that will handle distributions quite nicely. When the distribution is truly normal, the data points will define a straight line. The graph paper has either a log scale or an arithmetic scale for the ordinate and has probability on the abscissa. We get the information for the abscissa from areas under the probability curve.
