My first version of a marathon performance predictor. The height of the curve gives the probability of a marathon time given the input values. Note that the distribution is asymmetric, as a much faster time is less likely to occur than a much slower than predicted time.