The Y axis plots the actual values. The X axis plots predicted values. To calculate
these predicted values, first assume the distribution is Gaussian (with the mean and
SD of the data), and then select n points (the number in the data set) from that
Gaussian distribution with equally spaced percentiles. For example, if there are five
values in the sample choose values from the Gaussian distribution at the 10th, 30th,
0 5 10 15
0
5
10
15
Predicted Data
Actual Data
2
50th, 70th and 90th percentiles. Each predicted value becomes a Y value matched by
rank with one of the actual values which is the corresponding X value. This
algorithm is explained in more detail below.