Hi John. This is an excellent tool! I have a question. The .rtf accompanying the file lists 'prescription.PredictionUncertainty' as an option. The file itself, however, does not. Is there any way to tease out the uncertainty/confidence interval for the 'x' value chosen for a free knot? The verbose output lists 'Range of prediction errors', but this appears to be in 'y'. I'm using a linear 3-knot fit.