Brian;

It isn't necessary.

I evaluate whatever is predicted. If the predictor wants an archery target (like JOB) that's fine. Nearly everyone specifies a range for all 3 parameters.

I wrote a program called BESTFIT that selected the 4 best fitting quakes, one for each parameter plus one for probability, to be used as a training device, showing what they should have predicted.

No one was interested.

Roger