[daip] Re: POSSM with BPARM normalisation

[Eric Greisen]

You have not followed the rather simple argument:

If you take an unsmoothed data set and

1. Smooth the data and then apply a BP from no smoothing

2. Smooth the data to make a BP table but apply that to unsmoothed
data

You get more or less what you deserve - a poor result.

You must either smooth everything or not smooth everything.

I will change the code - non-trivial - to allow smoothing before or
after the BP application, but you are the first in 15 years to
complain of this so I take it that it is not urgent.

Eric Greisen