MBBEFD cumulative loss distribution should be used in your MBBEFD example, not G(x) exposure curve.

[hdavid333]

I think your MBBEFD example does not work. Seems it should use CDF F(x) of MBBEFD, not exposure curve G(x). Because I was unsure of this, I created aggregate that should be completely fixed at exactly 1 claim per exposure, close to 1.000 loss intensity, E(X). This worked as expected using CDF F(x), and not as expected using G(x) provided in your example.

I tried to test fix completely in build of .rst, but don't know how to build .rst. I did test however via copy paste of the code chunks.

Demonstration of test is in html files of "issue-raised" folder of MBBEFD-edit branch in my copy of the repo.

Getting right loss degree using MBBEFD CDF, F(x) https://htmlpreview.github.io/?https://github.com/hdavid333/aggregate/blob/MBBEFD-edit/issue_raised/agg%20MBBEFD%20using%20CDF.html

Not getting right loss degree using exposure curve, G(x) https://htmlpreview.github.io/?https://github.com/hdavid333/aggregate/blob/MBBEFD-edit/issue_raised/agg%20using%20G%20(which%20does%20not%20seem%20correct%20as%20used).html