I am not arguing for or against the merits of the paper.

I am simply pointing out that it provided a range of scenarios that produced a range of estimates and that you are only ever interested in the scenario that produced the highest estimate.

I can but hazard a guess as to why that might be.