When a financial model produces a number you don't like, the tempting move is to adjust an assumption until it does. This isn't analysis — it's ventriloquism. The 14th-century North African historian Ibn Khaldun noticed that court scholars reliably produced the conclusions their patrons wanted, not because they were corrupt, but because prolonged proximity to power reshapes what counts as 'reasonable' without anyone noticing the drift. Modern portfolio managers do the same thing: the discount rate nudges upward by half a point, the terminal growth rate stays stubbornly optimistic, and the model now confirms the deal you already wanted to do. The philosopher of science Imre Lakatos called this 'protective belt' reasoning — the practice of shielding a core belief from falsification by endlessly adjusting surrounding assumptions rather than questioning the belief itself. The discipline Lakatos was pointing toward isn't skepticism about your models; it's the habit of designating, in advance, which assumptions are genuinely load-bearing and then treating those as non-negotiable. Decide before you build what would have to be true for you to kill the deal — and write it down.

In the last project or deal you modeled, which single assumption, if reversed, would have changed your recommendation — and did you ever actually test it?

Drawing from Philosophy of Science (Lakatosian methodology) — Imre Lakatos (The Methodology of Scientific Research Programmes, 1978, on the 'protective belt' of auxiliary hypotheses)