Best-Guess (Hypothesis Enumeration)

NASA Goddard Space Flight Center - "New Evidence for Water at Mercury's Poles"

When you come up with an interesting hypothesis:

  • Look at the field where it applies.
  • Look at what assumptions the hypothesis breaks.
  • Enumerate all possible hypotheses that emerge from those broken assumptions.
  • Look at why your hypothesis is the best of these options.
  • Why not test the other ones too?

The best hypothesis can then be selected for priority of testing via:

  • Testability
  • Parsimony
  • Scope
  • Fruitfulness, and…
  • Conservatism

This approach can’t apply to all situations since not all hypotheses are of the type that extend a field, but I believe that there be similar enumeration principles for other types of hypothesis too. While this may seems obvious, I’m interested how often we put on blinders when we develop an interesting idea in the light of new contradictory evidence against prevailing wisdom. At this point it would be very useful to stop, take a breath, and enumerate.

To be clear - This is not a correctness measure. It’s all about optimisation.