Count the assumptions each explanation requires

When two explanations fit the facts, count how many unverified assumptions each one rests on.

Why it works

Each assumption in an explanation is an independent place the explanation can fail. An explanation with five assumptions has five places where the chain can break; one with two assumptions has two. Counting assumptions makes probability differences concrete — independent errors multiply, so the gap between explanations widens quickly.

How to do it

1. Write out the competing explanations as explicit chains: "this is true IF A, B, and C."
2. Count the unverified assumptions in each chain.
3. Ask which assumptions are actually checkable with current evidence.
4. Favor the explanation with fewer unverified steps while remaining open to update.

Evidence

The parsimony preference is foundational in philosophy of science (Sober, Quine) and is encoded in Bayesian probability: simpler models receive higher prior probability because they make fewer independent bets. The formalization is well established in statistical model selection (AIC, BIC penalize unnecessary parameters). (mechanistic)

Parsimony is a prior, not a guarantee. Nature is sometimes genuinely complex; simpler explanations can also be wrong. The razor guides starting points, not endings.

Common mistake

Treating simplicity as a trump card regardless of evidence — dismissing a complex but well-supported explanation because it feels complicated.