ANALYSIS TOOL BOX

Analysis of Competing Hypotheses (ACH)

Test competing explanations against your evidence instead of gathering evidence for a favorite. Rate each evidence item against every hypothesis, watch inconsistency scores rank the field, and save versions as new evidence arrives — the least-inconsistent hypothesis wins, with an auditable trail of how your judgment evolved.

How this technique works
Learn the technique

About this technique

What is Analysis of Competing Hypotheses?

Analysis of Competing Hypotheses (ACH) tests several explanations against the same body of evidence at once. Richards Heuer developed it at the CIA in the 1970s and published it in Psychology of Intelligence Analysis (1999). Its central idea runs against instinct: rather than collecting evidence that supports a favored explanation, you look for evidence that is inconsistent with each one. The hypothesis with the least evidence against it survives.

When to use it

Use ACH when a question has several plausible explanations and the stakes of picking the wrong one are high. Typical cases are attributing a cyber intrusion, explaining an unexpected policy shift, or judging whether reporting is genuine or deception. It works best once you have a real set of evidence to weigh, and after a Key Assumptions Check has tested what you're taking for granted.

How to use this tool

  1. Write the analytic question and list every reasonable hypothesis as a column. Keep each one short and falsifiable.
  2. Add evidence and arguments as rows, with a source citation for each.
  3. Rate each cell from Very Consistent to Very Inconsistent, or Not Applicable. Work evidence-first: rate one item against every hypothesis before moving on. The next-cell button follows that order.
  4. Give a one-line rationale for each rating so the matrix stays defensible.
  5. Note assumptions and deception risk on evidence that rests on either.
  6. Read the inconsistency scores. The lowest score is the strongest candidate, not a probability.
  7. Add diagnostic indicators that would shift the picture, and save a version before new evidence arrives.

Common pitfalls

  • Counting support instead of contradiction. Evidence consistent with every hypothesis doesn't help you choose. The tool marks those rows.
  • Scoring hypothesis by hypothesis, which invites confirmation bias.
  • Leaving out a hypothesis, especially deception or "none of the above."

Related