Critical Thinking

Fast opinions are cheap; calibrated judgment is earned. Slow down where stakes are high, steelman what you disagree with, and keep score on your own accuracy.

When to use

• Evaluating claims, proposals, articles, pitches — before repeating or acting on them.
• Disagreements where you feel certain.
• Reviews and retros: your own reasoning included.

Method

1. Claim → evidence: restate the claim precisely; what evidence would support it, what would refute it? Is any actually present?
2. Evidence hierarchy: measurements/data > documents/logs > expert synthesis > hearsay; note conflicts and why they exist (incentives!).
3. Base rates first: what usually happens in cases like this? Reference-class thinking beats compelling stories.
4. Steelman the other side: strongest version of the opposing view; if you can't build it, you don't understand it yet.
5. Fallacy spot-check: ad hominem, straw man, false dilemma, appeal to authority/antiquity, survivorship stories, motivated reasoning, cherry-picked windows.
6. Calibration habit: state confidence ("80% sure"); review outcomes; track hit rate by confidence band — most find they're overconfident.
7. Second-order thinking: and then what? Effects on incentives, timing, competition, future options.
8. Write the judgment: conclusion + confidence + key uncertainties — short, checkable later.

Quality gates

• Claims restated precisely; supporting AND refuting evidence sought.
• Base rates / reference class used where applicable.
• Strongest counterargument engaged (steelman, not strawman).
• Confidence stated; outcomes reviewed for calibration.

Red lines

• No conclusions from vibes or authority alone; no straw-manning in your own favor; don't let certainty outrun evidence.

Related

analytical-reasoning · decision-making · agent-evaluation.