• Define a Black Swan by its three marks, and explain why it is relative to the observer: a surprise for the turkey is a plan for the butcher.
  • Sort any domain into Mediocristan or Extremistan by asking one question: can a single observation move the total disproportionately?
  • Recognize the turkey problem in a long, calm track record, and check which kind of randomness produced it before you trust it.
  • Tell “no evidence of Black Swans” apart from “evidence of no Black Swans”, and know where confusing the two gets expensive.
  • Catch the narrative fallacy, memory's rewrites and the pull of anecdotes, and counter them with deliberate habits like a diary and a written base rate.
  • Ask who never made it into the sample before you trust a track record, a success story or a dataset.
  • Judge when a game, an exam or a clean model is a fair stand-in for a real decision, and where it stops fitting.
  • Calibrate your own forecasts with honest ranges, and trust experts by whether their field is learnable, not by their credentials.
  • Read a plan's range before its average, because the surprises in projects push cost and time in one direction.
  • Explain Taleb's barbell, a truly safe majority plus a small speculative part and nothing in the middle, and why it fails if the safe leg is not safe.
  • Explain where the bell curve is legitimate, and why fractal models turn some Black Swans gray but never all of them.
  • Find the fourth quadrant in your own decisions, and change the exposure instead of modeling the risk harder.