One miss tells you nothing. Fifteen might.

Back of the Envelope · Week 12

70%, and right

Compatible with excellent judgment. Compatible with terrible judgment. Compatible with chance.

One result cannot tell the three apart.

Open the one that has been sealed

Take out your Week 10 forecast — Canberra Airport’s daily maximum, 20 May 2027.

Check it against the exact product you named: the Bureau of Meteorology’s Daily Weather Observations for Canberra, ACT.

Record what it shows, honestly, before touching anything else in the diary.

Fifteen rows, misses marked

(illustrative structure — not real diary data)

  1. 75% — hit
  2. 60% — hit
  3. 80% — miss
  4. 75% — hit
  5. 90% — hit
  6. 65% — miss
  7. 75% — hit
  8. 70% — hit
  1. 85% — hit
  2. 75% — miss
  3. 60% — hit
  4. 80% — hit
  5. 75% — hit
  6. 90% — miss
  7. 70% — hit

No formula — a plain count, laid out so the shape is visible at once.

A pattern is not “I was often right”

Being right does not need explaining. A pattern is something like:

  • always missing on the same side
  • most overconfident about quantities read somewhere, not built
  • widening every range that involved someone else’s number

Name the actual shape — not a summary of how the semester felt.

Now go looking for the entry that ruins it

Confirming a pattern and searching for its exceptions use the record in opposite directions. Doing only the first produces a claim that only ever meets evidence that agrees with it.

Then swap diaries. Your partner reads for what would undermine your claim — not for whether it sounds right.

The method, aimed at its own author

Week 4 said to keep the sheet, because the line naming what would make your range miss “turns out to be worth something in week 12.”

This is that week — the same discipline, turned for the first time on the person who has been doing the estimating all along.

← Back to lecture