Y12W12WR When to stop looking

Design
The writing prompt

Apply the 37% rule to a real sequential search you’re in (or will soon face), and reason through what the rule tells you and where it misleads.

1Retrieval check

Q1.What does the secretary problem (37% rule) recommend?

  • ANever commit
  • BSample ~37% of options without choosing, then take the next option better than all seen so far
  • CAlways pick the first option
  • DPick the last option

Q2.What’s the article’s counter-thread on the 37% rule?

  • AIt has no limits
  • BThe rule optimises for ‘best possible’ outcome; satisficing (settling for good enough) often produces higher reported satisfaction
  • CIt only works for secretaries
  • DThe rule is mathematically wrong
Show answer key

Q1 → B. Sample ~37% of options without choosing, then take the next option better than all seen so far.The rule’s logic is ‘look before you commit, but commit before you exhaust all possibilities’ — a heuristic, not a formula.

Q2 → B. The rule optimises for ‘best possible’ outcome; satisficing (settling for good enough) often produces higher reported satisfaction.Use the rule as heuristic for look-then-decide, not as a formula to optimise the absolute best outcome.

2Prompt deconstruction

Command verb
APPLY the 37% rule to a real search, then REASON through its fit
Must reference
the 37% rule; the article’s caveats about real-world mismatch (options can be revisited, preferences shift)
Must assess
how many of the problem’s assumptions actually hold for your case
Close with
where the rule’s core insight (look before committing, commit before exhausting) still gives useful guidance

3Position nudge

Where on the range does your proposal sit?

Pole A
Pole B

Pole Astrict (take the 37% number as a hard stop)

Pole Bloose (use the rule only for shape, not for numbers)

Commit to a specific point; defend it in your planner.

4Planner — design the thing, then the trade-offs

The search
Named concretely — universities, flats, clubs, friends, something sequential.
Where I am in it
How far through the sequence, and how close to 37%.
What the rule says now
The specific recommendation at your position.
Where the assumptions break
Options you could revisit; preferences that have shifted; comparison noise.
The heuristic I’ll actually use
What you’ll take from the rule and what you’ll ignore.

5Sentence stems

  • My proposal is ___.
  • I am grounding this in [researcher]’s finding that ___.
  • The main trade-off is ___: this design gains ___ but loses ___.
  • The most predictable objection is ___, and my response is ___.
  • I would know it was working after [time] if ___.
  • What I am most likely to abandon is ___, so I will build in ___ to prevent that.

6Exemplar paragraph (not about this article)

(1) My proposal is to apply the 37% rule to my university-course shortlist. I am grounding this in the secretary-problem reasoning and in the article’s caveat that real-world search violates the problem’s assumptions. (2) The main trade-off is that my search is not strictly sequential — I can revisit options in theory, and comparison is noisy because prospectuses don’t align. (3) Applied at my current position (three courses seen out of an intended eight), the rule says to keep sampling without committing, and that matches my instinct. (4) The most predictable objection is that I am using the rule to delay deciding; my response is a pre-commitment: I will stop adding new courses after eight, which forces the rule to hit its trigger. (5) I would know the rule had been useful if I ended up with a choice I could defend without ‘I wish I had looked more.’ (6) What I am most likely to abandon is the strict 37% number, because my preferences are shifting during the search — a shift the secretary problem explicitly ignores. So I will build in one rule that the model doesn’t give me: if my ranking of the first three has changed after the fifth, I restart the count.

What this paragraph does, move by move

  1. Names the search and the framework.
  2. Specifies where you are in the sequence.
  3. Reports the rule’s recommendation at that position.
  4. Identifies the problem-assumption that doesn’t hold.
  5. Pre-commits to a stopping point to make the rule testable.
  6. Adds one rule the model doesn’t give, to handle shifting preferences.