Understanding the 37% Rule for Dating
This tool applies optimal stopping theory — specifically the classic secretary problem — to a dating timeline. The setup: you will meet a sequence of potential partners, one at a time, and you can rank each new person only against everyone you've dated so far. Once you pass on someone, you can't go back. Mathematics gives a provably optimal strategy for maximizing your odds of ending up with the single best-ranked person: reject a fixed fraction of your expected pool outright, then commit to the very next person who beats everyone who came before them.
The formula
If you expect to seriously date N people in total, the optimal number to reject first, r, is the integer that maximizes the classic secretary-problem success formula P(r) = ((r−1)/N) × Σ (from i=r to N) of 1/(i−1). As N grows, the optimal ratio r/N converges to 1/e ≈ 0.368 — the origin of the "37% rule" name — and the resulting success probability also converges to 1/e, about 36.8%. This calculator computes the exact optimal r and exact success probability for your specific pool size, rather than only the large-N approximation.
To translate the candidate count into a real calendar age, the same rejection fraction (r/N) is applied to your dating age window: cutoff age = start age + (end age − start age) × (r/N). Before the cutoff age you're in the "look phase" — date freely, but don't commit. After it, you're in the "leap phase" — commit to the next person who's better than everyone you've dated so far.
How to read your results
"People to reject first" is r, the number of early candidates you should date and learn from without committing. "Cutoff age" maps that same fraction onto your stated age range. "Success probability" is your exact chance, under this strategy, of ending up with the single best-ranked partner out of your full expected pool — not just "a good one," but the very best. "Current phase" compares how many people you've already dated to r and tells you whether you're still gathering data or should be ready to commit.
Assumptions worth stating plainly
The model assumes you can rank every candidate you meet from best to worst with no ties, that you know roughly how many people you'll date in total, that candidates arrive in random order, and that "success" means landing the single top-ranked person rather than any of your top few. Real dating rarely satisfies all of these perfectly — but the 37% rule is still a useful, mathematically grounded heuristic for pacing a search where you can't un-reject someone once you've moved on.