# Regression fallacy

*Also called: Regression to the mean ignored*

- **Cluster:** Causal errors — Sequence, coincidence and hindsight read as cause
- **Page:** https://fallacy.wiki/fallacy/regression/
- **In Arabic:** https://fallacy.wiki/ar/fallacy/regression.md
- **Licence:** CC BY 4.0 — https://creativecommons.org/licenses/by/4.0/

## What it is

Crediting an intervention for an improvement that was a return to normal after an extreme.

## In the wild

> He improved after the pep talk, not because he was having an unusually bad run.

## How to answer it

Ask what the baseline was and whether the starting point was unusually high or low.

## Origin and naming

Named for the statistical phenomenon identified by Francis Galton in the 1880s: measurements taken after an extreme value drift back toward the average on their own. The fallacy is to hand that drift to whatever happened in between — a lecture, a remedy, a punishment.

## More places it shows up

- The coach screamed at the team after their worst loss; they won the next game, so screaming works.
- My back pain was at its absolute worst when I booked the chiropractor, and a week later it was much better.
- We sent the underperforming branch to a training seminar and their numbers improved the next quarter.
- The stock plunged, the CEO gave a defiant interview, and it rebounded — the interview restored confidence.

## How to spot it

- The intervention always lands on a peak or a trough — nobody tries the remedy on an ordinary day.
- Improvement is measured from the worst point, not from a baseline.
- The same 'cure' keeps being rediscovered every time things are unusually bad.

## Nearby

- [Post hoc ergo propter hoc](https://fallacy.wiki/fallacy/post-hoc.md)
- [Hindsight bias](https://fallacy.wiki/fallacy/hindsight.md)
- [Anecdotal evidence](https://fallacy.wiki/fallacy/anecdotal.md)
- [Gambler's fallacy](https://fallacy.wiki/fallacy/gamblers.md)

## Read more

- [Wikipedia](https://en.wikipedia.org/wiki/Regression_toward_the_mean)

---

From fallacy.wiki — a bilingual field guide to 91 logical fallacies, published by phi holdings. Licensed CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/): reuse freely with attribution to fallacy.wiki.