Cherry-picking

Survivorship bias

Plotting only the planes that made it home.

a.k.a. the missing dead

If the dataset only contains winners — funds that still exist, companies that went public, planes that returned — every trend tilts rosy. The bullet holes you can see are the ones that didn’t matter.

How to spot it

Ask what something had to survive to end up in the dataset.
Suspiciously smooth success curves for risky endeavors.

The fix

Include the drop-outs, delistings, and wrecks — or state plainly that you can’t.

In the gallery

Line chart titled “Why your cat wants the top floor” plotting average injuries per cat against stories fallen; the line climbs from 0.76 injuries at 2 stories to a peak of 2.4 at 7–8 stories, then a red arrow plunges to 1.1 at 9–32 stories, beside a cartoon cat asking for “Top floor, please” and a caption reading “Higher is safer!”

The cat-fall chart that only counted the survivors

Survivorship bias

Bar chart titled “Where our bombers are getting hit” showing bullet holes per square foot on returning planes — rest of plane 1.80, fuselage 1.73, fuel system 1.55, engines 1.11 — with a red bracket over the two tallest bars labelled “add armor here”

The bomber-armor chart that pointed at the wrong parts

Survivorship bias