Where would you armor the bomber?
Survivorship bias: the data you can see may be pointing the wrong way.
It's 1943 and your bombers keep getting shot down over Europe. Armor is heavy, so you can add only a little, and it has to go where it matters most.
Your crews mapped every bullet hole on the planes that returned to base. You have 4 armor plates. Tap the plane to place them.
Armor where the holes aren't
Most officers wanted armor where the holes clustered: wings, mid-fuselage, tail. The statistician Abraham Wald said to armor the engines and cockpit, which had almost no holes.
The data only covered planes that made it home. A plane shot through the wings could still fly. A plane hit in the engine was at the bottom of the Channel. The clean spots on the survivors were the fatal ones.
Before you trust a dataset, ask what got filtered out on the way in.
Where else this shows up
- Founder advice. "We ignored the doubters" comes only from companies that survived. The failed ones ignored the doubters too.
- "They built things to last back then." The flimsy old buildings fell down.
- Fund performance. Indexes drop funds that closed, which flatters the average.
- Transformation photos. You see the one person a routine worked for, and none of the thousand it failed.