Where Would You
Add the Armor?
It's 1943. Bombers keep coming back from missions riddled with bullet holes. Military engineers want to bolt armor onto the places that got hit most. Before you scroll further, would you have agreed?
1943 · Bomber Damage Report
Where would you add armor?
Every dot below marks a real bullet hole, plotted from bombers that made it back from their missions. Based on this pattern, tap up to three places on the plane where you'd bolt on extra armor.
Act two
Run the actual numbers
Bullets don't aim for the survivable spots. Assume every zone on the plane is just as likely to get hit as any other. What changes everything is how likely a plane is to make it home once it's hit there. Drag the slider and watch what the "observed" damage report looks like once only survivors get counted.
Simulates 2,000 virtual planes, each taking one hit. Only the ones that make it home get counted in the "observed" report, exactly like the real one engineers studied in 1943.
The data was only telling half the story
This really happened. In 1943, the US military asked statistician Abraham Wald to help decide where to add armor to bomber aircraft. Engineers had plotted bullet holes from planes returning from combat, and the pattern was clear: wings and tail took the most hits. The obvious move was to reinforce those spots.
Wald said the opposite: reinforce the places with the fewest hits, especially the engine and cockpit area. His reasoning was simple once you hear it. The diagram only included planes that survived long enough to land. A hit to the wing was survivable, that's why it showed up on the chart. A hit to the engine or cockpit usually was not, so those planes never made it back to be counted at all.
The missing bullet holes weren't good news. They were the most important data point in the entire report, and it was invisible because it belonged to planes nobody could examine.
This pattern shows up everywhere once you know its name, survivorship bias: judging a group by only the members who made it through some filter, while the ones who didn't stay invisible.
- Business books study companies that succeeded and draw lessons from them, while the far more numerous companies that tried the exact same strategy and failed quietly disappear from the sample.
- "Every successful founder dropped out of college" ignores the much larger number of dropouts who did not become successful founders.