In WWII, the US Army Air Force wanted to reinforce planes against enemy fire. They studied bullet hole patterns on returning bombers and proposed armouring the most-hit areas.
Abraham Wald, a statistician, pointed out the opposite: the bombers they were studying had survived. The undamaged areas on returnees were where bullets killed planes — those got shot down and weren't in the sample. Armour the engines and cockpit, the unscarred regions on survivors.
The Army Air Force adopted his recommendation. It's the foundational example of survivorship bias in decision-making.