My favorite example of survivorship bias is the US Navy's statistical analysis of returning WW2 planes to determine what parts of the planes received the most damage from anti-aircraft fire.
The idea was to selectively add armor to the areas most likely to get hit.
But by analyzing returning planes, their sample excluded the planes that were lost.
So the better strategy was to add armor to the areas with the least damage instead.
The idea was to selectively add armor to the areas most likely to get hit.
But by analyzing returning planes, their sample excluded the planes that were lost.
So the better strategy was to add armor to the areas with the least damage instead.
https://en.wikipedia.org/wiki/Survivorship_bias#/media/File:...
Now I see survivorship bias everywhere.
Retail chains improve their comparable store sales metrics not by actually increasing sales, but by closing stores with declining sales.
Money managers and traders with the highest performance often have no special talent, they're just ones who survived.