In 1966, a mid-air collision between a USAF B-52 bomber and an air refueling tanker off the coast of Spain caused more damage than the destroyed aircraft. The bomber carried four atomic bombs. Three were recovered soon afterward on land, but the fourth was nowhere to be found, and presumed to be at the bottom of the Mediterranean.
Intelligence officers convened a group of top mathematicians to predict the landing position of the bomb. The mathematicians placed bets based on their calculations. Then, using Bayesian predictability analysis, the group pinpointed the most-likely positions the bomb could be found. Using the odds they created to assign probability quotients to several key locations, the mathematicians mapped the most probable spots where they believed the bomb would be found, choosing one ravine in particular. A deep sea search discovered the bomb resting in the very ravine where the mathematicians' calculations predicted it would be.