|
|
In 1966 a B-52 bomber crashed near a fishing village in Spain. Three of the nuclear bombs carried by the plane
were recovered in a field. The fourth bomb fell into the sea and became the object of an extensive search
by the United States.
The scientist in charge of the recovery used Bayesian statistics to locate the weapon. The Bayesian approach
combines knowledge and intuition of experts with experimental results; the experts placing "bets" on the most
likely outcomes before any action is taken. A validation of this approach comes from research showing that at
the racetrack bettors accurately predict the odds of a horse winning a race.
In the Spanish nuclear bomb case, one of the preliminary high-probability sites was a spot where a fisherman
claimed to have seen the bomb enter the water. As continuing negative results from the other likely search
areas caused a decline in their probabilities, the fisherman's site rose to most likely and resources were
diverted to it. In the end, he was proved correct and was able to collect a salvage award on the $2 billion recovery.
[ Back to Method ]
| |
|
