Fermi Problem

A type of problem in which insufficient information is available to accurately solve the problem. Instead, the solver makes educated guesses about subcomponents of the problem, and uses these to reach a rough order of magnitude solution. If several people independently reach a solution, you can achieve a more precise answer by combining their answers.

Named after Enrico Fermi.

Some sample Fermi Problems:

How many hairs are on your head?

Unknown factors: Hair density, surface area of scalp.

How many pennies are handing to or from supermarket checkout clerks in a given day?

Unknown factors: sales per day, percentage of cash sales, percentage of sales with precise change given to clerk.

Number of piano tuners in Pittsburgh

How many residents in Pittsburgh, how many pianos per capita, how many pianos serviceable by a given tuner.

In the 1989 Loma Prieta earthquake in California, approximately 2 million books fell off the shelves at the Stanford University library. If you were the library administrator and wanted to hire enough part-time student labor to put the books back on the shelves in order in 2 weeks, how many students would you have to hire? (You may assume that the books just fell off the shelves and got a bit mixed up but books in different aisles did NOT get shuffled together.)

How many atoms of Jesus do you eat every day?

And no, communion wafers do not count.

For more Fermi Problems, see: http://www.physics.umd.edu/rgroups/ripe/perg/fermi.html

It's a given when "solving" a Fermi Problem that your answer is only going to be accurate to a power of ten or so. In other words, if the calculated correct answer to "How many atoms of Jesus do you eat every day?" is 250 trillion, then the actual answer can be expected to be somewhere between 25 trillion and 2.5 quadrillion.

That's not terribly accurate for everyday engineering, of course, but for an educated guess it's pretty good.