As the world continues to wait with bated breath for the white-knuckle thrill ride that—should it ever get made—will be the Where’s Waldo? movie, a data scientist has just the thing to at least temporarily slake the public’s inexhaustible thirst for all things Waldo. Randal Olson, a graduate research assistant in Michigan State University’s computer science department, has run the numbers and found the best strategy for easily beating your little shit nephew in the hunt for the titular character in any given Where’s Waldo? book.
Starting from an earlier analysis of Waldo’s top hiding spots from Slate, Olson dives in with a kernel density estimate, a genetic algorithm and several other data analysis tools you’ll nod blankly at, a telltale bead of liberal-arts-degree flop sweat trickling down your temple. He ends up with the best possible path to follow in order to find Waldo (or Jonas, if you’re Lithuanian) in the least amount of time. That path looks like this:
An explanation of what that means, as well as some surprising details on just how hard it is to build an algorithm to track the predictable movements of some nerd in a red-and-white striped shirt, can be found on Olson’s blog.
