Darts for Geeks: Statistician Cracks the Game’s Secrets

Photo: Melissa Kaseman

Ryan Tibshirani couldn’t hit a bull’s-eye if his PhD depended on it. Last fall his roommate nailed up a dartboard and proceeded to trounce him game after game — even with his eyes closed. So Tibshirani (in green, above) did what any good geek would do: The Stanford statistics student used a Gaussian model to calculate where he should aim to maximize his points in a game of High Score.

In darts, the difference between cheers and boos can be a mere millimeter. Winning wedges are adjacent to losers — 20 is next to 1, for example. But some are more forgiving: If you miss 14, you still might land on the respectable 11 or 9. So, Tibshirani figured, your accuracy should dictate your target.

To find his own sweet spot, Tibshirani launched 50 darts at the bull’s-eye and ran his scores through an algorithm that computes a rank from 0 (flawless) to 100 (hopeless). He then wrote software to translate that number into a heat map of the board, highlighting areas that are his rank’s best bets.

Tibshirani has put his app online for pub athletes everywhere. And now he’s imagining something even better: a dart-flinging robot with an accuracy dial. Tibshirani may never prevail in the dorm rivalry, but he could program his bot to win — or at least scare the pants off his old roomie.

Find Your Sweet Spot

Aim 50 darts at the bull’s-eye. Then plug your scores into Tibshirani’s app (stat.stanford.edu/~ryantibs/darts) to get your heat map. here are three sample results: