Last month, Google rolled out March Madness spots touting the number-crunching capabilities of its cloud computing services. What better dataset to dig through than NCAA game stats?
Now that the tournament is over, it's clear that Google wasn't overestimating its own predictive prowess. At halftime during the Final Four games, Google fed data from the first half into its algorithms to create a prediction about how the second half would play out. Then, along with creative agency Eleven, they had 10 minutes to create a 30-second spot that ran on CBS before gameplay resumed. The accuracy is impressive.
During the Loyola vs. Michigan game, Google Cloud predicted 29 rebounds in the second half. There were 29. During the final game between Villanova and Michigan, the Cloud predicted a 76.4 percent chance of at least 21 three-pointer attempts. Final count: 24. Now as long as Google Cloud can't start placing bets on its own, we'll be all right.