Tech giant Google has long been a specialist in finding unique uses for its technological achievements, but using predictive analytics and the cloud to predict how the second half of a basketball game will go as soon as the first half is done is fairly new territory. Thanks to a partnership with the NCAA, Google was given access to a wealth of data on past basketball games, and used its own Google Cloud Services and some AI muscle to figure out trends and patterns on both an overarching basis and per-game. Using the same techniques, refined by past handling of vast swaths of data, Google plans to have a crew on site to pump data into its cloud servers during the NCAA Final Four.
The whole arrangement is going to be just as high-pressure and fast-paced as the game being analyzed. Google has a staff of Google Cloud employees that will be on hand for games, along with a number of basketball enthusiasts that Google is calling the Wolf Pack. Together, the bunch will work with Google’s own tools to analyze the first half of a basketball game as it happens. When halftime comes and the final numbers are crunched, the team will generate predictions for the second half of the game, and then produce a quick TV spot showing off those predictions. This will happen fast enough to hand off the ad tapes to CBS and Turner before halftime is over, and have them aired on TBS before the game comes back on.
The kind of number crunching required for such a feat is nothing to scoff at, and the idea to make an ad for immediate showing once the results are out only adds to the pressure. This is, essentially, a very public test run for predictive technology. Due to the human-driven nature of the sport, the chances that Google’s predictions could be wrong are astronomical, but so are the chances that they could be right. Audiences watching at home will be able to tune in to TBS live during tomorrow’s big March Madness conclusion to see the system for themselves, and see if Google’s technology is able to accurately predict the big winner.