Published: 15 January 2016
One big lesson from this year's MIT Sloan Sports Analytics Conference was that sports analytics has now fully grown up from its blogger-in-his-pajamas phase and is now big business. Major, global consulting firms that do most of their work for billion-dollar-budget federal agencies are advertising their services to teams and leagues. Much of the growth in the field is in player tracking, both in terms of physiological measurement and in digitizing player location information during games.
I have no doubt that both of those approaches can offer teams insights and benefits, but it remains to be seen just how big those benefits might be and how costly and difficult it will be to get them. I get the sense that there will be an avalanche of data and it will require large and expensive efforts to gain marginal benefits above what conventional methods offer.
That's why I remain convinced that the most direct, most demonstrable, and most actionable analytic approach in football is in-game decision support.
I mean that benefits of the analysis lead immediately to the thing that matters most--winning. Better physiological data could lead to healthier athletes, which would then lead to winning. But it's murky how healthier we can make players and how much the improvement could mean to a team's win total.
I mean that the analysis can directly quantify and verify its own impact. When a coach is faced with two alternatives and the traditional choice offers .05 points of win probability less than the choice the analysis recommends, we can credit the analysis with having that impact on the game.
I mean that the analysis doesn't exist for its own sake. There is a high-stakes decision to be made that is directly informed by the analysis. In short, it matters.
Perhaps the most important aspect of game strategy analysis is that it's the most cost-effective way to increase a team's win total, by far.
Consider the very best performing All Pro running backs each season. The last few were DeMarco Murray, LeSean Jackson (twice), Marshawn Lynch, Maurice Jones-Drew and Adrian Peterson. Those top runners averaged 0.08 Win Probability Added (WPA) per game during each of their All Pro seasons. In other words, if an average team added such a RB to its team, knowing he would be the very best in the league that season, it would mean taking their expected winning percentage from .500 to .580.
Compare that with an effective in-game analytic strategy. You can give me any game (that's not a blowout from the beginning), and I could easily find .07 points of squandered win probability that could have been saved by better decisions regarding clock management, play selection, two-point conversions, onside kicks, coach's replay challenges, and of course fourth downs.
Now how much money would a guaranteed All-Pro level running back command? Certainly something north of $10M per year. I think I can safely speak for most analysts and say we'd gladly take merely half of that. Just kidding...I'll take all 10.
Seriously, I believe that's exactly what this kind of analytics would be worth to any team that chooses to embrace it. And in reality a good analytics program would cost a tiny fraction of that amount, and it wouldn't count a dime against the cap.