#BigData invades the NFL Draft. Vikings are all in.
A month and a half ago, I wrote “#BigData Naysayers and Luddites in the National Football League. Laughter is NOT the best medicine!” mentioning how some NFL execs ridiculed the hiring by the sad-sack Cleveland Browns of Paul DePodesta. Mr. DePodesta, with no football experience but with some street cred as one of the earliest architects and adherents of ‘Moneyball’, was brought on board by the Browns as their ‘chief strategy officer’. Not because he could design clever schemes for the offense and defense, but because he and his cohorts were very good over time in using statistics and data to aid in finding “the most ‘predictive markers in player evaluation’.”
Today’s (4/28) Wall Street Journal offers a story “A Data Scientist Dissects the 2016 NFL Draft” which speaks of Jared Lander, hired last year by Rick Spielman, GM of the Vikings for one very specific reason. “I was always curious about what we could do with all the data we have on players,” said Spielman. “Could we spin it in a way that would help us predict how a guy would play at the NFL level?” No one laughed at Spielman when he then hired Lander, maybe because Spielman was a longtime football man, albeit with no professional playing experience, so presumably he got a pass when he showed a bit of out-of-the-box thinking.
The results were immediate and eye-popping. Every team gets a grade as to their draft picks, from prominent football analysts like Mel Kiper Jr and the Vikings got an A in 2015, with two of their picks making one guy’s all-rookie team. Lander’s methods were not complicated and the fact that they brought such success should surprise no one who embraces data analytics and the advantages of the Big Data ecosystem. In conjunction with “the Vikings scouting department, Lander used hundreds of variables to assess each player. He ran those variables thousands of times through an ensemble of machine-learning algorithms, allowing him to evaluate them in comparison to players over the last 10 years.”
This involved no time spent at the annual NFL meat market, officially known as the Scouting Combine, where promising prospective pros (say that three times fast with a mouthful of Saltines or Weetabix) are poked and prodded and timed and measured to see how they compare to their peers. The weakness there is that the Combine is a one day event and looks at how the kids do on that day. We all have bad days and woe to the college kid whose bad day is the day of the Combine. Of course, any team, especially the Vikings, should combine what they observed at the Combine with Lander’s work (and in-season observations) if they chose to use it to establish their draft ‘board’, or list of players they’d like to select.
As impressive as the effort by Lander when engaged by the Vikings, his CV is even more so. His little company has worked with the Centers for Disease Control and the Department of Defense and was also asked how many winners could be expected for that memorable $1.6 Billion Powerball drawing this past January. His answer was 3, arrived at using ‘rules of probability and Monte Carlo simulations’. Guess how many winners there were. First two answers don’t count. Yes, there were indeed three winners.
Though Mr Lander and I are both Columbia men, I must admit to not fully agreeing with his mantra. “The bottom line is, if you have enough data, you can come pretty close to predicting almost anything.” Yes, but. You can have too much data and the noise from the excess information can taint a prediction. And, you can properly use the data, but as I have said many times, you still have to look at the results with an experienced, somewhat jaundiced eye and also take into account intangibles such as character and intelligence, etc. (Johnny Manziel and Rick Leaf come to mind as those who might have slipped through the cracks.)
Yet another use case heard from. There will be many more, as statisticians, who likely do not chew tobacco and can’t throw a spiral, come up with more ways to slice and dice the information in front of them. Lander was not engaged by the Vikes this year but the team has asserted that they still believe in his methods.