Using data to try and predict the outcome of a football match – or any sport, for that matter – is a notoriously difficult science that can stump even the most talented of data analysts. This much is evident from the failure of Big Data whizz kid Nate Silver’s attempt to use analytics to predict who would progress to the Super Bowl next month. Silver, who correctly guessed the results of all fifty states in the presidential elections, came a cropper this time round when he tipped the Patriots to meet the Seahawks in Super Bowl XLVII.
Given the ridiculous number of unpredictable variables that come up in any football game – weather conditions, emotions, errors, illnesses, to name just a few – few people will have been surprised to learn this much. But just because we can’t predict the outcome of NFL matches by analyzing the data available, does that mean that football should just ignore the wealth of stats it has available? Hell no…
One area where Big Data definitely has potential in the NFL world is with talent scouting. Now, given the resources available to today’s NFL teams, and the awesome financial prizes at stake for those teams that enjoy a successful season, you might think that coaches and scouts are already employing the most cutting edge decision making processes when it comes to acquiring personnel.
But you’d be wrong. According to Ted Sundquist, ex-General Manager of the Denver Broncos, NFL teams largely tend to ignore the data available to them, instead relying on outdated techniques such as ‘personal judgment’:
“My experience over three different regimes with the Denver Broncos (1992-2008), in the capacity as both Director of College Scouting and General Manager, led me to the conclusion that “something is being missed”, writes Sundquist.
“That something is the proper utilization of statistical analysis as it relates directly to the identification, evaluation, and procurement of talent in the National Football League. I’ve written about this subject and related topics numerous times on TheFootballEducator.com. In a multi-billion dollar industry such as professional football, it seems rather odd that this would be the case.”
“But taking into account tendencies, trends, and correlations from an unbiased view (such as Advanced NFL Stats, Football Outsiders, Ourlads) you can gain a greater, more accurate indication of what consistently leads to winning in professional football. You can develop a better understanding of the parameters and indicators of predictive potential and productivity in the game’s players. And thus you should have a greater degree of confidence in choosing the right players for your club, and for the right reasons.”
Calling The Shots
There’s been a lot written about whether or not statistical analysis can benefit coaches when it comes to calling plays in crucial situations, and the overwhelming evidence seems to suggest that it can. Take punting for example. A recent study by Advanced NFL Stats suggests that football coaches consistently make the wrong call on fourth down, choosing to punt when the statistics show that this tactic leads to fewer wins than if they had decided to go for first down.
Other studies have come up with similar findings, with the New York Times citing Professor David Romer of the University of California, who determined from an analysis of all NFL games between 1998 to 2004 that teams are statistically always better off going for it when facing fourth-and four-yards or less
Further evidence to suggest that teams should ‘go for it’ comes from a recent game between the Detroit Lions and the Tennessee Titans this season. ESPN highlights the crucial mistake made by the Lions in attempting to draw the Titans offside on the last play of a 44-41 defeat, a play that ultimately backfired. In his analysis, Kevin Seifert explains that, statistically speaking, the Lions would have almost certainly tied the game had they attempted to go for the field goal, and they even had a 54% chance of winning had they gone for the touchdown – yet they chose to ‘get cute’ and paid the price for their antics.
Reading The Game
Perhaps the most likely to benefit from Big Data in NFL will be the players themselves. For all their talent and ability, players often find themselves caught out of position. This isn’t necessarily because they lack the ability to read the game – it’s just that, given the huge number of variables, sometimes it simply isn’t possible to know exactly where each player is best positioned on the field in a given moment.
A new technology solution may be able to help change that. SportVU.com provides optical tracking data that relies on cameras at the side of the field to follow and record the movements of players in real-time, as well as the three-dimensional movements of the ball. The technology is still in its infancy but its progressing by the day.
Tests in NBA have shown that it’s possible to discover the “non-linear relationship between shot location and its impact on offensive rebound rates” and judge with a reasonable degree of accuracy which part of the court a player should run to in order to grab the rebound if a shot misses the target. We’ve yet to see such tests made in NFL, but the potential is there to see – by analyzing previous games and the tactics of opposing teams, the technology could help players to understand where they need to be in any given situation.
Given how Big Data is being applied with success in other sports, it’s surely only a matter of time before someone in the NFL does the same. And as Ted Sundquist predicts, we can expect big things from whoever does so first:
“Those GM’s and head coaches that figure this out will put themselves well beyond the teams working with just magnetic nameplates stuck to a white board.”
Before joining SiliconANGLE, Mike was an editor at Argophilia Travel News, an occassional contributer to The Epoch Times, and has also dabbled in SEO and social media marketing. He usually bases himself in Bangkok, Thailand, though he can often be found roaming through the jungles or chilling on a beach.
Got a news story or tip? Email Mike@SiliconANGLE.com.
Latest posts by Mike Wheatley (see all)
- The 5 biggest markets the IoT is poised to disrupt - May 30, 2016
- Google intros “node pools” to support heterogeneous container clusters - May 30, 2016
- Amazon Web Services said to be secretly building a new AI service - May 27, 2016