The NBA, NFL, NHL, and Major League Baseball have embraced player-tracking solutions over the past three years and have begun integrating graphic visualizations of the data into broadcasts and streaming content. The four major U.S. sports leagues join a rapidly growing global bandwagon of leagues, teams, and broadcasters beefing up data-analytics teams and investing in tracking/visualization. Despite the treasure trove of data available, however, the key remains adding context to make it digestible for both fans and professionals alike.You have to be very, very careful to provide context, because data in and of itself is just an overwhelming amount of zeros and ones that don't mean anything, cautioned Perkins Miller, chief digital officer/head of media operations, NFL, at the MIT Sloan Sports Analytics Conference this month in Boston. That is where people will confuse data with insight. We spend a lot of time taking this data and making sure we provide the right context so you can develop the right insight. As Mark Twain said, There are lies, damn lies, and statistics.' You have to be very careful about how you provide that context and insight.
NFL, Zebra Report on Deployment of Next-Gen Stats
The NFL partnered with Zebra Technologies to install player-tracking systems at all 31 NFL stadiums (as well as Wembley Stadium in London and Aloha Stadium in Hawaii) last year. NFL players wore RFID chips that delivered data points to these systems, serving the league's Next-Gen Stats system. In all, Zebra's system tagged and tracked 2.500+ NFL players and collected more than 180 billion bites of position data during the 2015 campaign.
We spent a lot of time working with Zebra, said Miller, and they have done a great job to provide that level of fidelity so that the data that comes out is predictable and consistent. If it's not predictable and consistent, it's not trustworthy.
A screen from the NFL Next-Gen Stats system
With a full season of league-wide player-tracking in the books, coaches, scouts, and strength/conditioning staffs; the league's broadcast partners and NFL Media; players and even fans are beginning to reap the benefits of deeper data.
Not only does [data-analytics/visualization technology] work at the really granular level with player development, health, and safety and at the coaching level, we also look at it through the lens of how we can tell the story for our fans better on television? Miller added. How can you understand how hard it is from a quarterback's [perspective] to throw to an eligible receiver in a split-second decision?
He also stressed the importance of selecting specific scenarios in which to use the data within an NFL telecast, rather than forcing mountains of incomprehensible information on the fan.
You're not going to get someone really thrilled talking about a change-vector speed, but they might want to understand how much separation a wide receiver had on coverage and has he been able to maintain that separation throughout the first quarter and is that why his team has gotten so many first downs? Miller explained. If you can find a way to package that right, [such as] a separation index that is easy to digest, you can get to the vernacular, and it becomes part of the story.
NFL Hackathon Delivers Deeper Insights
In San Francisco last month, the NFL hosted a 24-hour hackathon, handing its Next-Gen Stats data to an army of software developers and challenging them to develop the next generation of engaging NFL stories. Given raw location data, developers were tasked with creating real-time, in-game pattern-recognition systems to provide deeper insight into the game. The winning group came up with a pattern capable of determining the probability of a successful pass immediately after the ball is snapped, based on the configuration of eligible receivers and the defensive line.
That is just a fascinating piece of evolution, said Miller. [It] makes the game more interesting if you are able to understand the decision-making the quarterback has to make as he drops back to throw. Now, how soon we can make that a reality is still many leaps to the future, but there are ways to make the game even more exciting than it already is on television by using the data to tell a story and showcase how hard it is for some of these players to perform on the field.
More on Sports Tracking: The Next-Generation Panel
In addition to Miller, a panel on sports-tracking solutions offered insights ranging from the benefits of data analytics for evaluating player performance and the potential impact on fantasy-football games to player strength, conditioning, and medical evaluation. Here are a few highlights from the session:
Dean Oliver, VP, data science, TruMedia, on the potential impact of data analytics on fantasy football for fans:
I think we can come up with all sorts of new fantasy stats for positions that have never really been measured. For [a player like New York Jets cornerback] Darrelle Revis, who is denying and clogging the passing lanes, we can add more to that conversation. Some of the players have rebelled against I didn't get the touchdown, but I ran 87 yards to the 1-yard line, and people were upset because I didn't get the touchdown.' I think we can add a few more meaningful statistics to that, and the consumer fantasy [market] is certainly going to be interested by some of these.
Matt Sheldon, director, football analytics and research, Chicago Bears, on where player-tracking systems must improve to deliver more-valuable data:
So many of these systems are really tuned towards open-field running. In American football, 45% of the players don't open-field run on every play. We would like to see orientation to know when a player is running forward or backward. In the passing gam
