8 ways analytics are changing pro sports

Tracking sports statistics was once viewed as the domain of the organization’s techies. But more and more, data is driving the action on and off the field.  As these examples from the recent MIT Sloan Sports Analytics Conference demonstrate, analytics have gone from bench player to emerging star.

Analytics may offer hope to fans frustrated by the length of professional sporting events. The most prominent example of slow play is Major League Baseball, whose critics suggest that the league may be losing casual and younger fans because games routinely last longer than three hours. “The time between actionable moments is too long for the casual fan. [The league] should be trying to shrink the time between interesting events,” said Dave Cameron, managing editor of FanGraphs, a sports statistics website, at the MIT Sloan Sports Analytics Conference last month.

The introduction of technology, such as MLB’s digital clock this season, could help shorten the game. But other experts at the conference said the debate is a double-edged sword, which brings the risk of losing the drama of America’s pastime. For example, Dan Brooks, founder of BrooksBaseball.net, a website that makes sports statistics digestible for sports fans, points to the tension of a hitter stepping out of the batter’s box to try to figure out the next pitch or a pitcher walking off the mound. “I don’t know if that would be any more fun if [the game] was 15 minutes faster.”

If Seattle Seahawks coach Pete Carroll could have used predictive analytics to anticipate the interception by Patriots cornerback Malcolm Butler in the final minute of this year’s Super Bowl, would he have called for quarterback Russell Wilson to throw a pass? With data collection and statistical analysis playing a bigger role in professional sports, there is potential for decision-making to be more precise than relying on a coach or player in the heat of a game. When asked whether a statistical model could replace a coach, New England Patriots director of player personnel, Nick Caserio, was skeptical. “Coaches spend an extraordinary amount of time preparing. You can formulate a plan but be prepared to change it in the course of a game. My opinion is no, the ability to adjust in the game [means] you may have to cast aside what you spend six days doing.”

Brian Burke, president of hockey operations at the Calgary Flames, agreed, saying in-game decisions are still a combination of data and gut instinct. “[Analytics] are useful for support, not illumination. [Sports] is still an eyeballs business. The notion that you can sit behind a computer and find athletes is BS.”

The hottest talent in pro sports used to be that coveted draft pick or talented free agent. Now it’s the statistical wiz who can manipulate numbers, create models, and provide guidance to players and coaches. For example, the Edmonton Oilers hired Tyler Dellow, who created his own website, mc79hockey.com, around hockey statistical analysis. He now works as an analyst with the team, helping to analyze games and mine data. Dallas Eakins, former NHL player and former coach for the Oilers, hired Dellow. “I had to get [it] pushed through management with the Oilers. It was something I thought was important,” he said.

The Texas Rangers offer another example of a sports organization investing in data analysts. The team formed an analytics task force, made up of people from IT, marketing and ticket sales, and hired an analytics intern to help improve in-stadium data analysis. The team brainstorms ways to mine data to sell more concessions, merchandise and tickets, and decrease unnecessary staffing during games, for example.

One similarity between sports organizations and other enterprises is the disconnect between IT and the business. Speakers at the MIT Sloan Sports Analytics Conference said statistical experts need to be able to communicate relevant information to players, coaches and staff in order for the organization to benefit from technology and analytics.

“Analytics are only useful [if] they can be communicated. People [who] can communicate what they find will be most helpful to the process,” says Scott Pioli, assistant general manager of the Atlanta Falcons. Besides the statistical language barrier, old school thinkers need to come into the fold of the new technology and systems that are becoming part of the game.

Eakins said he’s still an old-school coach, but has become more open to using data analysis to steer his decisions. “The most important thing is using every resource you can to get better, whether it’s managers evaluating talent or coaches putting together systems,” he said.

Kyle Dubas, assistant general manager of the Toronto Maple Leafs, said the team’s R&D group challenges him to be better at his job. “When all facets are working in lockstep, that’s when you can see the separation [from] using data and analytics. We’ll find ways to beat other teams, out-draft teams or do better in [player] development.”

Statistics aren’t just for players and coaches. Fans are getting in on the action by, for example, starting their own websites and doing ad-hoc statistical analysis. As a result, leagues are trying to figure out more ways to engage fans based on the level of enthusiasm and participation that has grown organically online.

The Oilers’ Dellow (slide 3) suggests getting involved in the conversation around sports and analytics on Twitter and creating your own blog, which is how he was noticed by Eakins. “Do good work, have an opinion and back it up,” Eakins said.

Wearable devices and sensors have created a world where teams and players can get real-time data during games and practices. They can then analyze that data to make better decisions, change game plans or prevent injuries. For example, the NHL tested Sportvision player and puck trackers at this year’s All Star Skills competition. “In hockey, we’re desperate for more data. There has been a great frustration of not being able to do [analysis] or pursue leads,” said James Mirtle, a hockey writer for The Globe and Mail. “That’s what the tracking data will do now.”

MLB has begun charting new statistical territory as well. For example, MLB Advanced Media rolled out Statcast, an interactive tool that tracks the baseball and players on the field. Statcast measures factors like trajectory of the ball and speed at which the player runs after it. “With radar guns we’ve focused on velocity because we had a tool to measure it,” said Sandy Alderson, general manager of the New York Mets. “Now we can measure spin rates and suddenly we have a handle on something other than velocity.”

Wearables can also help prevent, track and detect player injuries. But all the data collected from these new devices will need to be analyzed and translated into actionable insights, which will create the need for more analytics-driven positions on teams. Brian Burke, founder of the website, Advanced Football Analytics said if more statistics experts aren’t hired soon, leagues will “enter a state where there’s more data than analysis. “Teams need people to make sense of it and find inflection points.” Successful analysis could enable teams to pinpoint a reason for a poor performance on the field, such as if lack of sleep contributed to a player’s slower running speed. However, the effectiveness of analysis is connected back to the method of communication, said the Patriots’ Caserio. “In the end, what coaches care about are two to three bullet points.”

Analytics can guide better decision-making for teams during drafts, scouting, combines and trade considerations. One futuristic example is a single statistic that could represent the total performance of a given player — something that combines speed, success on certain plays, strength and other factors. “The future of analytics is a merger of quantitative and qualitative,” said of Advanced Football Analytics.

Burke described how in the future, if one player was twice as good as another, it could be an indicator of how much each player would impact the bottom-line revenue. “It’s the holy grail of football analysis. You can find inefficiencies, positions that are under- or overvalued and optimize your roster. That’s the future.”