Possibly the closest Super Bowl matchup in history also poses the question: how much does it mean when certain aspects of an NFL team improve dramatically in the second half of the season?
28 Aug 2013
by Rivers McCown
If you ever make the trek to Boston for the Sloan Sports Analytics Conference -- or even if all you do is watch the presentations online and seek out pieces about the powwow -- the word that comes up again and again is communication. Organized sports, as a whole, has struggled to deal with the integration of sports analytics. Whether it be a coach ignoring game theory or a general manager going against his data, one need not look far to see a basic principle of sports analytics being sullied. Benjamin Alamar's Sports Analytics: A Guide for Coaches, Managers, and Other Decision Makers, is a state of the industry from an analyst who has been on both sides of the fence (Alamar has worked or consulted for several NFL and NBA teams, most notably the Oklahoma City Thunder), and recognizes the inherent challenges in finding the right balance to the equation of analytics and traditional sports positions.
A majority of the book's findings are supported by the results of the Sports Analytics Use Survey (SAUS), which was answered by 27 "individuals" representing teams across the NFL, MLB, NBA, and EPL. For example, the idea that communication is a problem is buttressed by two responses to the survey, from different executives on the same team, that showed wildly different opinions on the usefulness of the in-house research. In response to the statement "Your analytical capabilities are stronger than your competitor's," a personnel man responded "somewhat disagree" and an IT man from the same team responded "strongly agree." Alamar draws on both theory, personal experiences, and anecdotes from others as he relays his ideas about the future of sports analytics and the way they can best be implemented on a chapter-by-chapter basis that takes on cases such as "Data and Data Management" and "New Metrics."
This book showcases some material that would be known quantities for frequent readers of this website. Things like small sample size, or the challenges inherent in statistical databases, are concerns that we've been over a time or twenty. Describing the process of trying to discern the probability that a given NFL coach is going to be successful based on his past job -- as well as what "successful" means -- is a subject that has been taken on multiple times by enterprising people. Alamar has several pattern-based model examples where he'll walk you through the steps of the development of basketball's PER metric, and how that can be applied to a theoretically new metric today. Still, football-minded readers will enjoy anecdotes like the one on page 60, where a defensive end's performance is broken down by several different people from the same staff and they continue digging at the question of how good he could be before eventually determining that most of his sacks came on plays where the quarterback held the ball much longer than the average sack. At its heart, this book is seeking to bridge the gap between analysts and "decision-makers," as Sports Analytics refers to them -- and the accessibility inherent in that decision means there aren't as many jump-off points for the sophisticated sports fan.
The two biggest problems that Sports Analytics deems worthy of its time are the better organization and display of statistics, and ways to bridge the communication gap between analysts and decision-makers. Mostly, the organization problems are problems of accessibility. Alamar starts out referring to a simple problem where different reports from different sections of a front office will return different player names and organization techniques, so that the decision-maker will have to sit down and mush the information together. Then it moves to ideas like integrating all the vast swarms of information -- from video highlights to video-capture data to metrics -- into one easily accessible system. Without teams focusing on this, Alamar posits that a series of reports with no real rhyme or reason to their organization will make for a lot of wasted time cross-referencing lists for decision-makers, and leave them unlikely to embrace the new departments fully. Another telling find from the SAUS was that so much of the data is inaccessible to a front office as a whole -- 93.7 percent of respondents listed that access to "some" or "most" data was dependent upon just one person. That's a lot of time thrown away or spent waiting because the organization did not develop an initial plan to streamline access.
The more interesting takeaway for me were Alamar's thoughts on the communication gap. Because many decision-makers don't have the specific traits that apply to analysts, many of them simply don't know where and how to hire them. To take one anecdote from the waning chapters, "A question often asked to analysts in the interview process is: How do I know if you are any good?" Decision-makers "get used to looking at a particular set of information, and unless they are motivated to expand that set ... it is unlikely that they will, no matter how well the analyst makes his case in the metric's documentation."
The dynamic at play here is not scouts versus stats -- this is an even older one: willingness to learn and adapt versus hubris. I won't go so far as to presume that simply having an analytics-based front office is the "smart" way to do things, but my takeaway from this book was that teams that are inclined to be on that path have common traits that would lead them to be innovators and reap competitive advantage from their innovations. Just about every sports team has some measure of analytics department in-house. The questions are: how much power do they have and what tasks are they being made to focus on? Alamar relates a story of a defensive coordinator not adhering to entering the data into the system and saying -- I'm paraphrasing -- "if they want my thoughts on how the team played they can come ask me."
The world of big data is here to stay. The teams that do the best job of cultivating and harvesting that data are going to have advantages as long as they can implement it as an organization-wide philosophy. We don't know how that will play out in the NFL yet -- the Jaguars are the first team to make football analytics a hallmark of their front office and they have a very deep hole to dig out of -- but my suspicion is that, much like in other sports, there will be front offices that are willing to challenge conventional thinking and ones that are only willing to pay lip service to the idea of doing so. Teams that are willing to innovate, and teams that are just starting to get a foothold on the statistical profile of football as of 2007.
So much of this book is spent on affirming how important it is to be able to relay a message to the decision-maker that the takeaway I got out of it was "just hire a decision-maker who does understand this and you're ahead of the game." Easier said than done, I know, in a business as incestuous as football. But that's the path. Football is still at the very start of its data revolution compared to other sports -- there's no better time than the present to take a risk like this and be ahead of the game for decades to come.
Sports Analytics (Amazon link) is, in many ways, a textbook for the new discipline. It's a breezy read at 124 pages, but it's a good get for someone looking to break into the sports industry through analytics. You'll definitely have a much better picture of the depth of the climb analytics faces to be fully embraced throughout sports.
2 comments, Last at 29 Aug 2013, 9:56am by Ryan D.