Here are some of the things I learned during the final day of the MIT Sloan Sports Analytics Conference in Boston. See the Day 1 column, in case you missed it.
Mark Cuban doesn't think you should try to work in sports. The conference ended with another taping of a live B.S. Report, this time with Bill Simmons and Mark Cuban. The Mavericks owner mentioned that young people are constantly asking him for advice regarding a career in sports. His answer: "Don't. It doesn't pay shit, and 1,000 people apply for every job."
Deadbeat gamblers get preferential treatment.
Most surprising reveal of #ssac: big time gamblers can get a different line than the one on the board.— Dashiell Bennett (@dashb0t) March 3, 2012
At the gambling panel with all the other degenerates. Vegas often ahead of curve on predictive tools. #ssac— John Hollinger (@johnhollinger) March 3, 2012
But I guess we already knew that.
Matt Roth doesn't want to hear about numbers. The player representative on the "Football Analytics" panel alongside Grantland's Bill Barnwell, former NFL head coach Eric Mangini, San Francisco 49ers Chief Operating Officer Paraag Marathe and moderator Andrea Kremer, Roth, a defensive end for the Jacksonville Jaguars, held his own. But when asked what role he wants numbers to play on the sidelines during the game, his found himself thinking out loud. He started with a hesitant "It has its place," moved on through a number of "Players just want to play" rationales, and, when pressed a bit, finished with a simple "I do not want to hear the numbers on the sideline. No."
The San Francisco 49ers might have some injury trouble next year. Part of the "Football Analytics" discussion focused on analytics beyond performance numbers. On the topic of injury prevention, Barnwell mentioned how some teams are consistently more healthy than others (Tennessee Titans) while others seemed to always be battling a series of injuries (St. Louis Rams). Maraathe said that on average, teams lose 60 starter games a year to injury, the equivalent of almost four starters being lost for the entire season. If you can figure out methods for either preventing injuries or minimizing recovery time, you could find an edge. (Mangini mentioned ways to track reps in practice to make sure recovering players aren't getting overworked.) Marathe pointed out that his 49ers only lost 3-4 starter games on defense, which helped them advance to the NFC Championship game. Perhaps the Niners have cracked the code a bit … or perhaps they were a bit lucky this year.
This was a basketball conference. As well as organizers did in varying the topics and giving most sports a chance at drawing a crowd, the hoard of basketball junkies was dominant. Good things are coming your way, hoops fans.
Drew Carey was Day 2's Jeff Van Gundy. Wearing his "Owner Of The Seattle Sounders" hat, Carey was featured on two panels -- "Franchises In Transition" and "Soccer Analytics" -- and held his own on both, taking conversations to interesting and funny places, much like Jeff Van Gundy did in Day 1.
Drew Carey: "We could sell cocaine and make money. We want to win." #ssac— charlie widdoes (@charliewiddoes) March 3, 2012
"In Seattle, we want to win, but we want to sleep at night, too, so we want to treat our fans right. " -@DrewFromTV , my new God. #SSAC— Hardwood Paroxysm (@HPbasketball) March 3, 2012
Gotta be a smart shopper, man. Carey managed to be both entertaining and enlightening. In the "Soccer Analytics" discussion, he freely acknowledged that America's college soccer structure is not a good thing for the MLS, and how youth academies, longer seasons and better competition would serve the league better than a system that supplies talent that requires quite a bit of seasoning after college before it typically thrives. He also talked about the ways that analytics can play a role in identifying talent in the "lower middle of the [talent] pyramid," from which most MLS teams are choosing. "It doesn't take a lot to understand that Didier Drogba is really good," but there are a lot of similar players for MLS teams to choose, and how they go about it determines their success. "Gotta be a smart shopper, man."
Soccer executives have not spent a lot of time at FourFourTwo.com. Everybody on the "Soccer Analytics" panel -- Carey, Alexi Lalas, Steven Houston (Head of Technical Scouting, Hamburg FC), Scott McLachlan (Head of International Scouts, Chelsea FC), Steve Brown (Senior First Team Analyst, Everton FC) -- acknowledged that the typical soccer box score is atrocious and nearly useless. They also discussed how very little performance data was publicly available for fans to use and analyze. That is true, but … have you seen the FourFourTwo Stats Zone? It is incredible. Yes, some of the data featured is simple physical data -- where players lined up and how far they ran (which is data the panel felt was a bit overrated in terms of anything predictive) -- but there is wonderful optical data regarding the length of passes and shots, and the pass completion percentage data is as useful as the same type of data in American football … which is to say, it is pretty useful.
On a side note, if you are looking for the soccer version of Scott Hatteberg (the perfect example of a player analytics proclaimed as useful before conventional wisdom did), and I know you are, look no further than Shinji Kagawa, attacking midfielder for the Bundesliga's Borussia Dortmund. Unheard of before Dortmund executives uncovered film of him playing at Japan's Cerezo Osaka, Kagawa helped Dortmund to a league title last year and is in the process of doing the same this year. As Houston pointed out, he was a no-name, and now he could probably draw a 20-25 million pound transfer fee soon.
Analytics cannot save U.S. Tennis. Most "_____ Analytics" panels began the same way, with executives, current or former players, and ESPN personalities all smiling and telling the crowd that analytics are surging in the given sport (then they spend 40-50 minutes attempting to prove it with varying degrees of success). The "Tennis Analytics" panel, on the other hand? Not so much. In ESPN the Magazine's Analytics Issue, major sports were ranked according to analytical advancement. Tennis was second to last, ahead of just boxing. Former pro Todd Martin, on the panel with coach Paul Annacone, "tennis analytics pioneer," New York Times contributor Craig O'Shannessy and moderator Marc Stein, discussed how, when you watch tennis on television, you are seeing the same stats you saw 20 years ago: first and second serve percentage, winners, and unforced errors. Meanwhile, as I learned during the Australian Open, you cannot really even find consistent data online regarding winners and errors, which was mind-boggling.
O'Shannessy discussed the ways in which he is using data and analytics to help professional players -- and yes, I joined his website, The Brain Game, immediately after the discussion -- and mentioned that since tennis is basically "50 percent chess, 50 percent poker" the potential for tennis analytics is still quite high. They all agreed, however, that analytics cannot help the sorry state of U.S. Tennis at the moment. Martin briefly channeled Bill Murray: "If the game's not taught well at the beginning levels, it just doesn't matter." Americans are being taught how to hit, but not where to hit, and, as Annacone said, "You get 16-20 year olds trying to figure out how to play after eight years or so of bad habits." Exciting, huh?
Behold the Flying 20. Back at "Football Analytics," while Roth was begging to dump the 40-yard dash forever and always, Marathe pointed out that there does appear to be a drill better suited for evaluating receivers than the 40: the "Flying 20," which is the second half of your 40, once you have reached full speed. He said it will tell you a lot about their ability to separate from a defender. I personally would have to think that acceleration time (the first 20) would still matter, too, but it was an interesting point, one that I am sure will be discussed at Football Outsiders.
ESPN is by far the best thing going for sports statistics today … and also kind of the worst. Their influence was inescapable. The Worldwide Leader in Sports had employees on every panel, banners and tables in the hallways … they were omnipresent. And in some ways, that is fantastic. The bottom line is that analytics are not exactly an amazingly profitable topic but the company is pursuing it anyway, in part because it is good for the fans. To be sure, without ESPN's (and, in particular, Bill Simmons') influence, the conference would not have grown at its current near-exponential rate (from 175 attendees five years ago to 2,200 with a waiting list this year). Every nerd should appreciate what ESPN has done in this regard. At the same time, however, there was also an inescapable heavy-handedness at play, both in the omnipresent nature of the ESPN logo and in the way fans have been pummeled by new ESPN measures like Total QBR and the new college Basketball Performance Index (BPI). The approaches to these measures has been rather sound -- ESPN's analytics team, led by Dean Oliver, discussed the makeup of and lessons learned from the QBR rating in a panel Saturday afternoon -- but the delivery of these measures was so forceful, so absolute ("This is better than any measure you've ever seen, and you must use it") that it rather quickly turned off much of its target audience, nerd skeptics. When you are omnipresent and influential, you can get away with this, but it is not difficult to shift from benefactor to overlord, and ESPN toes the line every day. Still, their influence on this conference was almost certainly a net positive.
Every Dunkin Donuts employee in the country looks at me funny when I ask for only a little bit of cream in my extra large coffee. But I digress.