Advanced sports analytics still face massive pushback - NBC Sports

Advanced sports analytics still face massive pushback
Teams must move beyond gut instinct - yet when there's resistance to new information at a conference for advanced analysis, it's a sign of incredible resistance to change
Reuters
Ravens kicker Justin Tucker was stopped during a fake field goal attempt in Super Bowl XLVII. Would Baltimore have been better off simply trying the kick? Advanced analytics helps answer that question.
March 6, 2013, 1:03 pm

The overriding story, of course, was how big the numbers have become. Here it was, the Seventh Annual MIT/Sloan Sports Analytics Conference. And it's astounding how much this thing has grown.

A quick history: In 2007, Darryl Morey (now Houston Rockets GM) and Jessica Gelman (now vice president for marketing at the Kraft Sports Group) thought it would be fun to put together a little conference to talk about sports analytics. At the time, it was a pretty out-there idea. They connected it with MIT, for obvious reasons.

There were nine panels. The keynote speakers were J.P. Ricciardi, since fired as Toronto Blue Jays GM, and Jamie McCourt, since divorced from Frank McCourt, who subsequently took the Dodgers into bankruptcy. About 100 people showed up.

This year? More than 2,700 people attended. Everyone was there, from Dallas Mavericks owner Mark Cuban to political Nostradamus Nate Silver to Moneyball author Michael Lewis (and, for the record, I was on a panel as well). There were more than 100 speakers (including two named Brian Burke, one a senior advisor for the Toronto Maple Leafs, the other the founder of Advanced NFL Stats). There were more than 75 panel discussions, including such provocative titles as these:

  • The Hidden Foundation of Field Vision in English Premier League Soccer Players.
  • Going for Three: Predicting Field Goal Success with Logistic Regression.
  • Now Analytics and Big Data are Impacting the Evolution of the Fan Experience.
  • KORE Software: Data Mining and Analytics: Turning Insight into Action using CRM.

Yes, the clear and principal story of the conference is how analytics are advancing and changing everything you can imagine about sports - from stuff on the field to the fan experience to TV coverage to this column. Every single team in professional sports has access to previously unimaginable information and statistics, and all of them are making at least a token effort to understanding it. Some teams are a lot more with it than others, of course.

The hottest trend is called "Big Data," where cameras and various software programs can collect and break down in breathtaking detail literally each second and each action of a sporting event. You can determine precisely how proficient basketball players are at setting screens and what football formations yield the best results and how fast you need to hit a ground ball for it to become a base hit.

But the little story was as fascinating as the big one. Many people are still fighting back against the numbers - even at the SLOAN SPORTS ANALYTICS CONFERENCE.

"KYP!" ESPN commentator and former NFL head coach Herm Edwards said at a panel discussion called "Football Analytics." "Know Your Personnel. That's what it comes down to. The numbers are the numbers. But you better know your players."

"This isn't a video game," former Orlando Magic coach Stan Van Gundy said as he explained that coaching isn't about numbers as much as it is about getting players to play their best.

"You can't measure that," ESPN baseball writer Buster Olney said when talking about the idea that former Tigers pitcher Jack Morris would "pitch to the score." The pitch-to-the-score concept theorizes that Morris, being a bulldog of a pitcher, would pitch his best when the game was close and ease up when the game was out of hand.

This kind of pitching, you might expect, would have led to numerous one-run victories and fairly clear patterns. Instead, several detailed game-by-game review by several analysts showed there is little to no evidence that Morris pitched to the score at all. Olney said that you can't trust the numbers on that because pitching to the score might only involved one or two pitches in any particular game.

And so on.

Point is, even at a conference about using the numbers and analytics in new ways, there is significant backlash against using the numbers even in the current ways.

< Prev | 1 | 2

This is natural. Sports are physical. At the highest level, they demand a kind of physicality - speed, strength, power, willpower, commitment, nerve, tolerance of pain, countless hours of practice - that eludes analytics and defies systematic thinking. No amount of number-crunching can inspire someone to stand in the pocket against the rush or foul off a 99-mph fastball when they were looking for the curve.

You could understand, to an extent, athletes in the arena and coaches under the microscope being dismissive of those people calculating odds and percentages and best practices while sitting far from the fray in what Herm Edwards dismissively calls their "air-conditioned offices."

"Nobody gets fired for typing in the wrong number," Edwards said.

OK. But ignore the math at your peril. One of the great conceits of sports has always been this idea that it's better to trust the eyes and the gut. But the eyes are often looking the wrong way - Bill James points out that the difference between a .275 hitter and .300 hitters is one hit every two weeks or so.

Eleven hits in 40 at-bats equals a .275 batting average. Twelve hits in 40 at-bats equals a .300 batting average.

You can't see that - not without keeping a running total. And that's batting average, one of the simplest and least useful statistic. If you couldn't legitimately tell a .275 hitter from a .300 hitter without keeping totals, how in the heck are you going to see which center fielder has the best range or which NBA player has the best on-ball defense or which cornerback is best in man-to-man coverage.

The eyes might see things the analytics miss; sure, nobody denies that. But the analytics definitely see a whole lot the eye misses.

And what the eyes miss, the gut miscalculates.

The data is getting more and more comprehensive - and, yes, there are dangers in that too. It's easy to become too reliant on data or misread the numbers or lose some of the humanity of the games. People chime those warnings all the time - repeatedly at Sloan.

But in the end, the greater danger is being too sure and getting left behind. Some of the new analytics are complicated and scary. But most of it is sensible if you are willing to open your mind.

An example: On the Football Analytics panel, NFL Stats founder Brian Burke explained what the numbers said about the Baltimore Ravens' fake field-goal attempt near the end of the first half of the Super Bowl. It was 4th and 9 and Baltimore led San Francisco 14-3.

But Herm Edwards and former Jacksonville Jaguars coach Jack Del Rio had questions. Lots of them. What if instead of a fake field goal, they had tried a conventional fourth-down play? Would that have made a difference? What if the Ravens had been able to block Patrick Willis (who made the tackle)? What if the Ravens had a different quarterback? What if the 49ers had a different quarterback?

Burke looked bewildered. He tried to explain that he wasn't trying to coach nor call the play nor make judgments about the Ravens or 49ers. He just wanted to give that bit of information - based on the league's history, unless you are 65 percent sure you will get the first down, kicking would give your team its best chance to win. (It should be noted - and was noted more than once by Herm Edwards - that the Ravens did win the game.)

It seems like coaches should love to have that kind of information to use however they see fit. Many do. But, for all the talk about the changing world, it's pretty clear that many still don't.

Joe Posnanski is the national columnist for NBC Sports. Follow him on Twitter @JPosnanski. Click here to subscribe to Joe's stories.

< Prev | 1 | 2

Joe Posnanski is the national columnist for NBC Sports. Follow him on twitter @JPosnanski



Channel Finder