My Journey to the MIT Sloan Sports Analytics Conference
(If you missed part one, click here)
One of the best parts of the weekend was seeing all of the other bloggers, sports writers and other sports figures who were attending the conference. I mentioned Masai Ujiri and some of the other team employees in the last article. There were people like Kevin Pelton, Sebastian Pruiti and Ben Golliver; people whose work I had read for a while, but had never gotten to meet in person. There were also ESPN guys like David Thorpe, Henry Abbott and John Hollinger.
One funny story: At the end of the conference, the line for the coat check was pretty long. I got in line and remarked to the gentleman in front of me that we might be here for another day just to get our coats back. We started chatting and had a very nice 10-minute conversation . I didn't recognize him, but he kept mentioning Indianapolis, which I thought was a little random. Then I asked him what he did for a living. He gave me a quizzical look and said, "I work for the Colts."
"Really? That's cool, what do you do there?"
"I coach the team."
I had been talking to Indianapolis Colts head coach Jim Caldwell the entire time without knowing it. I'm not a huge football fan, but I almost feel compelled to root for the Colts from now on because of how nice a guy Mr. Caldwell was.
Anyhow, after lunch with Apolo Anton Ohno, I had three more panels to attend that day:
Panel Three: The Decision. How Players and Teams Will Choose in the Future
Moderator: ESPN writer and PTI co-host Michael Wilbon
Panelists: San Antonio Spurs General Manager RC Buford; President and GM of the Toronto Maple Leafs, Brian Burke; CEO and Founder of Priority Sports & Entertainment, Mark Bartelstein; and former player Donny Marshall
(Bill Polian of the Indianapolis Colts was supposed to be there, but since he had to attend to CBA negotiations that day, Marshall took his place)
This panel was intended to discuss the future of sports, how players will move between teams and how those moves should be covered by the sports media. I thought Wilbon was a great moderator, and although I know next-to-nothing about hockey, Brian Burke may have been the star of the conference.
Did 'The Decision' change what will happen with players in the future?
Donny Marshall remarked that the NBA is a "softer league" today. He said that players in his day "wanted to compete against the best", but today, athletes are asking, "what's the easiest way to get [to a championship]?"
Bartelstein added that, "It's a different culture now. In the AAU world, the top guys already know each other and have played together for years. Players on the other team are very friendly after games, when it definitely was not that way in the past."
With a smile, Burke interject that, "the NHL is still the toughest league. Hostility is at a premium. There is more fraternization than before, but much less than other leagues." He told a story about how a rookie player got yelled at during halftime of a game by team captains for shaking hands with an opposing player before the game started.
R.C. Buford was not as critical. "Free agency is good for professional sports. There will not be many times that 'The Decision' can happen." He said that while teams like Cleveland and Denver may suffer in the short term, that trying to overcorrect for something that does not happen very often would do more harm than good.
Brian Burke said, "I admire when guys take less money to win. Teemu Selanne took pay cuts to play for the [Colorado] Avalanche. It doesn't always work." He added that he did not like the 'theater' of 'The Decision', but that its very rare for anyone to take less money to win."
As an agent, Mark Bartelstein had a unique perspective with athletes. He said that there are three things that players want:
Typically in that order. He says he gets complaints from athletes he represents after every game. A player hates his coach. A player doesn't get along with his teammates. A player thinks he should've gotten more touches.
"There is a sense of entitlement with certain players. I give advice everyday, but I work for players, not the other way around. I'll tell them the situation and give advice, but if they make a different decision, I have to comply with it."
Marshall remarked that even players at the end of the bench still feel that they are entitled, especially to money. They will complain just as much as the superstars on the team.
Are small markets at a disadvantage?
Marshall: "The economics are huge [when talking about small market vs. big market]. If Dwyane Wade was playing in Minnesota, I'm not sure LeBron and Bosh join him."
Burke: "Over time, the small market stuff is lethal. If this is a trend, small markets will be killed. Only a few teams will be shown on TV. In the NFL, every team believes it can win at the beginning of the season and nearly every game is up for grabs. In the NBA, there is not as much anticipation. Fans want to buy a ticket and believe that their team has a chance."
Bartelstein made a good point that, "the NFL and the NBA are very different. In the NFL, players are trying to earn a job every day. Most players are replaceable. It's not the same in the NBA."
How much is the media responsible for the current situation?
Wilbon said that while the media has a responsibility to entertain, more and more athletes have a thirst for controlling their own image. With Twitter and other social media, many players are doing just that and are doing it without traditional media. They are going right to their fans.
Bartelstein added that, "Players want to know about their branding. I've had backup left tackles want to know. There is a lot more that you can do, too. When we had Kurt Warner on Dancing With the Stars, it gave him a lot of exposure and helped him gain fans among people who had never even seen him play football."
"The media can also be dangerous. There is so much competition and a race for information. Some guys are starting to give generic answers because they are scared."
Talking about Donny Marshall's "soft comment".
Wilbon said to Marshall that "not everything that's old is good and vice versa." He talked about how he enjoyed Olympic hockey much more than NHL hockey because there was less fighting.
Burke quipped that "the guys who are best at fighting didn't get invited."
Marshall said that his statement wasn't necessarily about physical toughness and that guys are not as mentally tough as they were in his day either.
More Burke, "I want a coach who is difficult. In my days as a [minor league hockey player], I got 2 1/2 years of largely negative feedback. The game was never designed to be fun. We want it to be difficult. [In Toronto], this is not a fun place to play, but if we win enough, that will be fun."
Panel Four: Basketball Analytics
Moderator: Senior ESPN NBA writer Marc Stein
Panelists: ESPN columnist John Hollinger; Former NBA GM Kevin Pritchard; NBA owner Mark Cuban; and Assistant Executive Director of the Boston Celtics, Mike Zarren
Marc Stein opened the panel by saying that, despite some people who are still holding out, basketball analytics can no longer be ignored. He then talked about the mid-season trades that went down and asked how much of a factor analytics are in those types of deals.
Cuban, who was definitely the biggest talker of the group, answered first. "You look at analytics way before a trade. Of course, that is for trades trying to improve a team and not all trades are [trying to improve a team]. Unfortunately, the NBA is not an efficient market for trades. Teams have different relationships and trust factors. Some teams will not talk to one another."
MIke Zarren added that, "we might hear about a deal and think 'we could've done better than that' but the team never called to find out."
Cuban: "Things go down to the last minute [at the trade deadline] due to lack of trust of analytics." He said that teams with analytics already know what players are good and don't have to make snap decisions at the deadline about adding or subtracting elements from a trade.
Kevin Pritchard, half-jokingly, said that "it might be better to have an eBay system" for trading rather than the current one. "Communication between teams is garbled. I rank opportunities and try to make the best one happen. If the first opportunity is gone, go to the second one. By the time you get to the third or fourth, you might be at a deal that you don't want to make."
He also joked that when he is confused, "I go to ESPN trade checker. If I win the trade, I do it."
What goes into the equation?
"Five years ago," Pritchard added, "only about 5% of the league discussed analytics in trades. How do you talk analytics to a regular GM without a math background? Analytics is a part of it but the gut is still a large part as well."
Cuban: "It's about risk. You're looking at three things: Current performance, future performance and financials. It's all becomes part of the algorithm, not just the current impact. Some GMs just want to keep their jobs, though, so the future isn't always as important."
Zerran: "We're always introducing new stuff into the equation. Scouting is still data."
Cuban: "It has to pass the eye approach, too, since not all data fits. You have to be comfortable in each piece."
Zerran mentioned that team chemistry is also an important element and that is where coaching comes in.
Pritchard: "We had many people doing analytics. After talking with other 'stats' teams, we're mostly after the same players."
John Hollinger, who was fairly quiet for most of the panel with all of the big dogs talking back and forth, quipped, "have [Daryl] Morey explain why he trades with Memphis each year."
How does this figure into analyzing prospects?
Cuban answered that "teams fall in and out of love with prospects and players. At deadline, teams have more current data."
(Talking about Corey Brewer) "Bad teams can test people. On a good team, Brewer's role will be different. He's going to get more open shots. He can focus on man defense because he has Tyson Chandler behind him to block shots. He'll look better on both ends because of his teammates. You have to plug in numbers with a new role in mind. I love advanced +/- for some of that. Win Shares are good for player evaluation, but not very good for lineups.
"If I want a guy in a trade, I'll often watch and root for him to do poorly [so that the other GM undervalues him]."
How long do you get before you say, "projections are off"?
Pritchard: "Depends on age. Brandon Roy has a shorter window at 22 than Derrick Favors at 19. Big men tend to develop slower than guards. Look at LaMarcus Aldridge. However, there's no way to be good without taking risks...unless you get Tim Duncan."
How do analytics apply in real time? Will we eventually see assistants with iPads on the bench?
Cuban: "We have a guy behind the bench who takes notes with analytics and then breaks it down for Rick Carlisle. The coaches don't have to know analytics themselves."
Zerran: "Analytics matter, but you need to know what the coach does. You shouldn't make any decisions based on 3/4 games."
Cuban: "There are so many sportsmanship and strategic elements to the game. That's why there isn't a 'Moneyball' option for basketball. For every action, there's a reaction that is very difficult to quantify. The best stuff comes over a period of time."
Zerran: "People don't know what variance is. That's very important. What's the impact %? It's not Kevin Garnett vs. Tim Duncan. It's how well will a player help help our team.
Pritchard: "Analytics have to be prepared to change with the game. Changes in the hand check rule changed things. We have to take that into account."
Is there still resistance to analytics among NBA teams?
Zerran says that there are 10 teams that don't devote major resources to advanced analytics. "You still need people who watch basketball for a long time because the numbers are noisy. Scouting is not going away. Moneyball created a false impression that the stats community ignores everything else."
Cuban offered that, "we tried that and it doesn't work. It's not the data, its how you use it and take advantage of it. Take veteran teams. Sometimes they get bored and might not play up to a certain level. How do you rank that stuff? We put as much emphasis on psycho analytics as statistical analytics."
Hollinger, who hadn't been heard from in a while, joked that, "if you torture a stat long enough, it will tell you anything."
The panel opened up to questions. I asked the question about David Kahn's "analytics are less important for rebuilding teams" quote and whether they would agree or disagree.
Cuban quickly said that, "I'm glad he thinks that way." He came back to say that, "in the Timberwolves' defense, the worse place you can be in the NBA is mired in mediocrity."
Prtichard called it the "treadmill of mediocrity".
Zerran offered that it can be tougher for rebuilding teams, since analytics will not help you find the next Kevin Garnett, which is ultimately what will help a team to get to the next level.
Cuban and Hollinger both thought that analytics could be more helpful for rebuilding teams since they could find great role players. Cuban brought up Brewer again, "we signed Brewer because we think he'll help this year and in future years. He's a bargain even with the luxury tax."
The panel ended with a question about why NBA players have such difficulties with free throws.
Kevin Pritchard seemed irked by the question and responded to the person by saying, "Here. Get in front of 20,000 fans, TVs and lights. Make sure you have a high heart rate after running up and down a court. It'll be a little different. It's not 'just a free throw'. (He asked the guy how many he thinks he'd make out of 10 free throws. The guy responded, '6'.) "Now, you've made your 6. What if I gave you 10 more shots and said, 'you have to make 6 or more again or you'll die'? How well would you do?"
Zerran closed the panel with his favorite Shaq quote about free throws: "Maybe God does it to keep me humble."
Panel Five: Gut vs. Data: How Do Coaches Make Decisions?
Moderator: New York Times columnist Howard Beck
Panelists: Former NBA head coach Del Harris; co-owner of the Boston Celtics Steve Pagliuca; NFL head coach Eric Mangini; and former Texas Tech head coach Mike Leach
The final panel of the day discussed how coaches make decisions. Beck opened with the obvious question of how much coaches use data vs. gut.
Leach opened the comments. "I try to assemble as much as possible. I use more data than gut, but not all data. I still break down game film. Gut is more useful as the game unfolds, while data is more useful to prepare for the game. Since sports is played by people, you have to account for emotions and approach things aggressively."
Harris: "There is less prep time in the NBA than in football. Decisions come quicker. But you still have to decide ahead of time what your major decisions will be. Practice end-of-game stuff. Practice game scenarios. I never decide about intentional fouling at the time. We already know under what conditions we will foul. We know in close games who will take last shots."
Mangini countered that there are a lot of decisions that come up that data can wrongly influenced. He brought up onside kicking in certain football situations, and that while the data says not to do that, the element of surprise can work by going against the data.
R.C. Buford believes that a team should, "weigh strategy early and follow it every time. When you do not have an alignment with your data and your ears, you need a philosophy to fall back on. Fouling up 3 late in a game, for instance. It works according to the data. When it doesn't work, its probably at the worst time. Most coaches have a philosophy of not to foul because its easy to remember the times when it doesn't work."
Are confidence and momentum quantifiable and real?
Leach said that while he can't define momentum, it clearly exists. He sees it as a sort of 'collective confidence'.
Harris believes that preparation gives a player or a team confidence. He said that he'll often run last-second shot plays with no defense so that players see the shot go down and get more confidence that it will work.
Del Harris also said that back in the 1960s, he had an assistant keep points per possession data. He would then use that data to figure out pace and momentum. Any time his team failed to score 3 or more times in a row, he had an assistant tell him. Likewise, if the other team scored 3 times or more in a row. He believes that is when momentum shifts. When his team has momentum, he said that it will override almost everything with one exception: you need finishers at the end of the game, no matter how well the bench guys are playing."
Buford added that, "[the Spurs] are known for 'diamonds in the rough'. I feel confident in our processes. We're not doing the most or the least. Data is better at elimination than prediction. You fit data into the system that helps you make gut decisions. With DeJuan Blair, we believed that 3 or 4 years of Blair on a rookie scale contract, with the risk of losing him to injury after that, is better than overpaying a worse player for longer."
Finally, a rant from Del Harris after Kobe's rookie playoff game against the Utah Jazz came up:
"I want to dispel a myth. One of my claims to fame is I was Kobe's first coach. A common myth is that Kobe's airballs in that game caused us to lose. That's not true. We got the ball three times in the last minute. The first time, I went with Eddie Jones. He missed the shot. The second time, we went to Jones again and Greg Ostertag blocked his layup. The third time, well, I can't go to Eddie again. We wanted Shaq down low for the rebound. Horry is out of the game with an injury. We know Kobe can get a shot, so we go to him. He gets the shot, but it rims in and out. Tie game and we go to OT. In OT, the Jazz went up on us and the game wasn't much in doubt anymore. At that point, Kobe shot his two airballs. But not in the clutch."
After the Panels...
When the panels ended for the day, the organizers had appetizers and drinks for everyone. It was a good spread, too. Some fancy cheeses, mini-quiche and figs. We're not talking hot wings and mozzarella cheese sticks. I chatted with some of the other attendees and then left a little early. I had a game to catch.
The Boston Celtics Game
Since I was in Boston, and the Celtics had a game that night, I figured I'd get a ticket and enjoy a game. The Celtics were playing the Golden State Warriors and I was able to find an upper deck seat. It was already a weird feeling going to the game. Instead of walking through a mostly-empty Minneapolis skyway, I walked through the streets of Boston with a crowd of fans who were already excited about the game.
Have you ever seen a line to enter the Target Center that was this long?
Although its not the original Boston Garden, I was still impressed by the place. How could anyone not be with that many championship banners hanging from the rafters? It was definitely a different atmosphere.
The game was a good one. The Celtics jumped out early after Ray Allen started hitting every shot he took. Jeff Green played well for them and Nenad Krstic had a few good moments. Troy Murphy looked a little lost. Rajon Rondo is a brilliant playmaker. For the Warriors, David Lee and Monta Ellis helped lead a comeback and the game was close at the end. In typically Celtics fashion, however, Paul Pierce closed it out with a three of his own.
It was really strange to see a completely full upper deck in a game against the Warriors.
Also surprising? Someone willing to wear an Antoine Walker jersey in public.
(To Be Continued...)
In Part Three: The last day of the conference and the geek after party.