So it’s Friday before the Super Bowl, and you know what that means - hundreds of thousands of people pretending to work while getting the latest updates on what Bridget Moynahan may or may not have done to Tom Brady’s ankle, arguing about whether we’re going to see the “Manning Face” early in the game, and - of course - whether SpyGate will be an anecdote or an asteriks on New England’s potentially historic season.
Now on the SpyGate thing (I know this has nothing to do with the title of this post and is a little outdated, so you can skip the next few paragraphs and I won’t be offended), I’m on the side that thinks it got a little overblown, but there is still something very, very fishy about it.
For those that may have forgotten, here’s how the NFL investigation went: after New England was caught illegally taping the Jets defensive signals (for non-football fans, note knowing what the defense is going to do makes thinks a tad easier for the offense), and news started trickling out about how persavive the problem might have been, so the league ordered New England to turn over all of their tapes. Football fans waited to find out just how much cheating had happened, how many games might have been tainted, and what might be done about it. The NFL, within about 12 hours or so, decided to declare the case closed, and destroy all the tapes while issuing a small penalty.
Now that just seems odd doesn’t it? In what bizarre universe is immediately destroying the tapes synoymous with conducting an investigation? Can you imagine any other organization collecting all the evidence into what could be a systematic cheating method, and then just destroying it and saying their investigation was complete? If it really wasn’t that bad, wouldn’t it have been better to let people know what they had? Or maybe keep it around and, I don’t know, investigate for a few days? Suspicious - very, very suspicious. Let’s just say we won’t by featuring the NFL in any “transparency” presentations anytime soon.
Anyways, we’re (to put it mildly) not immune to talking about sports in this office, and Brendan and I were discussing the Patriots success, and the apparent involvement of a certain mathematical genius that New England employs to legally get an edge on their competition. Such analytics are slowly making their way into football, following the path that MoneyBall blazed in baseball. There are also people trying to get it better integrated into professional basketball, and here’s where I finally get to the point of my post.
The NY Times has a great article on Houston Rocket’s GM Daryl Morey (a new analytical whiz kid), and in it is a quote from Grizzlies GM Chris Wallace that puts the challenges faced by basketball in a light that might be interesting to wikinomics readers:
“The difficult thing to factor in is basketball is more of a collaborative, chemistry sport than baseball. Obviously, you need star power and great players. But a tremendous amount of your success is dictated by your interplay and synergy with the team — continuity, chemistry, feel for each other, those type of intangibles, which are very difficult to quantify.”
The jist of the issue is simple - while baseball is a team game, it lends it self neatly to statistical analysis because (for the most part) everyone performs in isolation. For example, when you’re up to bat, you’re pretty much on your own. In basketball, every action and outcome is dependent on a group of people working together (i.e. collaboration), which makes employing statistical and analytical tools much more difficult.
Sounds a lot like the problems many companies and government organizations face in the age of wikinomics, doesn’t it? When you’re trying to get everyone to work together as one, how do you effectively measure the contributions of each piece and figure out how to optimize everything simutaneously?
To further illustrate the basketball issue, I’m going to refer to my extensive experience as a B division player in the Mississauga Recreational Basketball league, where I’m the starting point guard on the returning champs… not so much because I’m good at “dribbling” and “passing” but more because I’m “too short to play anywhere else.”
Now the point of basketball is of course to score more points than other team, and many teams we play have a “black hole” on offense - when this player gets the ball, they’re almost always going to shoot as they try to “be like Mike”, and these teams are very easy to defend. However, almost everyone ridicules such players, and many have reacted by trying to “be like Lebron” - racking up both points and assists, with the latter being seen as indicative of playing within the team concept - i.e. collaborating.
However, a funny thing happens and these teams get easier to defend, as there is less real teamwork going on. While this seems counterintuitive, you know everyone on the other team will shoot as soon as they get a pass from the “black hole”, else they’ll get a stare down and/or never see the ball again. It’s what I like to call the selfish assist - it’s not about getting the best shot for the team, but ensuring that the shot that’s taken creates an assist for the passer. In turn, it’s not really collaboration and teamwork at all. In the former scenario, the other players were at least a little less predictable when they were sharing amongst themselves.
By far, what we find to be the hardest play to guard (and leads to the most success) is when a player in the post kicks it out to a guy on the wing, who swings it to the top, who immediately passes to one of the players on the other side, who almost invariably gets a good look - real teamwork. Who’s the “star” of this play? Arguably, the guy that passed to the guy who passed to the guy who passed to the guy who scored. While arguably you could measure “4th assists” to try to capture this, such a thing would almost certainly become a meaningless stat, and there was probably a couple of important picks in there that may or may not need to be accounted for, in addition to a myriad of other factors… hence the problem.
Or to sum it up another way, I played in a league that didn’t track stats and the ball movement by most teams would make almost any high school coach proud. Once I started playing in a league that tracked and published individual points, assists and rebounds, people started playing very differently.
This has become a reasonably well known (but difficult to quantify) issue in the NBA, as teams that have a single player dominating both assists and points typically do poorly (see Tracy McGrady never getting out of the first round of the playoffs, the Wizards doing better this year without Arenas, any team Stephan Marbury has played on, and many more examples.) While Lebron managed to surprise a lot of people by leading his team to the finals last year, it remains to be seen if this can be repeated (and notably he benefited from an extremely weak Eastern conference, and would likely have been first round and out in the West).
In fact, who they lost to in the finals - the San Antonio Spurs - are a great example of how stats can’t neccesarily measure true player value. Their star - Tim Duncan - is making a strong case for being the best big man of his generation among basketball purists. However, for fans that value players based on their statistical contributions to fantasy teams (which try to account for overall contributions by looking at numerous stat categories), he often ranks about 14th among PFs, behind lesser players such as Gasol, several Wallaces, and now even David West. Even worse, their amazing PG Tony Parker ranks 30th, which more or less makes him the “worst” starting PG (statistically) in the league. He’s amazing at running the team, but it’s run in a way that doesn’t neccesarily lead to him getting assists, which is generally seen as the key measure of a PG. In other words, he plays the position in extremely well, but the way we measure success at the position doesn’t always capture it.
Again, pretty easy to make the connection here to the problems many companies face in creating incentive structures tied to performance, non?
Now I’ve used basic fantasy stats here to make a point, but as the article notes teams are getting more sophisticated in the types of stats they measure - accounting for second assists, success rates on shots from passes, including substitution patterns, quality of competition, etc. As Morey says:
“We track everything imaginable. Each pick-and-roll, what’s the result of it? Each guy on the floor, how efficient they are. A lot of it, we end up not using. But we track it so that we have it available in case the question comes up where it becomes relevant.”
He also notes the team has made a significant investment in people and millions of dollars - but it remains to be seen how effective these stats are. According to what he has, Tracy McGrady (Houston’s star) led the league in passes that led to high-percentage shots; Houston has a dominant center in Yao Ming and many good supporting players; everything should add up to success one would think - but as noted, Tracy McGrady is known as Mr. “First round and out”, having never won a playoff series. There are things about basketball and the value of teamwork that are just hard to measure - as anyone who’s watched a team of European no-names whollop a group of U.S. stars can attest to.
This doesn’t mean I think the measurement is useless - in fact, I’m a huge believer in the ability of statistical analysis as a tool for optimizing all kinds of different decisions. But I am interested to see if a team ever gets to the holy grail of measuring collaborative contributions in something like the NBA. If they do, not only would it be a huge advantage, it would likely be a template that a myriad of different types of organizations should try to replicate.
Comments RSS
| TrackBack URI
Using advanced math to improve performance of pro sports is covered well in Competing on Analytics by Tom Davenport and Jeanne Harris.
http://www.amazon.com/Competing-Analytics-New-Science-Winning/dp/1422103323/ref=pd_bbs_sr_1?ie=UTF8&s=books&qid=1201893386&sr=8-1
Comment by Mike Dover - February 1, 2008 3:18 pm