« Behind The Basket Power Rankings 2.0 | Main | Why Isn’t Anyone Talking About… »
Saturday
Mar032012

It's Christmas Time For Analytics Fans

This color-coded chart from Kirk Goldsberry makes it clear how harmful shooting lots of long-2's can be. Someone should show this to league-leaders Kobe Bryant (8.2/gm), Gerald Henderson (6.4), and LaMarcus Aldridge (6.4).MIT’s Sloan School of Management is hosting their sixth annual Sloan Sports Analytics Conference (SSAC) this weekend, an event that brings together panelists and guests who discuss the importance of the increasing role of analytics (ie. all that Moneyball stuff) in sports. The conference has attracted hundreds of NBA owners, GM’s, coaches, and even some players throughout the years, not to mention numerous columnists and media personalities. The event has sold out 2,200 tickets this year and is now hosted in downtown Boston, away from MIT’s campus, in order to handle larger crowds. ESPN sponsors it, and the conference is even streamed live on-line. Bill Simmons dubbed the SSAC “Dorkapalooza,” but still attends—in fact, he’s actually been a panelist more times (5) than either Rockets GM Daryl Morey who co-chairs the event (4) and baseball Sabermetrics godfather Bill James (3). 

So what have you been missing? Here’s a complete list of the panels, a complete list of research papers that are being presented, and a complete list of the Evolution of Sport presentations. What will interest most fans reading a pro basketball site, however, will be links to what’s already been written about the presentations and discussions that were specifically geared toward the NBA. Here you go:

Basketball Prospectus’ Why We Use Stats: BP’s Kevin Pelton explains that “Using statistics is a tool to understand basketball better--and one of many. In my opinion, that's where the dividing line is truly drawn: Not between those who use stats and those who don't, but between people who are interested in learning more about the game and those who are not.” Here here.

Sporting News’ Building an NBA team at the Sloan Sports Analytics Conference: SN’s Sean Deveney discusses the findings of a research paper presented by Robert Ayer of MIT, who broke down teams’ top-3 players into specific roles (clusters) and analyzes which combinations produce the best team results. It turns out Lin, Carmelo, and Amar’e don’t fare very well in this regard, whereas a dominant center paired with a “multi-faceted, high scoring wing, with high assists for his position and a great 3-point shooter” is a terrific duo to start with.

Off the Dribble’s Mapping the NBA: OTD’s Joshua Brustein discusses the work of Harvard geography scholar Kirk Goldsberry, who turned every shot taken in the NBA for the past 5 years into a color-coded chart based on the points/shot for each of the 1,284 shot locations. The real insight to be gained from such a chart comes when teams use it to answer questions like “Where do the most steals occur? Where are the Pacers bad at defending against offensive rebounds? Where does Kevin Garnett tend to commit fouls?”.

ESPN’s Fact or fiction: Basketball analytics: There’s a 3-minute video with Henry Abbot and John Hollinger being really awkward while discussing free throw distractions, but the fact-or-fiction below it with 5 panelists is worth a read. They discuss wether fans should be unhappy if their teams aren’t investing heavily in analytics (5/5 fact), whether communicating analytics to coaches trumps the most cutting-edge data (4/5 fact), and whether analytics is getting out of control to the point fans should ignore it and just watch the games (5/5 fiction), among other topics.

Boston Globe’s Numbers don’t lie, teams use advanced analysis: BG’s Christopher Gasper looks at the role of advanced analytics within the Celtics (huge), mentions that 27 NBA teams have representatives at the SSAC, says that the behind-the-scenes “Moneyball” guys employed by teams “are no longer outside-the-box thinkers. They’re the box,” and points out that hockey is much further behind basketball in using analytics.

PrintView Printer Friendly Version

EmailEmail Article to Friend

References (10)

References allow you to track sources for this article, as well as articles that were written in response to this article.
  • Response
    Response: taxiskara.se
    cpace mcorus reformation clustering nasdaq rkjhk hkamkjksa andb infer extend real
  • Response
    Response: evabiotech.com
    a little something written here was absolutely Lots of great . an incredible I just want
  • Response
    Response: www.prpr.se
    There are some interesting points in this article but I dont know if I see all of them middle to heart. There is some validity but I will take hold judgment until I look into it further. Good article , thanks and we want more! Added to FeedBurner too.
  • Response
    Response: Pick 6 Leak Proof
    It's Christmas Time For Analytics Fans - Behind the Basket - The Antidote for Conventional Wisdom
  • Response
    It's Christmas Time For Analytics Fans - Behind the Basket - The Antidote for Conventional Wisdom
  • Response
    It's Christmas Time For Analytics Fans - Behind the Basket - The Antidote for Conventional Wisdom
  • Response
    It's Christmas Time For Analytics Fans - Behind the Basket - The Antidote for Conventional Wisdom
  • Response
    Response: wedding advice
    It's Christmas Time For Analytics Fans - Behind the Basket - The Antidote for Conventional Wisdom
  • Response
    Response: wedding advice
    It's Christmas Time For Analytics Fans - Behind the Basket - The Antidote for Conventional Wisdom
  • Response
    It's Christmas Time For Analytics Fans - Behind the Basket - The Antidote for Conventional Wisdom

Reader Comments (26)

Lets hope those that passionately utilize uber and over analysis of past performance from the box score instead of watching games understand that watching the game closely will always remain the best way to understand basketball. WHile these statistics, when properly used, are a great tool for measuring role and supporting players, they should always be used as a tertiary tool in building a team. After all, how were great teams built and great players discovered before these stats? Watching games carefully and understanding the system you want to run I would guess.

Imagine someone building an NBA team using only advanced stats and never watching games?

March 5, 2012 | Unregistered CommenterJamesD

Great links and will check out later when I have the time.

@JamesD,

I don't think anyone is suggesting building a team solely off stats or solely off "eye ball test" rather using both. In fact that is what is being used now. Scouts learn about good players from....you guessed it stats. Then they visit and use their "eye balls" to see if there are circumstances leading to those stats (bad competition, etc).

March 5, 2012 | Unregistered Commentergav

I would disagree with your post. I think there is a segment that wants to go strictly with advanced stats. I also don't think scouts learn by looking at stats first. They learn by going to games to check out player A. then they watch the games and maybe see another player.

Bottom line for me is when you analyze just the boxscore you miss context. I like to see how a player performs rather than looking at simply what he does. In advanced stats 6/10 is 6/10. Maybe you disagree but to me 6/10 in Boston is different from 6/10 in Sacremento. Its different when 20,00 people are screaming for you to miss then when 10,000 people are just there.watching.

March 5, 2012 | Unregistered CommenterJamesD

As I've preached forever, one has to use the proper (most appropriate and telling) context when evaluating a player or team by any method. Unfortunately 95% of fans barely consider context at all, let alone intelligently, so we end up with some crazy opinions that fans ignore all future facts/insight about.

March 5, 2012 | Registered CommenterZachariah Blott

Perhaps they ignore then because the people using the advanced stats say and write things like "95% of fans barely consider context, let alone intelligently". Fans that love the game and have watched for years probably don't like to be told they are watching it stupidly or incorrectly.

Its like the debate about clutchness. Henry Abbott gave a definition of clutch. Maybe fans who watch the game don't agree with that specific definition.Maybe they see "clutch" as a knack for hiting big shots when the team needs them. I don't know but I do know no one that has watched the game for years wants to be told what is good or bad by someone that simply crunches numbers and then creates definitions to fit their argument. Or maybe they simply dont care. Or maybe they like the "performance" aspect of it.

If someone likes watching DRose play, do you think care if his efg% and orb% are lower than another guy. So. They watch Drose and get excited when he has the ball, they don't get excited when the other person has it. That excitment, which CAN never be measured using statistical analysis, is perhaps what those 95% of fans watch for.

So its not necessarily a rejection of advanced stats its a rejection of those that claim using them makes them a superior watcher of the game.

March 5, 2012 | Unregistered CommenterJamesD

I still believe scouts regularly start with stats (rather than only hype) to see who to evaluate. Perhaps I"m wrong, just a hunch.

Also stats for a single game can be misleading, I admit, context (as Zach pointed out) is king. However when looking at the breadth of an entire season or career the stats will even themselves out. If player X is allegedly a great ball handler but is A/T ratio is horrible, there is obviously a disconnect between perception and reality.

Ironically at my place of work coworkers and I have been discussing stats vs. eyeball. One is strongly in favor of the eyeball test. The problem is this only works on games you've actually seen (which no one sees them all, no one), people's memories fade and only remember really good/really bad (think about past GF, they are all monsters, funny eh?), or they are looking through w/o any objectivity. We've debated on numerous occasions a certain player and their playoff performance. The eyeball test got brought up, the problem was I saw all the same games and had a different conclusions. What to do? I went to the stats, and lo and behold one of our memories was incorrect, the others was more accurate. The stats verify the eyeball test with objectivity.

Again, both are useful. I'm not suggesting to stop watching games and just look at box scores and advanced stats, what I am saying is those stats can prove/disprove subjective perceptions about games/teams/seasons/etc.

March 5, 2012 | Unregistered Commentergav

@James D,

I would never say someone isn't a fan of basketball if they didn't' look at stats. Or you can't like player X w/o looking at stats.

For me, I look at stats for three main reasons:

1. If I don't get to see a game.
2. I love numbers and statics.
3. When comparing players/teams.

I don't use stats to determine who I like the best, or who I like to watch.

March 5, 2012 | Unregistered Commentergav

@JamesD
You've hit on an important point; fans who don't care about advanced stats are those who like the performance aspect of the game. And that's fine, and that explains why street ball and dunk contests and All-Star Games are so popular. But it also makes these types of fans lousy evaluators of the game objectively. Everyone who enjoys watching the game does so for their own reason, and that's great, but I don't want fans who value cool looking baskets (which are often bad shots) over "boring" baskets (often the result of smart play) telling someone who seeks out more information to evaluate the game that they don't know what they're talking about. Less is not more when it comes to information.

March 5, 2012 | Registered CommenterZachariah Blott

Gav:

Respectfully, I dont think you and your friends (for example) discuss clutchness as only defined by Henry Abbott. You discuss it, my guess, like all hoops fans, within various vacums. not just one used to fit a particular argument.

Obviously stats to check historical accuracy is appropriate. But there is a difference between assist to turnover ratio- a standard definition used across the board, then "clutchness" for example which is not so clear. Checking if you are correct is not the same as defining a variable based on the theory you which to argue.

As to your next post I use numbers almost the same way. The exception, as I tried to illustrate in my last post, is the difference in achieving numbers given the environment. Advandced stats tend to make every player equal. In the NBA, with the emphaisi on entertainment, the differences in salaries, and the nature of the business (ex Kobe/Lebron ALWAYS playing on holidays since the league markets them vs. a Durant or Kevin Love that don't have to deal with being the face of the league yet), its hard for me, and others, to go too much on the statistical end because those types of non basketball related pressures don't apply to all.

March 5, 2012 | Unregistered CommenterJamesD

@Zach

The problem with your statement is that there are fans and scouts who never relied on advanced stats and yet do view the game objectively. You (and some like you) in essence are telling people how to view the game objectively correctly when you are not watching the game. Think people are going to embrace that?

There is nothing wrong with seeking more information about the game. There is something wrong, IMO, with using advanced metrics to decide certain awards like MVP, since that award isn't just about stats or with defining a variable to suit the argument you plan to make or your individual biasies.

March 5, 2012 | Unregistered CommenterJamesD

@JamesD,

When we discussed "clutch" I tried to get a working definition but the other parties were reluctant to bring any forth.

I brought forth two separate sites that did exhaustive "clutch" statistics, both defining it in different ways, both disproving their assertions. Of course they were simply rejected out of hand.

Finally I went to the play by play for games showing my point, again rejected for "eye ball test" when I watched those exact games myself.

This is one of those cases, that stats are necessarily. This isn't a popularity contest where stats mean thing, this was who was better at certain parts of the game. This can be objectively verified and proven by many measures.

Again, I look to stats when comparing players, not determining likes/dislikes/favorites and the likes. I think they are very valuable in this area. Eyeball tests are great, but very subjective and suddenly cancel themselves out real quick when two people saw the same game and came to wildly different conclusions.

March 5, 2012 | Unregistered Commentergav

@JamesD

Speaking of MVP, you are right it has nothing to do with stats, that is why I put it in the "popularity" awards and I put little to no value to it when comparing players.

To go on an aside, I wish they'd do away with the award altogether and replace it to Most Valuable Position X. So they'd have most valuable center, most valuable point guard, etc. I hate comparing centers to shooting guards to power forwards when they have such different roles and different ways to determine they are great in their positions.

March 5, 2012 | Unregistered Commentergav

@JamesD
There are indeed scouts who evaluate the game with no stats whatsover, and we also have a complete jumbled mess of a draft most years (not to mention really stupid free agent signings), and that's chosen by the eyeball experts.

I watch a ton of games, so don't make up information about the opposite.

What did eyeball fans think of Shane Battier for the longest time? Not too much, that's what. Then that article came out a few years ago describing the immense impact he's making on teams, particularly at the defensive end, and now a bunch of fans actually talk about the "no-stats All-Stars" and look for and recognize what Battier is doing so well. It took those advanced matrices for people to even know what to look for, that's why it scares me that many fans want to totally dismiss this data simply because they're not seeing the small things materialize into points on a floor with 10 players and 1 ball all moving at once.

March 5, 2012 | Registered CommenterZachariah Blott

@Zach

There were smart fans and scouts before advanced stats and there will be after as well. It absolutely did not take advanced stats for anyone who truly watched and understands the game reshape their opinion of Shane Battier. He is a very good defensive player can make the open shot. His value to a team is as much financial as it is statistical. He can't create his own shot, he doesn't create for others, and is not someone who makes others around him better. If Shane wanted a max salary contract for what he contributes he surely wouldn't recuive it and the GM who gave it to him would get fired.

March 5, 2012 | Unregistered CommenterJamesD

1) There were smart scouts before advanced stats, but they were the few who knew what they were looking for when it came to finding players who helped their teams win. Read some of the books about Red, and you'll see what I mean.
2) There will be no after.
3) Fans as a whole never talked about Battier in a positive light before that article. They now love him and discuss the little things he does. It took that article for the massive turnaround. You can't pretend that he's not talked about and watched totally differently now than he was before that article.

March 5, 2012 | Registered CommenterZachariah Blott

@Gav

I don't disagree with your theory about using stats to verify eyewitness accounts of events. When comparing players its of course useful, but you still lose context because advanced metrics treat every shot by every player in every arena the same. Advacned stats create as many false conclusions as they do enhanced information.

A point guard whose advanced stats are not statistically different (within error) than DROSE adv stats can't just be plugged into his position and the team will do as well or better. A 2 guard with a better asst/TO ratio isn't going to make the Lakers better than they are with Kobe Bryant.

March 5, 2012 | Unregistered CommenterJamesD

@JamesD

Swapping players in and out of teams and expecting same results is something completely different than what I was discussing. This doesn't work not because stats aren't reliable, but since practicing with your team is essential in a team sport.

If you want to look at players impacts on their team, the best (not perfect) way would to look how the team does w/ and w/o them and/or other key players.

March 5, 2012 | Unregistered Commentergav

@Zach

Shane was picked like 6th in the draft and was put in a position he couldn't excel in, that being a starter and the 2nd scorer behind Pau. So fans talked more about his failure in that role because it was relevant to how he performed as an NBA player. He got to Houston, took less shots and took on the role as the defensive stopper. They talk about the little things because thats really all he does.

March 5, 2012 | Unregistered CommenterJamesD

@Gav

Not saying that was your argument, but this is what is implied when Zach and others use certain advanced metrics. Have no problem with using stats to check historical accuracy or verify what happened during a game.

But the player that does better on advanced metrics may not be a better player. Think John Stockton and Gary Payton.

March 5, 2012 | Unregistered CommenterJamesD

@JamesD
I think you're oversimplifying what stats and computations are available based on your comment that "advanced metrics treat every shot by every player in every arena the same." You told me a while ago that Kobe's FG% was low because he took lots of shots late in the shot clock, but then I was able to show you the data that his teammates took the vast majority of those shots (and it's becoming more clear and well-known recently that Kobe often dribbles out the clock and then passes it to someone else in the final couple seconds, ensuring a bad Laker shot/possession). As for Battier, I'll repeat what I said before: "You can't pretend that he's not talked about and watched totally differently now than he was before that article." Fans talk about the little things he does WAY more than the other guys who do the little things, and it's because of that article and all the stats they used for him. If the media hypes Kevin Love's long outlet passes before he enters the league, people will magically notice them more than those by other players, even if Love isn't the best at it.

March 5, 2012 | Registered CommenterZachariah Blott

@ Zach

That is patently untrue. I never attributed Kobe's FG% to ONLY taking tough shots at the end of the shot clock, I used that as an example as to why it is. And I never said nor implied it was lawyas the other guys fault. You interpretd that way unfortunately because of your bias toward Kobe.

Keep in mind you said the same thing about how I characjhterized Amare SToudamire for MVP.

You are confusing sportscasters and basketball TV people with all fans. The guys I talk to about ball now who Shane is, Again, I mind you to be careful of how you label people in order to make your prefernce seem more enlightened.

March 5, 2012 | Unregistered CommenterJamesD

Of course the guys you talk to about ball know who Shane is; he was an NCAA champion, a lottery pick, and is on the Heat. I'm saying that the way fans talked about him and pay attention to the little things he does when he's not getting stats is FAR different than it was before that article.

So it was only supposed to be part of the reason you felt his FG% and eFG% were low. It still turned out he does the exact opposite of what you thought he did, which has to really make one question the use of the eyeball test to judge greatness. If you're a fan and watch a bunch of his games and got the opposite impression of how Kobe performs than it turns out is the truth...well let's just say it wouldn't be the first time someone had to tell a Kobe fan what the truth is.

March 5, 2012 | Registered CommenterZachariah Blott

There you go with the "Hey we use stats so we see the game better stuff." People who pay attention paid attention long before the advanced stats. You simply glossed over what I said about him in Memphis-he was unsuited for the role they had him playing. It didn't take advanced stats to make that clear.

Of course this isn't a Kobe discussion but of course you try to make it one. When it comes to the eyeball test Kobe is one of the ten greatest players who has ever played. The eyeball test tells me he has 5 rings and 2 finals MVP's. I'll go with what other great players say about him rather than a haters advanced stats. Besides, its not a Kobe discussion.

And please stop acting as if every fan read a shane battier article.

March 6, 2012 | Unregistered CommenterJamesD

So Shane's defense wasn't showing up as MUCH better in the advanced stats than the normal stats (steals, blocks) when he was in Memphis? Again, you're pretending that article didn't completely change how normal fans view what he and a few other role players the media have latched onto (Nick Collison being the latest du jour) are doing. The term "no-stats All-Stars" certainly wasn't used before then.

Kobe only comes up in a discussion about the eyeball test because you're a huge fan, you obviously watch a lot of his games, yet your eyeball test about how he plays is so far off you honestly believed a statement about his shot selection that was THE EXACT OPPOSITE of the truth. It's very similar to the whole clutch thing, where literally not a single objective measure of his clutch (and it's been defined in a zillion ways, with Kobe fans never providing an alternate) supports the eyeball tests that seem to be based on a reputation based on a couple overplayed highlights. Of course he'd come up in this type of conversation: 1) his eyeball vs facts comparison contains the largest gulf, and b) you've seen a ton of his games, so your eyeball test is worth noting when it's completely wrong if you're trying to support it.

March 6, 2012 | Registered CommenterZachariah Blott

His stats and how people watch the game are two completely different things. Shane played good defense Sunday but his stats didn't reflect it. But if you watched the game you saw this. If you are only a stats checker well you missed it. At the end of the day, he got torched by a player whose offense was better than his defense. If the advanced stats tell you anything else then they are worthless.

That term isn't in every day discussions so why you feel the need to over emphasize that one article is beyond me.

I am a Lakers fan, who isn't a big Kobe fan, who deosn't like to see people like you let your hatred of him cause you to minimize his contributions. I am not a huge Kobe fan at all and have made that clear to you. You label me that because you are a Kobe hater. Your method, which the history of articles reflects, is you label anyone who disagrees with you a "huge kobe fan" to shift the bias from you to them. You have an unfortunate tendency to purposely mischarachterize a poster's point when it doesn't agree with yours. I never believed anything of the sort. I pointed that out as one reason during a game Kobe may take a bad shot. As for clutchness, I simply said if you use Abbott's definition, then you'd ultimately have to choose the person whose FG% is the highest.

Like I said this isn't a kobe discussion. So there is no need to continually make it one.

March 7, 2012 | Unregistered CommenterJamesD

@JamesD
By mischaracterizing it with facts and stats they don't like because they're normally happy with the surface level ones that tell us a very small piece of the puzzle with no context? You've made it clear you're not a Kobe fan by a) being a huge Laker fan, and b) defending him to the hilt against all factual evidence, including LA's better record when he sits (which you'd think strictly an LA fan would be fine with) -- sorry I got the wrong impression from that kind of track record. And don't downplay it like you were talking about A bad shot he MAY have taken during A game: you said, "Think about how many times at the end of the shot clock or quarter Kobe shoots a 3 with the clock winding down. Each time he does this and misses it negatively effects his efg%." OK, sounds like you think it happens a lot. Guess what, stats prove that it doesn't. In fact, the way he plays actually causes all of his teammates to be forced into those bad shots much more regularly than for him. If the way you watch the games made you believe the opposite of that, yet the eyeball test of a knowledgeable fan is your go-to evaluation, then you might want to re-evaluate how you get your impression of which players are actually contributing value to their team. And that's assuming you watch a really high percentage of all the games (1 entire game every night = 11% of all NBA games) so your eyeball test can be considered valid/significant.

March 7, 2012 | Registered CommenterZachariah Blott

PostPost a New Comment

Enter your information below to add a new comment.

My response is on my own website »
Author Email (optional):
Author URL (optional):
Post:
 
Some HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>