Bonanza for NBA Stat-Heads

[tommy]

So the NBA is going to install motion tracking cameras at all of its arenas starting this coming year, with the intent to be able to provide extremely detailed statistical analysis of all aspects of the game of basketball. Including refereeing. For example, from the article:

Want to see how successful Ricky Rubio (http://espn.go.com/nba/player/_/id/4011/ricky-rubio) was at guarding Russell Westbrook (http://espn.go.com/nba/player/_/id/3468/russell-westbrook)? The system could break down the shooting percentages and results of each head-to-head possession.
Want to get an idea how close to 100 percent Kevin Love (http://espn.go.com/nba/player/_/id/3449/kevin-love) was in his first few games back from a broken hand? The system could send information to the team trainers and doctors about his endurance and how quickly he is tiring during a game, thereby painting the most accurate picture possible of his recovery.
Want to see how many times Al Jefferson (http://espn.go.com/nba/player/_/id/2389/al-jefferson) touched the ball on the left block in the first half? The system could send information to an iPad that showed the location of every one of his possessions and allowed coaches to make adjustments on the fly.
"What this can measure is both shooting efficiency and shooting frequency," Hellmuth said. "When a defender's in the paint or in the area, players can always choose not to shoot when he's in the area. And then also how much does that defender reduce the shooting average of the players he's defending. These are some of the secrets this unlocks."
The algorithms that are used to interpret the data are constantly being refined, Kopp said, and now can identify certain plays -- like a pick-and-roll -- and defensive rotations. Now that they are poised to have even more data, the strength and breadth of the information the system can provide should only improve, he said.
"I really do think we've just scratched the surface on how we can use this," Kopp said. "I think the next few years will be fun as we have more data to work with.

Link to the article is here (http://espn.go.com/nba/story/_/id/9639224/nba-install-motion-tracking-cameras-every-arena).

[loran16]

So the NBA is going to install motion tracking cameras at all of its arenas starting this coming year, with the intent to be able to provide extremely detailed statistical analysis of all aspects of the game of basketball. Including refereeing. For example, from the article:

Want to see how successful Ricky Rubio (http://espn.go.com/nba/player/_/id/4011/ricky-rubio) was at guarding Russell Westbrook (http://espn.go.com/nba/player/_/id/3468/russell-westbrook)? The system could break down the shooting percentages and results of each head-to-head possession.
Want to get an idea how close to 100 percent Kevin Love (http://espn.go.com/nba/player/_/id/3449/kevin-love) was in his first few games back from a broken hand? The system could send information to the team trainers and doctors about his endurance and how quickly he is tiring during a game, thereby painting the most accurate picture possible of his recovery.
Want to see how many times Al Jefferson (http://espn.go.com/nba/player/_/id/2389/al-jefferson) touched the ball on the left block in the first half? The system could send information to an iPad that showed the location of every one of his possessions and allowed coaches to make adjustments on the fly.
"What this can measure is both shooting efficiency and shooting frequency," Hellmuth said. "When a defender's in the paint or in the area, players can always choose not to shoot when he's in the area. And then also how much does that defender reduce the shooting average of the players he's defending. These are some of the secrets this unlocks."
The algorithms that are used to interpret the data are constantly being refined, Kopp said, and now can identify certain plays -- like a pick-and-roll -- and defensive rotations. Now that they are poised to have even more data, the strength and breadth of the information the system can provide should only improve, he said.
"I really do think we've just scratched the surface on how we can use this," Kopp said. "I think the next few years will be fun as we have more data to work with.

Link to the article is here (http://espn.go.com/nba/story/_/id/9639224/nba-install-motion-tracking-cameras-every-arena).




Of note: it remains to be seen how much of this is actually a boon for "stat heads". While NBA teams will certainly use the data, it's unclear how much of it will ever be available to the public for use. Baseball for example, has a similar system called FieldFX, which tracks every fielder on the field at once. However, that system's data, along with a system set up to track how hard and what direction the ball goes off the bat (HitFX), is unavailable for the public to use, resulting in fielding metrics in baseball having to rely on much crappier data.

The BBall guys are saying nice things so far....but so did the baseball guys behind FieldFX, and there's no incentive for NBA teams to share the data publicly.

[toooskies]

The BBall guys are saying nice things so far....but so did the baseball guys behind FieldFX, and there's no incentive for NBA teams to share the data publicly.

The NBA is much more fan-friendly than MLB. They've already released the most advanced stat platform of the major sports' websites, and Kirk Goldsberry has already presented plenty of articles based on the partial data sets already. Among reasons to release at least parts of this info include:
- the league is better off having an even playing field in terms of stats, and more public knowledge means more knowledge parity.
- fans want it, and fan service is a good thing when it is cheap or free.
- more qualified applicants for front-office staff jobs.

[jipops]

Interesting in the video that the stats/data guys contend that NBA teams should be shooting more 3's, "way more". Hmm, could be a very interesting retort to the accusation of 'live by the 3, die by the 3' that our Duke teams have gotten many times in the past. Obviously this stat claim is directed towards nba teams but wonder if this claim could be directed towards basketball in general?

[WillJ]

I think the trend towards more analytics by teams is great and, as a fan, more discussion of analytics by commentators and writers is wonderful - I already love some of the analysts on espn.com, such as Zach Lowe, and this new data may make that type of analysis more widely accessible. I think of it as the anti-Brent Musburger school of analysis, in that Musburger, more than anyone I can think of, liked to couch things in terms of personal history, personalities, psychology, etc.

[loran16]

The NBA is much more fan-friendly than MLB. They've already released the most advanced stat platform of the major sports' websites, and Kirk Goldsberry has already presented plenty of articles based on the partial data sets already. Among reasons to release at least parts of this info include:
- the league is better off having an even playing field in terms of stats, and more public knowledge means more knowledge parity.
- fans want it, and fan service is a good thing when it is cheap or free.
- more qualified applicants for front-office staff jobs.

A few things.
1. MLB's Sportsvision also has released partial data sets to journalists to write about. That's not exactly SHARING THE INFO PUBLICLY, it's sharing bits and pieces to write pieces that make the league look advanced, without helping the understanding of amateurs.
2. No, see the League HAS an even playing field with the data available to every team. Then each team makes their own PRIVATE proprietary (redundant word use i know) stats based on the data. We see this already in the NBA - each team has their own guys who are NOT sharing info or their own data.
3. "Fans want it" - I think you overrate this point massively. Most fans don't care about analytics, and even those who do won't stop watching the NBA without these things provided. They don't add anything that can be monetized. Again, the same is true in baseball, where the sabermetrics movement is BY FAR larger than basketball and more established. It has nothing to do with being fan-friendly, as much as lacking a profit incentive.
4. Baseball without releasing this data has tons of applicants looking for jobs, and as such the salaries for analysts is basically peanuts. Suspect the same is true of the NBA.

I hope i'm wrong. But I don't get celebrating a development until you know it's actually going to be public.

EDIT: Note the NBA putting advanced stats stuff on their website is completely different, because those #s were ALREADY AVAILABLE elsewhere on the internet and based upon public data. The NBA had a profit incentive to put them on their own site, which is to draw viewers away from other sites which might display them. The same is NOT true of private data.

[Des Esseintes]

A few things.
1. MLB's Sportsvision also has released partial data sets to journalists to write about. That's not exactly SHARING THE INFO PUBLICLY, it's sharing bits and pieces to write pieces that make the league look advanced, without helping the understanding of amateurs.
2. No, see the League HAS an even playing field with the data available to every team. Then each team makes their own PRIVATE proprietary (redundant word use i know) stats based on the data. We see this already in the NBA - each team has their own guys who are NOT sharing info or their own data.
3. "Fans want it" - I think you overrate this point massively. Most fans don't care about analytics, and even those who do won't stop watching the NBA without these things provided. They don't add anything that can be monetized. Again, the same is true in baseball, where the sabermetrics movement is BY FAR larger than basketball and more established. It has nothing to do with being fan-friendly, as much as lacking a profit incentive.
4. Baseball without releasing this data has tons of applicants looking for jobs, and as such the salaries for analysts is basically peanuts. Suspect the same is true of the NBA.

I hope i'm wrong. But I don't get celebrating a development until you know it's actually going to be public.

EDIT: Note the NBA putting advanced stats stuff on their website is completely different, because those #s were ALREADY AVAILABLE elsewhere on the internet and based upon public data. The NBA had a profit incentive to put them on their own site, which is to draw viewers away from other sites which might display them. The same is NOT true of private data.

We'll see, but as this Zach Lowe article (http://www.grantland.com/blog/the-triangle/post/_/id/73501/seven-ways-the-nbas-new-camera-system-can-change-the-future-of-basketball) suggests, advanced camera data is potentially value added to the next tv contract. Apps for the hardcore fan set are absolutely a possibility. If the league makes the data publicly available, it won't be about not losing viewers; it will be about offering a subset of fans a greater investment than before. The fact that baseball has not been forward-thinking about revenue streams in this regard is not a guarantee that the NBA will be similarly blinkered. But who knows?

I also think it's easy to understand the excitement around a development such as this even if it's not made public. The quality of information and commentary available to fans has improved dramatically over the past decade, despite public info frequently being inferior to what teams enjoy. What has been crucial has been the general atmosphere of opening new vistas for understanding the game, and the league embracing this new technology only leads us further down the road. I imagine things will continue to get better.

[Des Esseintes]

Interesting in the video that the stats/data guys contend that NBA teams should be shooting more 3's, "way more". Hmm, could be a very interesting retort to the accusation of 'live by the 3, die by the 3' that our Duke teams have gotten many times in the past. Obviously this stat claim is directed towards nba teams but wonder if this claim could be directed towards basketball in general?

Absolutely. NBA teams in the bottom half of taking threes last year would have led the league 20 years ago. Threes have been and are an ever-increasing part of the game. I'm sure a limit will be approached eventually, but the trends suggest we aren't close to it yet. In general, if you hear the phrase "live by the 3, die by the 3," you can stop listening and walk away. Your interlocutor does not know much about the sport of basketball.

[g-money]

Interesting in the video that the stats/data guys contend that NBA teams should be shooting more 3's, "way more". Hmm, could be a very interesting retort to the accusation of 'live by the 3, die by the 3' that our Duke teams have gotten many times in the past. Obviously this stat claim is directed towards nba teams but wonder if this claim could be directed towards basketball in general?


Absolutely. NBA teams in the bottom half of taking threes last year would have led the league 20 years ago. Threes have been and are an ever-increasing part of the game. I'm sure a limit will be approached eventually, but the trends suggest we aren't close to it yet. In general, if you hear the phrase "live by the 3, die by the 3," you can stop listening and walk away. Your interlocutor does not know much about the sport of basketball.

Yep. And credit to Coach K for never letting the critics of Duke's aggressive use of the three pointer get to him. He figured out the simple math that a 3-pt shot with a 40% chance of going in has a higher expected value than a 2-pt shot with 50% odds long before the stat gurus arrived on the scene.

At the NBA level, what is truly amazing to me is the ability of many players to shoot a 9" diameter ball to a distance of 23 ft, 9" +/- 4.5" (slightly over 24 ft if you include vertical offset), at a 40-50% clip - while being guarded on many of those shots.