Everything we know about basketball is a lot

April 14, 2012 8:23 AM

"Ever wondered why there are only 5 positions in basketball or how a player’s position is determined?" Maybe not. But analytics are becoming more and more important in basketball, to the point where some are questioning some fundamental 'facts' about the game. After the MIT Sloan Sports conference this year specifically addressed the role of analytics in basketball, there has been a bit of a backlash against the practice among commentators, coaches and fans. Yet the projects just keep coming, including this recently updated web project using some amazing mapping analysis: Courtvision

Is basketball headed for a moneyball moment? In some ways the 2011 Championship Mavericks prove it's already there.

More APBRmetrics resources.
Intro to APBRmetrics
"A Starting Point for Analyzing Basketball Statistics"
Play Index+
Simple Ratings from 82game.net
Basketball Prospectus
Hoop Analyst on Bad Shooting
PopcornMachine tracks player efficiency

posted by Potomac Avenue (16 comments total)

That Courtvision is just wow.
posted by Stonestock Relentless at 8:35 AM on April 14, 2012 [2 favorites]

It's a good thing I didn't have anything to do this morning, because this post is a wonderful Internet black hole. CourtVision is pretty excellent.
posted by sc114 at 8:39 AM on April 14, 2012

Being that there are five players on the court, I'd wonder a lot more if basketball had six positions.
posted by tylerkaraszewski at 8:39 AM on April 14, 2012

Excellent reading, leading into the playoffs! I can't find the quote, but after a loss in a interview a few years ago, in reply to a typical "what went wrong" question, Tony Parker replied "We did not put the ball into the basket." That sort of typifies my level of understanding of the game.
posted by Devils Rancher at 8:44 AM on April 14, 2012 [2 favorites]

I suspect that every other sports efforts to add statistical rigor would be greatly helped by completely ignoring what the baseball guys have done. Start on a blank page.

The statistical analysis has to look completely different from baseball in almost every way. Not to say all of this stuff is of questionable value - pace, eFG, and Usage all seem pretty useful, but usage is a key driver there and that's a function of who else there is on the team to take shots.

Courtvision is seriously just fancy parsing of good shooting data.
posted by JPD at 8:46 AM on April 14, 2012

Tyler: according to that video basketball should have 13 positions! Just like how soccer has 11 men on the field but many specialized positions that come and go, like striker and sweeper.
posted by Potomac Avenue at 8:46 AM on April 14, 2012

Also the fact the sloan award for is called the "alpha" award greatly amuses me and angers me.
posted by JPD at 8:54 AM on April 14, 2012

As a total stathead for baseball, I support this completely. Soon, there will be no more games, just spreadsheets. Glorious spreadsheets.
posted by Ghostride The Whip at 10:15 AM on April 14, 2012 [3 favorites]

For awhile in Seattle, George Karl seemed to have abandoned the two tweeter, two midrange, subwoofer traditional approach for a Bose-like set of five similar players, and had considerable success, but not great success.

I found it interesting and fun to watch.
posted by jamjam at 11:22 AM on April 14, 2012 [1 favorite]

One thing I wish my team's coach would do more of is go with small fast bench guys at the end of the game when the other team is probably too tired to full-out sprint much. That strategy seems to work very well when I see it tried.
posted by bukvich at 11:52 AM on April 14, 2012

You guys may know I'm an unapologetic Spurs fan, but there's more to it than stats.
posted by Devils Rancher at 12:46 PM on April 14, 2012

Actually DR, one of the stats up there says that a superstar coach can have just as much of an impact as a player like Lebron. So, yeah. Stats rule!
posted by Potomac Avenue at 1:17 PM on April 14, 2012

How many nerd sentences can we begin with "Actually" let's find out hahaha.
posted by Potomac Avenue at 1:18 PM on April 14, 2012

We feel certain that by applying science and higher mathematics to basketball, we can quickly master the game and beat Alvania!

That image is from The Hot-Shot Hoopsters (in the comic book Strange Sports Stories, #46, February-March 1963). You can find a review of the story here. It's pretty funny.
posted by twoleftfeet at 4:12 PM on April 14, 2012 [1 favorite]

I've been following this, mostly because Henry Abbott, who started Truehoop, is very, very into making the game better.

He's currently running a series called Hoop Idea that focuses on the new stats, as well as ways of rethinking the game. At the moment, he seems to be obsessed with flopping, but there's also a lot about the differences between newer, more successful GMs (Morey of the Rockets, Rich Cho, who used to be with the Blazers) and the more traditional thinking GMs are doing. Since we're near the end of the season, there's a lot about rebuilding a team, and the misconceptions involved in tanking.

It's interesting that this stuff is coming to the forefront. I wonder if, in some ways, this is going to radically alter our definition of a superstar. Would the Hawks have given Joe Johnson such a huge contract if they'd been looking at more modern stats?
posted by Ghidorah at 6:39 PM on April 14, 2012 [1 favorite]

I haven't had time to read every link above, but I'm a huge proponent of advanced stats in basketball. I think there are enough resources at hand to really turn this into a statistical revolution, and people like Mark Cuban and Daryl Morey are doing that (and keeping the results hidden from the rest of us - they have no interest in sharing their formulas).

Once you get started on the basic aggregates of stats (like PER, Win Shares, etc.), then you can get into some of the cooler stuff. And as much as I like delving into the formulas behind each of those two, the way I picked my favorite was much more basic - sort from high to low, and compare which chart had players I liked more near the top. That way I can combine my own "eye for the game" with those of stat-heads who do this for a living and that's my analysis on who is doing a better job.

My other favorite "advanced" stat is how basketball reference tabulates the percentages of a certain event happening. Rebounding percentage is really useful too - for example, NBA fans know that the Utah Jazz have four outstanding big men this year (five if you count Jeremy Evans) - so adding the percentage of rebounds grabbed when a player is on the floor is just another useful angle to consider when evaluating these players.

Anyway, thanks for the collection of links - this kind of stuff turns a lazy Sunday into a "I will create the ultimate basketball analysis formula!" Sunday...
posted by antonymous at 8:26 AM on April 15, 2012