↓ Login/Logout ↓
Schedule/Results
↓ Roster ↓
Salaries
↓ Archives ↓
↓ About ↓

Making advanced stats friendly and useful

One of the things I’m highly interested in is helping close the so-called divide between basketball fans who embrace advanced stats and those who are just fine using only their eyes thank you very much.

This week at BallinMichigan.com, my site that covers college and high school basketball in Michigan as well as keeps tabs on pro players from the state and other unique ties Michigan has to the basketball world, I interviewed three people with interesting connections to both Michigan and advanced statistical work in basketball who all offered great insights.

Kirk Goldsberry  (follow him on Twitter), a visiting scholar at the Harvard Center for Geographic Analysis and an assistant professor in the Department of Geography at Michigan State University, started Court Vision Analytics, a project that combines Goldsberry’s academic and professional background in mapping and imaging with his love of basketball to create easy to understand infographics that tell complex statistical stories. Goldsberry presented his work at the MIT Sloan Sports Analytics Conference, he’s contributed to the New York Times NBA coverage and his work has been featured on many different NBA sites.

I asked Goldsberry about how in-depth infographics and mapping can help explain advanced stats to people who may not be interested in learning complex formulas behind the stats:

I know it can (help). We’ve seen it in other domains. For instance in chemistry, the Periodic Table has made a huge set of chemical elements understandable in new ways. The power of graphics to simplify or translate statistical information into knowledge is one of the huge pillars of the project. My project is not unique in that sense, but it is unique in the context of basketball. I think that’s what it has the potential to do. Every one of those charts you look at is thousands of numbers encoded visually as opposed to encoded in a spreadsheet.

These spacial structures are immediately understandable to the human eye. You can take advantage of the most powerful sense we have as human beings, which is vision. I can show that chart to a basketball coach in four seconds and circle key areas with a Sharpie and walk away, and that coach has just understood the product of a really sophisticated statistical analysis in a few seconds without ever seeing brief notation, without ever seeing a decimal, without ever seeing some obtuse numerical jargon that, let’s face it, most basketball people and most human beings don’t communicate in that statistical language.

Ben Gulker, a veteran of Detroit Bad Boys and a Wages of Wins proponent, is well known for his stats-based writing among fans who read DBB, PistonPowered or talk about the Pistons on Twitter. A lesser known fact: Gulker is also a former college basketball player. From my conversation with him:

One of the things I encourage skeptics of statistical analysis of sports to consider is the relationship between music and mathematics. Obviously, I’m going to oversimplify for the sake of brevity, but an art like music can be expressed in very meaningful ways by numbers. There is an incredibly complex relationship between the two, and learning about that relationship can enhance one’s appreciation for both.

Analogously, basketball is in many ways an art, and I’d venture to say that all basketball fans appreciate the sport for that reason. For example, watching the USA Olympic Basketball teams decimate their competition in style truly is a thing of beauty. When LeBron James is running the break, and executes a perfect no-look alley oop pass to Durant, you can’t help but ooh and ahh.

But at the end of the day, success and failure are expressed by the numbers on the scoreboard, not by how impressive your crossover or step-back jumpshot looks.

My third interview was with Marc Andreas, a high school coach in west Michigan who recently received funding for his company Net Value through Start Garden, a Grand Rapids-based venture capitol fund founded by Rick DeVos that awards seed money to entrepreneurs who make proposals for business startups. Net Value as a company is looking for new ways to measure individual defensive performance statistically. Andreas is doing some preliminary statistical work for the Orlando Magic and Cornerstone University. From our conversation:

In the sports world, I think it’s more rare to find the ‘numbers people.’ Most coaches aren’t numbers experts. I have found great reception from the coaches I’ve worked with so far. Cornerstone’s new women’s coach is Katie Feenstra-Mattera, she just retired from the WNBA. She’s not a numbers person or doesn’t have a statistics background, but the reason she got behind the concept and asked me to help the team is these statistics from Net Value shine a light on what goes on on the defensive end of the court. When you compare traditional statistics, it’s mostly geared toward the offensive end of the court. Our advanced stats don’t measure the offense. That’s already tracked by other stats. We integrate our statistics to show the other half of basketball. Coaches want to win and understand you have to play good defense to win.

This helps them know what players to scout, what players to give more or less playing time to, which players to scout or trade for. Net Value just illuminates this other half of the basketball court, and when you explain it to them like that, it makes a lot of sense to them.

I know a lot of PistonPowered readers are interested in advanced stats, including some of the work Dan Feldman does on this very site. If you’re interested in reading more, check out the full interviews with Goldsberry, Gulker and Andreas.

4 Comments

  • Aug 24, 20123:22 pm
    by brgulker

    Reply

    Thanks for your good work on this, PH. It proved to be some really good reading… well, at least what the other guys had to say!

  • Aug 24, 20123:46 pm
    by gordbrown

    Reply

    I personally don’t understand why anyone would argue for advanced stats over the “eye test.” The answer is always to incorporate both. What advanced stats are really good at are finding players who are over-rated. If you just look at points scored, you can say “wow this guy is a great player.” If however he is cranking up shot after shot and wasting possessions (which are becoming more and more precious) then the eye test has failed. Having said that basketball is a team game and sometimes a players stats are affected by that. A point guard on a team where nobody else can score (or even catch the ball) is going to have bad stats. Also I detected in five minutes of watching Maxiel live and in person that he cannot move laterally at all and will never be able to defend players on the wing (something that seemed to elude John Kuester but there you are).  But the real problem with advanced stats is that line you see on every mutual fund prospectus ever printed “past performance does not necessarily translate to future gains.” Hence the interminable debates over very young players thrown in to sink or swim as best they can.

  • Aug 26, 20121:59 pm
    by Max

    Reply

    The music analogy to math reminds me of the one frequently made with basketball which is jazz.   This seems useful to me because stats, advanced or otherwise, completely fail to take in the factors which make basketball comparable to jazz and they are: flow, synergy, creativity and spontaneity.  These are relevant markers which distinguish, for me at least, who some of the best players actually are because stats are also not very useful at projecting whether a change in role, team, teammates or coach would significantly alter a given player’s stats. 
    I’ve said it before but stats fail to take in context.   Cedric Ceballos scored 20 points a game in his prime but he did it as a garbage man and had hardly any plays run for him and yet what distinguishes his stat line from a go to scorer who takes the same number of shots and shoots the same percentage?   Ceballos’ stats could theoretically be identical to an incredibly different player.

    • Aug 26, 20127:27 pm
      by Max

      Reply

      And I should have thrown tempo and harmony into the jazz analogy to boot.  Harmony, tempo, synergy, pitch, flow, spontaneity and creativity actually are some of the factors that really distinguished players like Magic, Bird, Cousy, Monroe, Ginobili and Billups.   
      The stats do agree that these players were all great but they fail to separate them from the more mechanical players and players who just were always “the man” on bad teams and not on great teams that were winning a lot of games.  
      Stats are fine and all.  I just don’t like some of the retroactive reassessments of history and use of them to justify absurd comparisons like Kobe Bryant versus Kevin Martin in the clutch,   I’ve been hearing guys like Henry Abbott and the like talk for forty minutes at a time on these subjects and it’s been driving me crazy. 

  • Leave a Reply

    Your Ad Here