Statistical Analysis. Humor. Knicks.

Monday, July 28, 2014

Open Feedback To ESPN Stats

[FOLLOWUP: Apparently the entire email was too big to send using their form, so instead I requested that ESPN read it off my site. Unfortunately I think my chances of them reading it has decreased greatly.]

ESPN.com created a feedback form for their NBA stat coverage department. So, I sent them this letter:


To Whom It May Concern:

First I would like to thank you for putting a feedback form on your web page. Since you’re asking for opinions, I’ve noticed that the baseball stat page on ESPN has a advanced statistics while basketball lags sorely behind. The only “advanced” statistic I can find is aFG% and PPS. There are some easy ones you could use that would be beneficial to all NBA fans.

You could start with possessions. A stat heavily used by every serious NBA author from Dean Oliver to John Hollinger (and I believe created by or used by Dean Smith?). Possessions are important because some teams run a fast paced offense thereby giving them (and their opponent) more chances to score. This makes points per game less important because a team might be scoring a lot of points in a game, but only because they are taking more chances (think early 90s “run at all costs” Nuggets). Possessions are easily calculated using stats you already have in your database.

Poss = FGA + TO – OReb +.4*FTA

Of course another stat that would go hand in hand with this is pPTS or points per 100 possessions. It’s simply PTS*100/Poss. Last year, according to my calculations Indiana had the 8th best offense, however due to their turtle slow pace they scored few points per game. There is no doubt that Indiana was a good offensive team last year, but you wouldn’t have thought so due if you only looked at ESPN’s PTS/G which ranked them at a lowly 21st!

Rebounding is not affected by possessions, since you already wisely use REB%. However turnovers should be measured by possessions as well: TO/Poss (turnovers per 100 possessions). The last team stat I will request is FTM/FGA. This shows a team’s ability to get to the foul line score. There is a correlation between winning and the following four factors of basketball: shooting, turnovers, offensive rebounding, and scoring from the free throw line. Shooting and rebounding are already covered on your web page (aFG% & OREB%), while the other two (TO/Poss & FTM/FGA) are simple equations that could be added to cover all 4 major aspects.

On the individual level, I would start at John Hollinger’s PER, a great measure of a player’s offensive contribution. You only need to pickup a copy of Basketball Forecast to get the equation (but being in the stat business I assume you already own a copy, right? ;-) Of course why stop there when you can also keep track of Dean Oliver’s excellent stats: Floor percentage, scoring possessions, etc.

If you’re still reading, I’ll pitch two more ideas to you. One is keeping track of charges. Some players are very good at taking charges which creates a lot of turnovers. Shouldn’t they get some credit for this? Finally why not keep track of the game using the Possession Scoring System? All you would need would be 2 or 3 interns at every home court to do this for a whole season. Actually if they had access to the games taped on video only 1 or 2 would suffice. Dean describes the notation in detail in his book “Basketball on Paper” (which you also own, right?) and can also be read online at: http://www.powerbasketball.com/theywin2.html.

Thanks for taking the time to read this & I hope to hear your response.
Mike
KnickerBlogger.NET

2 comments on “Open Feedback To ESPN Stats

  1. Pingback: property and casualty license

  2. Pingback: Autism Articles

Comments are closed.