Tempo-free analysis isn't really catching on with traditional media types.
But college basketball's so-called sabermetrics are catching on in other places. It's no longer just Brad Stevens and "Internet People" that rely on the enhanced metrics provided by tempo-free analysis, but increasingly large audiences are willing to pay to access such information on as they do on Ken Pomeroy's website and for the annual Basketball Prospectus books. There's also the success of Luke Winn's blog, John Ezekowitz and the occasional CBS foray into the tempo-free realm.
Still, for bloggers like myself that are wholly convinced that tempo-free stats are in every way more equitable, more useful and more meaningful for comparing players and teams on a basis that is far more exhaustive (I don't really have space to explain why, but you can read more about that here), it's altogether frustrating to see on-screen graphics, print media or school athletics releases that herald a player's points per game average, or a team's rebounding margin or, worse, a player's plus/minus calculations. These numbers mean nothing to enlightened bloggers and basketball fans like us.
I think I know what the problem is, however: Basketball doesn't like change. PPG, RPG, APG, SPG, TPG, +/- and the like have been used for generations without any problems, so why change things now? Plus, those numbers are easy to explain; just saying "points per game" exactly tells you what the stat measures. Turnover percentage, on the other hand, requires some explaining. But there's a better way, people, and I think easing the viewers and readers into the new generation of tempo-free analysis is the only way to bring change.
To that end I have created an easily digestable metric that does a lot but is easy to explain: Points per player possession. This stat essentially measures how many points a player scores when he executes the end of a possession, and is related to team points per possession calculations.
It's entirely useful and equitable across the entire sport and it incorporates a lot of tempo-free stats all in one number, because a player's shooting ability, his scoring ability and his turnover rate all are incorporated. It's a metric that fans and writers can both rally behind, because it's tangible. People like to quantify scoring ability, and up to now there's been no easy way to do that for individuals in the tempo-free realm. In that way pppp is better than offensive rating (although they are certainly different measures of a player), which seems mysterious and out of sight for most people.
Points per player possession isn't a total evaluation of a player, rather it shows how big a scoring threat a player is when he has the ball in his hands. It's one part of an overall tempo-free breakdown that I think has been the missing link.
The calculation of pppp essentially is performed in the same way that team points per possession calculations are made. Simply divide the number of points scored by the number of possessions.
Points Per Player Possession = Points Scored / Play-Ending Possessions
The problem with doing this on a player-by-player basis, however, comes when you try to count player possessions. Nobody keeps track of that stuff. But I've come up with a method for coming close without actually going back and sifting through the play-by-play charts.
In order to find, say, the number of possessions that ended with Doug McDermott's hands last season, I first added up all the possessions that Creighton played as a team in the 2011-2012 season. Those numbers are available on Pomeroy's website (73+70+71...= 2,331). I then apply McDermott's individual minute percentage and possession percentage to that number (both of those numbers are also available on Pomeroy) by multiplying. So, 2,331 team possessions * 0.792 * 0.287 = ~530 player possessions for the season. That's not a perfect number but it's as close as we can easily get. It's a method of filtering out possessions that the player in question wasn't on the floor for in addition to possessions that they weren't involved in.
So, to extrapolate, the overall equation for pppp looks something like this:
Points Per Player Possession = Ind. Points Scored / (Team Possessions * Ind. Min% * Ind. Poss%)
In the case of McDermott, here's how it works out:
801 points / (2,331 team posessions * 0.792 * 0.287) = 1.512 points per player possession
McDermott scored 1.512 points on average during last season when Creighton's offensive plays ended with the ball in his hands. That's pretty phenomenal.
I applied this method to the 200 Division I players that scored the most overall points in the 2011-2012 season, and I'm pleased with the results. To make the cut a player had to score more than 480 points during the entire last season; 155 schools are represented.
Here are the top 25 players by PPPP with PPG and Offensive Rating thrown in for purposes of comparison:
|3||Nick Barbour||High Point||Sr||1.463||20.42||121.5|
|5||Langston Galloway||St. Joseph's||So||1.426||15.50||124.9|
|6||Kevin Jones||West Virginia||Sr||1.409||19.91||124.3|
|t11||Dominique Morrison||Oral Roberts||Sr||1.385||19.82||122.2|
|14||Noah Hartsock||Brigham Young||Sr||1.378||16.82||119.5|
|15||Andrew Nicholson||St. Bonaventure||Sr||1.377||18.47||116.7|
|16||Shane Gibson||Sacred Heart||Jr||1.369||22.03||112.7|
|t20||Deshaun Thomas||Ohio State||So||1.346||15.92||122.1|
|t20||Robert Covington||Tennessee State||Jr||1.346||17.85||117.8|
|22||Damian Lillard||Weber State||Jr||1.345||24.50||124.4|
|24||Rodney McGruder||Kansas State||Jr||1.339||15.85||118.2|
|25||Preston Medlin||Utah State||So||1.336||16.97||121.7|
The most efficient team in the nation last season as far as offenses went, Missouri (1.254 ppp), features three players in the top 25: Marcus Denmon (1.407 ppp), Ricardo Ratliffe (1.396) and Kim English (1.385). Of course, Kentucky is also well represented with Doron Lamb (1.451) and Anthony Davis (1.385) both also in the top eighth of my dataset.
John Jenkins, one of the most threatening rangy two-guards to enter the NBA draft this season, finally gets noticed in these numbers. He hits a lot of threes and doesn't turn the ball over much, earning him the top overall spot.
There's a correlation between high PPPP and high offensive rating scores (in the larger dataset) while PPG seems to have no correlation whatsoever, as predicted.
The best part of this equation is that it accounts for players that are excellent scoring threats but get few possessions and/or minutes on the floor. They are efficient scorers. These are the best players to pass to.
I posted the entire dataset on the HL Hoops website.
Jared Sullinger (1.228 pppp), Trey Zeigler (0.976) and Tim Frazier (0.964) show why this isn't a one-size-fits-all metric. Sullinger is a tremendous scoring threat, of course, but scored 92.6 percent of his points inside the arc, something pppp doesn't account for, nor should it. Frazier had the second-best assist rate in the nation last season (45.3 percent) which had an obvious negative effect on his pppp. Zeigler is a combination of both of those. Both Ziegler and Frasier were in the bottom 10 of the list of 200 while Sullinger came in at number 64.
Oh, and you're probably wondering who Nick Barbour and Preston Medlin are.
Barbour's 48.4 percent three-point average last season was the third best mark in High Point's history. He knocked down 108 from range last season. Barbour also got to the stripe a lot with a free throw rate of 45.1. That's all notable by my standards, and I'm glad it showed up in the ppp rankings. That kind of sharpshooting dangerousness needs noting.
Medlin has essentially the same story: Murderous from range, gets to the stripe a lot. Drew Cannon just also called him "an outstanding all-around scorer."