Stats of the Day, Dean Oliver Edition
You might remember a few posts I wrote a while back where I played around with some offbeat statistics and named them after recent Iowa players. Today's post will cover more numbers you don't see every day, but the Dean Oliver in the title isn't the former Hawkeye. Rather, he's the author of Basketball on Paper, a very impressive book about basketball performance analysis that I read during my trip to Oregon.
Oliver stressed the importance of understanding possessions in analyzing team efficiency. As you might recall, he calculated team possessions as
FGA - OReb + TO + (0.4 x FTA)
A team's offensive or defensive efficiency was simply the points it scored or allowed per possession.
*** Related Info ************************************************
The four components of the possessions formula comprise Oliver's Four Factors (click the "articles" link, it should be at the top), or the four most important factors for winning basketball games. Shooting percentage, as you might have guessed, is the most important element of winning. The team that makes more shots than its opponent is more likely to win. Taking care of the basketball by limiting turnovers is next most important. Keeping possessions alive through offensive rebounds helps teams win, as does getting to the free throw line often. For a more detailed look at this, you should really check out Oliver's Four Factors page.
*** End of Tangent *********************************************
Oliver's years of studying basketball eventually led to what I view as the most intelligent player rating system I've seen. He repeatedly mentions in his book that there is no one "holy grail of statistics," or one number that represents the total value of a single player, but he does come up with a few numbers that, when viewed together, give you a good idea of a player's style and efficiency.
The numbers from the book that I'm going to use are floor percentage and offensive rating.
Both numbers use individual possessions in their calcualtion. Individual possessions occur when a player has a scoring possession (see below), misses a field goal or free throw that the defense rebounds, or commits a turnover.
Floor percentage measures the percent of a player's or team's possessions that end in a scoring possession. A team scoring possession is a possession in which the team scores at least one point; individuals have scoring possessions when they contribute to a team scoring possession, by making a field goal, assisting on it, getting the offensive rebound that leads to it, or by making free throws.
Offensive rating is a ratio of a player's points produced to his individual possessions. Points produced is a number Oliver created that gives players credit for contributing to the offense through assists, field goals, free throws, and offensive rebounds.
I hope that's understandable, because I don't plan on writing out the formulas. They're too lengthy and complex to lay out here. I assume that if you were actually interested in them, you'd already have a copy of Oliver's book (or will soon), and if you don't care about the computations, there's obviously no point in writing them here.
I'll provide numbers for a few of the better players of varying styles (titles taken from Athlon Sports's basketball annual) to let you see what numbers are "good" for different types of players. I'm also including another Oliver idea, the percent of a team's possessions that a player uses while he is in the game, as opposed to percent of a team's total possessions. (Numbers current through Monday, 1/10/04).
|Chris Paul||Wake Forest||.550||137||22.1|
|Jarrett Jack||Georgia Tech||.510||129||21.3|
|Raymond Felton||North Carolina||.493||123||21.1|
|John Lucas||Oklahoma State||.562||146||21.8|
|Ben Jacobson||Northern Iowa||.528||131||24.4|
|Julius Hodge||North Carolina State||.597||128||30.1|
|Rashad McCants||North Carolina||.524||132||23.3|
|Ike Diogu||Arizona State||.618||137||27.8|
|Joey Graham||Oklahoma State||.558||125||28.0|
|Paul Davis||Michigan State||.576||123||25.3|
|Eric Williams||Wake Forest||.607||125||23.9|
|Sean May||North Carolina||.610||128||28.2|
|Luke Schenscher||Georgia Tech||.555||117||19.7|
Note I. One thing that's nice about these ratings is that they're rate stats, based on possessions, so they're not biased by a team's tempo or an individual's playing time.
Note II. Before you get too excited about Haluska's high offensive rating, realize that for each player there is an inverse relationship between offensive efficiency and amount of possessions used. I'll talk more about that soon. For now, just know that players become less efficient as they take on bigger roles in the offense.
Note III. As always, be aware that differing levels of competition can affect each player's performance.
That's about enough for today. I plan to talk a little more about what the numbers mean (or at least what I think they mean) later in the week. For now, I hope you can get a feel for how different players do in each category (e.g., post players tend to have high floor %, good shooters who do little else tend to have high offensive ratings, etc). If you find this type of analysis interesting and want to learn more, I again recommend Oliver's book, Basketball on Paper.
I've pored over the numbers of all the Big Ten players the past couple days, so sometime soon I'll be looking at the leaders and how Iowa fits in. If there are any individual players you would like to see ratings for (on any D-I team), just leave a comment or send me an email.
One thing that caught my attention was the statement at the end that "for each player there is an inverse relationship between offensive efficiency and amount of possessions used." That's kind of weird, because John Hollinger says almost exactly the opposite thing. From Pro Basketball Forecast 2004-05, page V:
"There's a supposition that some players' production will decrease with increased minutes, but within reason that's completely untrue. The first Prospectus emphatically proved this with research sowhing that most player's given performance improves with greater playing time."
Just something to think about I guess...
I haven't read any Hollinger yet, but let me take a stab at that anyway. If you're talking about absolute numbers, they would almost always increase with increased playing time. If given two players of equal ability, you'd expect more points (or whatever stat you're measuring) from the one with more playing time. Again, I haven't read Hollinger, so I can only assume that's what he's getting at.
What Oliver measures through his offensive rating is efficiency (a rate stat, not an absolute figure) - how well a player can produce points through his use of possessions. Here's a quote from basketball on paper:
"Generally, the only way for players to improve their [offensive] rating is to take smarter and smarter - hence, fewer and fewer - shots (and passes)."
The opposite of that (I think) is that players trying to take on more of the offense have a tendency to force bad shots or "do too much" with the ball (see Pierre Pierce).
Hope that helps answer the question.