Having read through all 350+ pages of “Basketball on Paper,” I thought I should try to put a little of what I learned to use. Let’s take a look at MSU’s individual offensive rating and usage rates. Both these stats are available at kenpom.com.
Offensive rating is basically an attempt to use all the individual basketball stats currently available to measure a player’s efficiency in using possessions to score points. The rating uses the same scale as team offensive efficiency (points scored per 100 possessions).
Usage rate is determined by the number of shots taken and turnovers committed by each offensive player (i.e., the number of possessions consumed by the player).
Ken Pomeroy provides a good summary of how to look at the two numbers:
A very important aspect of offensive rating is that it must be used in conjunction with the possession usage (%Poss) column to have any value. The average player will use 20% of his team’s possessions while he is on the court. The majority of players fall between 15% and 25%. A player that has a high offensive rating and uses a lot of possessions is especially valuable (example: Adam Morrison, 122.8 ORtg, 31.4% possessions used).
While statistics can’t capture every nuance of a basketball team’s offensive performance (setting screens, passes that lead to assists, etc.), this is a pretty good way to get an overview of how a team’s players contribute to its overall offensive efficiency. Here’s a scatter plot of the offensive ratings and usage number for each MSU player receiving substantial playing time this past season:
Ideally, a player wants to be in the upper, right-hand quadrant of this scatter plot. That would indicate a player is both efficient (above the team average offensive rating, which was 110.7 for MSU last season) and contributing on a high percentage of possessions. Neitzel was the only player in that quadrant, although Morgan and Suton were both just outside it.
This chart emphasizes the perimeter-oriented nature of MSU’s offense. All the perimeter players except Walton used more than 20% of the possessions while they were on the floor. All the big men used less than 20%. (I’m counting Morgan as a perimeter player.)
Going into next season, MSU will need to replace its two most efficient players. Replacing Naymick on offense isn’t an enormous issue, as he took a limited number of shots. (Replacing him on defense will be a much larger issue.)
Replacing Neitzel will be tougher. Allen and Lucas will need to become more efficient. Allen’s 3-point shooting percentage of 36.0% this past season was decent; he’ll need to increase his 2-point shooting percentage of 41.5% and boost his assist rate of 8.9%. I think Lucas’s efficiency will naturally rise as he becomes more comfortable knowing when to create his shot and when to distribute the ball. Summers was pretty efficient this past season; he’ll need to maintain that efficiency as he’s given more minutes against tougher opponents.
Dean Oliver shows in “Basketball on Paper” that offensive efficiency tends to decline for NBA players as they’re asked to make bigger contributions to the offense. That equation is a little different in the college game, I think, as the development of players across their careers is going to be more pronounced. The development of the freshman should hopefully offset the loss of Neitzel.
A second major concern will be whether Walton and Gray can get to at least a 100+ level of offensive efficiency, given that both are likely to see significant minutes. A reduction in their respective turnover rates of 31.3% and 30.7% would at least make their lack of efficiency less costly in terms of sucking up possessions. (Note that this turnover rate is calculated based on the number of possessions a player uses; Walton’s rate is considerably higher than Lucas’s rate of 19.8%, since Lucas shoots a lot more.)
As a final note, Oliver also calculates individual defensive ratings in his book. The weakness in those ratings are that we have very little available in terms of individual defensive stats: blocks, steals, and personal fouls. Everything not covered by those three stats basically has to be divvied up equally among a team’s player, so the individual ratings tend to mirror overall team defensive efficiency. For that reason, presumably, kenpom.com doesn’t include individual defensive ratings.