Stats Primer

Based on the work of John Hollinger (ESPN.com), Dean Oliver (Sonics consultant), and other smarter men than myself, I’m seeking to provide the best objective basketball analysis you can find on any college basketball team, regardless of the level of play. I admit that I’m a stat geek through and through, so you might not find this interesting at all if you don’t have a genuine interest in basketball analysis. However, if you are like me, you might love it.  I hope you do.

Possession-Based Statistics

Team stats, such as points scored and allowed, are helpful in determining a team’s strength, but most of the stats you find in mainstream basketball analysis overlook the simple concept of the possession.  A basketball possession can result in one of three outcomes, a basket made (a field goal or free throw), a turnover, or a defensive rebound.

Teams essentially have an equal number of possessions in a given game, so the main idea in basketball is to be as efficient as possible with these possessions. Possessions are easily estimated from a box score based on the above outcomes, and the number of possessions a team generally gets in a game is referred to as the team’s “pace.”

If a team plays faster, it has more possessions, which usually means more points and more of the other counting stats. As a result, it is necessary to adjust for a team’s pace when figuring offensive and defensive stats. Once you adjust for that, you can figure out a team’s offensive and defensive efficiency ratings, which amount to points per possession.

Four Factors/Four Keys/The Keys

Branching out from this concept, Dean Oliver developed four keys (also called “factors”) to winning a basketball game: shooting for a high percentage, not turning the ball over, grabbing offensive rebounds, and getting to the free throw line (and making the shots). A team that manages to do these four things better than its opponent is almost certainly going to win the game. While I have revised the way that I look at them (see my original study and the revised one), I track these aspects of the game, and I refer to them when analyzing a single-game team performance.

Rate Stats

The concepts of pace and possessions affect all aspects of the game. Accounting for these can bring new insights into the value of an assist, a rebound, a block, or a steal. Thus, stats like the assist ratio, the turnover ratio, and the rebound rate are born. These provide a better idea of a player’s pure passing, ballhandling, and rebounding skill than do the traditional numbers. I won’t get into the formulas here, since that’s how Hollinger makes his money. By the way, his book is an essential read if you’re into this kind of thing.

True shooting percentage is another great individual or team stat. It basically takes the regular field goal percentage and accounts for three pointers and free throws as sort of an all-in-one shooting stat. If you’ve ever had trouble discerning who is a really good shooter by looking at all of his percentages, look no further.  A similar measure you might read about elsewhere is called “effective” field goal percentage or eFG%.

Pure Point Rating and Brick Index are the last two “simple” stats Hollinger uses, and they are fairly self-explanatory. PPR uses assists and turnovers in a way that more accurately approximates a point guard’s true ability as a distributor than does the regular assist-to-turnover ratio. NBA leaders in PPR include players like Steve Nash. BI uses pace factors and true shooting percentages to measure how much a poor shooter is hurting his team by shooting too much.  Chris Webber has been a perennial NBA leader in this category.

The last of John Hollinger’s stats that I use is his all-encompassing Player Efficiency Rating. PER is Hollinger’s attempt to accurately account for all the measurable events linked to a player. Offensive ability carries a much heavier weight than defensive ability because there are far more helpful offensive stats than defensive stats. PER carries a weight for each of these events and is set to a league average of 15.00. NBA player PERs in 2004-05 ranged from Theron Smith at 5.10 to Kevin Garnett at 28.35. Great on-ball defenders are still undervalued in this formula, but it does a better job of approximating a player’s overall contribution than any other stat. I refer to it often.

Along the lines of PER is the Game Score, which is a back-of-the-envelope style calculation that oversimplifies the formula to make it relevant on a single-game basis. You’ll probably catch me referring to it a lot in single-game contexts.

Dean Oliver and Efficiency

Dean Oliver’s stats take on a slightly more complicated form. While nothing is normalized to the league like in Hollinger’s PER, Oliver’s efficiency stats are a more precise calculation of a player’s value, particularly on offense. The offensive rating I track for each player is a measure of how many points the player produces per 100 possessions, much like the team rating. There is also a corresponding defensive rating, but since individual defensive stats are limited to blocks, steals, and defensive rebounds, these ratings are not considered as accurate as their offensive counterpart.

Other Oliver stats that I track include some of the numbers used to get offensive rating, including floor percentage (floor%), which simply divides a player’s estimated number of scoring possessions by his total possessions. Offensive rating takes more of a player’s value into account, but floor% is easier to calculate. Stop percentage is the defensive counterpart to floor percentage and is estimated based on the aforementioned three individual defensive stats.

That should take care of most of the technical stuff. You can always comment or e-mail me with questions.