December 5, 2007
Interpreting Individual Stats
An Introduction

by Ken Pomeroy

The saying that grates me as much as any other is that tired quote attributed to Mark Twain: “There are lies, damned lies, and statistics.” This of course implies that statistics are somewhat untruthful. But statistics are the historic record. It’s in the interpretation where deception can be introduced. With that warning serving as an introduction, it’s time to look at the first calculation of individual rate stats. Given all of the new readers around here this season, it’s also a good time to explain the value of some of these metrics while examining some of the notable players at or near the top in these categories. So call your friends, and gather ‘round--here’s an explanation of some of those terms that might have confused you up until now, complete with 100 percent of the daily recommended supply of context.

Offensive Rating and Usage (ORtg and %Poss)

I know it’s cool to have one magical rating that defines a player, but there are two stats that define a player’s offensive value, and you really can’t consider one without the other. That’s why when I create leaderboards for offensive rating, I stratify the players by usage. Players that maintain a high offensive rating while using a lot of possessions are more valuable than the efficient players that use few possessions. Through Monday’s games, Butler big man Matt Howard has been the most efficient high-usage player in America, with an offensive rating of 126.5 while using 29.7 percent of Butler’s possessions. It’s worth noting that none of my player statistics are adjusted for competition, and unlike the team stats I compute, games against non D-I competition are included. However, Butler hasn’t been feasting on weak opponents, and Howard had his best game against his team's toughest opponent, Ohio State. So understand this: Matt Howard, who’s only a freshman, is very good. Matt Howard should be starting, and assuming his coach Brad Stevens is no dummy, Howard will be starting (or at least getting starter’s minutes) soon.

Percentage of shots taken (%Shots)

This is similar to usage, but just focusing on shots. The early leader in this category is Santa Clara’s John Bryant, who by this estimate has taken 45.2 percent of Santa Clara’s shots while he’s been on the court. That’s something else to remember–-the rate stats are estimates based on box score data, and these estimates get better with more data. Bryant’s rate is unsustainable, as only two players over the past four seasons have been able to maintain a shooting rate of over 40 percent. However, I bring up Bryant because he has always been one of the statistical wonders of the college hoops world. Two seasons ago, he grabbed 15 rebounds in 16 minutes of action during his collegiate debut. He’s an excellent rebounder, shot-blocker, and scorer, at least on the West Coast Conference level. Sadly, a matchup with 7’2” center Luke Nevill earlier this season didn’t materialize when Bryant picked up two fouls in the first minute of SCU’s game against Utah.

In other news, the most frequent shooting freshman is not named Beasley, Mayo, Love, Gordon, or Rose. As I alluded to last week, it’s Michigan State guard Chris Allen, who has launched 37 percent of the Spartans’ shots while on the court.

Offensive rebounding percentage (OR%)

Offensive rebounding percentage simply measures the percentage of possible offensive rebounds a player gets. Players on poor shooting teams and teams that play at a fast pace tend to have more opportunities to grab offensive rebounds. The purpose of OR% is to determine who gets the most offensive rebounds given the same number of opportunities.

Currently sitting atop the OR% standings is Pitt freshman center DeJuan Blair. Blair inherited the position from arguably the country’s best overall rebounder over the past two seasons, Aaron Gray. The Panthers' schedule has been characteristically weak, so it’s too early to take Blair’s numbers in any category seriously. Effectively deal with Duquesne’s Shawn James on Wednesday night, and I’ll have a lot more respect for you, DeJuan.

Another notable freshman on the leaderboard is Texas A&M’s DeAndre Jordan, who is ranked 14^th with an OR% of 18.4. So far, Jordan’s offensive game has been disappointing, especially compared to the other impact freshman, but early indications are that he can keep possessions alive with his work on the boards.

Defensive rebounding percentage (DR%)

Defensive rebounding percentage is similar to OR%. Players usually have higher values of DR% since defensive rebounds are about twice as common as offensive rebounds.

Kansas State’s Michael Beasley ranks fifth in the country with a DR% of 32.3. Beasley’s case is unique because his four most frequently playing teammates basically get defensive rebounds only by accident, with each having a DR% less than 8. This advances the prospect that Beasley’s defensive prowess is aided by a lack of competition with his teammates for defensive boards. It’s a chicken/egg argument, because it’s possible that his teammates DR percentages are being suppressed by Beasley’s activity. However, the Wildcats are not a great defensive rebounding team, ranking 108^th in the country, which suggests that there are more defensive rebounds to be had. And since Beasley is beyond what is normally capable by one human being, it tells me that his teammates have some room for improvement. Beasley is a great rebounder, don’t get me wrong, but it takes a special set of circumstances to lead the nation in rebounding. In Beasley’s case, it helps to play on what has been a fast-paced team so far, and it helps to have a lot of teammates that aren’t going to take any rebounds from you.

Block percentage (%Blocks)

The tallest player in D-I happens to be the early leader in block percentage. UNC Asheville’s Kenny George has blocked 28.8 percent of his opponents’ two-point shots while he’s been on the floor. However, like most very tall men (George is listed at 7’7”), he has trouble just staying on the floor due to a variety of chronic injuries. Sadly, it would be a surprise if George is able to meet my qualifying threshold of playing at least 40 percent of UNCA’s minutes for much longer. George has recorded 39 blocks in 134 minutes this season, though half of those minutes were against sub D-I competition.

Among players you might hear about in March, I find Hasheem Thabeet’s story the most noteworthy. He has a block rate of nine percent, which ranks 64^th in the country. The problem is Thabeet is the third-tallest player in the country at 7’3” and he had the nation’s fifth-best block rate last season, at 15.3 percent. This is where some interpretation comes into play. Has Thabeet been a less effective shot blocker so far, or have his teammates been less effective in forcing opposing players to drive the baseline, providing easy pickings for Thabeet? Regardless, there has to be some concern there for UConn fans, because UConn’s perimeter-phobic offense doesn’t appear to be too much different than last season. Therefore, any slip in the defense will likely mean another rough season in the Big East.

Turnover rate (TORate)

Turnovers are a passion of mine, because while even the most simplistic analyst will acknowledge their importance to a team’s performance, they rarely are considered when determining an individual’s value. However, it’s not so simple to devise a metric that accurately measures a player’s ability to take care of the ball. One could use turnovers per 100 team possessions. But guys that rarely participate in the offense will come out looking like ball-security mavens. The metric I’ve settled on is the percentage of individual possessions that result in a turnover. This isn’t perfect either because high usage players tend to have a lower TO percentage than low usage players. So I tend to focus on the guys that have a high usage and a high turnover rate. These are the players that tend to be overrated. The most obvious player to pick on at this early stage is NC State’s Brandon Costner. Costner was one of the ten best freshmen in the nation last season. But he’s off to a miserable start this season with a 25.0 percent turnover rate (up from 20.4 percent last season) while racking up a 25.4 percent usage. Additionally, though, he’s shooting just 40.6 percent (eFG), so his turnover issues are not his most glaring problem so far. Fortunately for State, freshman J.J. Hickson has been as good as advertised (if a little turnover-prone himself), and has picked up some of the slack resulting from Costner’s slump.

---

That’s an introduction to individual stats. We’ll use these, and others, more to our advantage in the coming weeks. More so than the team stats, right now our sample size is still too limited to develop reliable conclusions in most cases. However, even with a large sample, we should always consider context, because basketball stats don’t usually speak for themselves.

Ken Pomeroy is an author of Basketball Prospectus. You can contact Ken by clicking here or click here to see Ken's other articles.