Tuesday, November 11, 2008
Usage is defined as the percentage of a team's possessions that an individual player uses. Basically, it approximates the "space" a player takes up in the team's offence.
If a team only had five players that were exact clones of each other, each would have the exact same usage rate: 20 percent. The stat is estimated by:
usage = 40 x (field goal attempts + (0.44 x free throw attempts) + (0.33 x assists) + turnovers) x (team pace / league pace) / minutes played
Unfortunately, calculating usage requires two other stats - team pace and league pace. The problem with calculating college usage is determining what "league" pace is. I've decided that when I use usage in evaluating college players, I'll ignore the "team pace / league place" part - in short, by assuming that team pace/league pace = 1.00. This introduces an error in a statistic which is essentially an estimation, but as long as the rules are laid out ahead of time, we'll go forward.
According to Ken Pomeroy at Basketball Prospectus, it seems to be an iron law that usage doesn't change much over time. If a player is a role player in college, she'll be a role player in the WNBA.
Furthermore, there's another rule about usage, this one formulated by Dean Oliver. Players who have high usage rates - somewhat over 25 percent - are performing under their maximum efficiency. It makes sense, as they have to carry more of a team's offense. You're not seeing the player at their best.
On the other hand, players with low usage rates - 15 percent or less - are already performing at their peak. If their usage goes above at what they're used to doing, they'll perform less well. Low usage players are performing at better than their maximum efficiency.
The end result is that players with high usage should be rewarded, and players with low usage should be penalized. I went to my Senior Prospects Metric and decided to reward any player with a usage above 25 percent, and to penalize players with usage below 15.
Players with Usage Greater than 25 Percent
Angel McCoughtry, Louisville - 32.69 percent
Robyn Fairbanks, Utah Valley State - 31.35 percent
Krystal Ellis, Marquette - 29.01 percent
Sade Logan, Robert Morris - 28.27 percent
Kendra Appling, Tennessee State - 27.16 percent
Shavonte Zellous, Pittsburgh - 26.36 percent
Megan Frazee, Liberty - 26.28 percent
Kristi Little, Duquesne - 25.86 percent
Obiageli Okafor, Tennessee State - 25.79 percent
Shantia Grace, South Florida - 25.21 percent
Erin Kerner, Quinnipiac - 25.20 percent
That doesn't mean any of the above players are going to be great WNBA players. Which players had the same unadjusted-for-pace average of the 2008 Draft Class? Valeriya Berezhynska with 30.76 percent and Jolene Anderson of Wisconsin with 30.55 percent. (Candace Parker had a respectable 28.60 percent at Tennessee.)
Adding in usage to the SPM moves Angel McCoughtry up from #8 to #5 on my list. And here's my newest Top 25: