Wednesday, October 24, 2007

Villanova Efficiency and Usage

While looking around over the weekend I checked into the Hoya Prospectus Blog and discovered he applied the analysis discussed recently by Ken Pomeroy over at The Basketball Prospectus to the Georgetown teams of the past 3 years. I decided to crunch the numbers for the Wildcats and see what I could come up with...

S Clark------21.512.2104.060.116.3106.8

Pomeroy provides individual possession-based stats going back to the 2004-05 season. Lowry was a freshman and Fraser, Foye, Ray and Sumpter were just beginning to get some (post recruiting class) national attention. I decided to add playing time (Min%) to the analysis. While Pomeroy suggests that playing time will not by and large, have an impact on ORtg and Poss% (growth/decline in Poss% is determined by the slope of the curve, not the amount of time on the court), a player receives more or less playing time based on (among a number of factors) how well he meets the staff's expectations for play at that position. The rapid growth (or decline) in playing time from one season to the next can suggest a significant change in the player/team relationship. The players above earned 10% or more of the PT at their position for the season listed. I have ordered them (roughly) in the years they played.

Pomeroy, in his latest article on the subject of player possession rates, ("Effective Usage: Putting Individual Efficiency in Context" over at Basketball Prospectus), used a loose classification system:

Superstars combine high efficiency and high usage. Outstanding role players provide high efficiency without high usage. Durant is an example of the former. He used 31.6% of Texas' possessions while he was on the floor. With five players on the floor, an average player will use 20% of his team's possessions. Hoops being a game where the team dynamic is so important to success, most players are clustered around the average usage. Roughly half of all players have a usage between 17% and 23%. Only a select few get over 25%, and only 16 players in the nation had a larger role in their team's offense than Durant did.
- Ken Pomeroy, "Effective Usage: Putting Individual Efficiency in Context", 10/10/07

Pomeroy uses 2 elements -- Poss% (percent of a team's possessions a player uses when on the floor) and ORtg (offensive efficiency) to identify the type of role any player might have with respect to his team. According to his "Individual Stats Primer" over at, looking at possession percentage only there are 4 types of players...

  • go-to guy/star - Poss% > 25.0. By his observation few exceed 30.0
  • regular/starter - Poss% is around 20.0 (give or take 2 - 3 points). Gets about 1 in 5 possessions, is sometimes called an "average" player (though obviously ORtg is not considered when using that designation).
  • role player - Poss% is around 15.0 (give or take 2 points). The player has a limited role in the offense. Sometimes a specialized player?
  • marginal player - Poss% is around 10.0 (give or take 3 or so points). Often a freshmen whose Min% >10.0, but has not quite integrated into the offense. Though not listed specifically (and rarely referred to in Pomeroy's writings) there are quite a few players tracked by Pomeroy (because they log > 10.0 of the minutes) who gain < 13.0 of the available possessions.

Pomeroy also wrote about expectations/development -- a player's Poss% (and therefore role within the offense) for the next year could be predicted to a degree of certainty, by the Poss% of the previous year. I decided to test Pomeroy's hypothesis, "Once a player demonstrates himself to be a role player, it's unlikely he'll ever be a go-to guy and, therefore, a superstar. It's not quite a law in college basketball, but players who are not very involved in the offense tend to stay that way. Any major changes in a player's usage are usually the result of filling the hole left by a departing possession eater." against the Wildcats.

A few items of interest caught my eye during review. First and foremost, Pomeroy's assertion is largely true. With a few exceptions, each player's Poss% the next year fell within the 5-95 percentile values (and most, predictably within the 25-75 percentile values). Notice Jason Fraser's Poss% decline from 2005 to 2006. The accumulation of injuries clearly began to take a toll on his minutes (they declined...) and his ability to operate within the offense. Mike Nardi's year-over-year increase (2005 - 2007) was strong and consistent. 2005 was a low point for Mike as his role vis a vis the team was being redefined to facilitate Kyle Lowry's integration into the offense. By 2007 Mike has picked up a good deal of the possessions left by Foye, Ray and Lowry. Note also the growth in Poss% for Curt Sumpter and Dante Cunningham last season. Both are actually slightly beyond the 95th percentile projected by Pomeroy. But given the loss of Ray, Foye, Lowry (and Fraser...), the growth was not entirely shocking. That Curt lost some efficiency even as his Poss% increased is very common, and practically expected when the player is moving from "regular/starter" status to "go-to guy" status (which was indeed the way most looked to Sumpter last season). The Poss% & ORtg numbers for Bump are pretty much what Pomeroy would have predicted. Bump was not a big part of the offense in his sophomore year, there should have been no expectation that he would be his junior and (or) senior years either. Other observations from the data:

1. Those who criticize Villanova's lack of front court offense have some great data here to back their argument. Notice the Poss% for Marcus Austin & Chris Charles in 2005. While neither logged large amounts of minutes, neither did they consume large percentages of possessions when they were on the court. While one can argue that the lack of possessions (and playing time...) may have been due to their low ORtg (98.0 & 95.0 respectively), check all of the front court players, for 2005 and 2006 -- Fraser, Sheridan, Cunningham and Clark. None had more than "role player" status in the offense, despite some promising ORtgs (Clark in particular). The sole exception is Curtis Sumpter. His 2005 stats were very good, but while he played the #4 in Villanov's system, he was clearly a #3 at the next level.
2. The numbers affirm how special Randy Foye was to the team. While his Min% increased, so did his Poss% and his ORtg. The last part is especially difficult. Usually as a player moves from the "regular/starter" role to the "go-to guy" role he loses scoring efficiency (his ORtg takes a hit -- see Alan Ray's ORtg in 2006 & 2007). Foye increased his ORtg by 3.5% even as he increased his Poss% by 3.4%. Extremely unusual when Poss% reach that level. It suggests he had not yet found his optimum level.
3. Dante Cunningham's potential for growth this coming season is pretty good. His 5-95 percentile range is (approximately) 14.5 - 24 (per the charts in "Effective Usage: Putting Individual Efficiency in Context"). The large growth in his ORtg (98.4 - 110.2) even as his Min% and Poss% increased bode well for his continued development. While he will not attain "go-to guy" status in the offense, he is the best candidate (among the returning front court players) to establish himself as a "regular/starter" - type player in the offense.
4. The prospects for Shane Clark are almost as promising. Shane also increased his Poss% and ORtg even as his minutes increased. While the ORtg increase was not a eye-catching as Dante's, his "normal" growth expectation would put him at about 18.0 for Poss%; the 5 - 05 percentile projection yields a range of 14 - 23. The projection may be on the low side given that Shane's ORtg was declining at the end of the season, most likely due to his injury. He lost about 6% over the course of the last month of the season.
5. Scottie's prospects are most interesting. He finished last season with a Poss% (27.7) that suggested he was a borderline "star" - type player. The expectation for this season is that his Poss% may actually decline slightly, to about 25.5-26.0. Even the 25-75 range and the 5-95 range suggest little upward or downard movement (at most a range of 22-32). Should his Poss% decline, his ORtg will most likely increase. He will not be the sole/primary focus of the opponent's defense, and that should give him very good opportunities to score (and do it efficiently).

No comments: