Zebras and Cats: 2007-08 Edition

For a few weeks in February there was a strong chance that referees Jim Burr and Bob Donato, arms outstretched to signal foul and clock stoppage in the (literally) very last seconds of one possession games, would become the icons for Villanova's season. Count on a Sweet Sixteen run to soften the memory, but limping out of the MCI Center saddled with a 14-9 record, 2 of those losses coming complements of quick whistles, it seemed that garnering a fourth consecutive dance bid was slipping just beyond reach. A zebra conspiracy against all things Wildcat? Self-involved fans fixing on excuses for a young and over-hyped team's inability to perform to expectations? Fans who sensed "things were different..." with the referees were not hallucinating fanatics (those are the other teams' supporters...). But there is no statistical evidence (that I have found anyway) that suggests Zebras hate Wildcats. They just called more fouls on them last season. As the chart below suggests, the 'Cats reversed a 2 year trend that had opponents fouling at higher (per game) rates than Villanova. While the disparity was nothing on the scale of the 2004-05 season (also ironically a season that ended in a Sweet Sixteen game...), the balance tipped slightly against the 'Cats.

The very slight average per game advantage to opponents is taken on all games this past season. If the average was less than a single foul per game, the standard deviation suggested the number of fouls per game varied a bit more than expected. For opponents, the standard deviation was 3.39 on an average of 20.1 fouls, while fouls for the 'Cats, the standard deviation was 4.63 on 20.8 fouls, a bit higher than expected, and a little surprising when compared to opponents. Given changes in the starting lineup and rotation throughout the season, combined with the lack of collegiate experience the wide variation in fouls calling is understandable. Ken Pomeroy, per his Villanova Scouting Report calculated Villanova's "experience" at 1.0 years and ranked the Wildcats 324/341, a good indication of the team's relative youth (as a yardstick, consider that Pomeroy calculated the 2007 Connecticut Huskies at 0.5 years and ranked them 335/336).

Breaking the games out by location (home, away, neutral -- see graph below) reveals more detail, and some variations that I did not expect. "Home Cooking" was a myth last season. Villanova held a paltry 0.5 average foul advantage over opponents over the 15 game home slate. This is especially interesting because a team with a home court record like Villanova's last season, 12-3, ought to expect their opponent's foul total to be padded by a few extra change of possession fouls at the end of the game. The average margin of victory, 13.2, may have been large enough to discourage the practice. Despite the record and the modest advantage in average fouls the standard deviation revealed a similar gap in the range of the number of fouls called. Surprisingly, it came to nearly 2.5 fouls. The 'Cats actually did better on neutral courts, holding an advantage of nearly 1 foul (0.9) average over their opponents over the course of 9 games. The gap for the 11 away games explains the fouling disadvantage for Villanova. The Wildcats averaged nearly 3.5 more fouls per game than their opponents. Home Cooking in reverse (Road Curse?) seems appropriate. The Wildcats had a losing road record, 4-7, which may explain some of that difference, if one assumes that the team that is behind fouls 4 - 6 times in the closing minutes as a strategy to force possession changes and give themselves more scoring opportunities. The gap in standard deviations actually favored Villanova by a very modest 0.5 fouls.

Given the relative inexperience of the squad the struggles on the road could have been anticipated. The large inconsistency from game-to-game, even at home is puzzling though. Any comfort to be drawn from these numbers will have to come from the differential in the neutral court games. Villanova was able to hold their average below their opponents, despite the unfamiliar setting. That the neutral court games were played near the beginning of the season (Old Spice Classic) and at the very end of the season (GET and NCAA) should be a good sign for the squad next season.

Methods & terms...
I counted a game as "home" if it was played at the Pavilion or the Wachovia Center, as "away" if it was played on a home court of the opponent and as "neutral" if the game was played at a location not the home court of either team. This season neutral games were played at the Milk House for the Old Spice Classic (though one could argue that the proximity of UCF made the Milk House a virtual home court for the Golden Knights...), Madison Square Garden for the Big East Tournament (though I counted it as "away" for the St. John's game) and the arenas in Tampa and Detroit where the 'Cats played 3 NCAA games.

Thoughts on Villanova's 2009 Big East Match-ups

I ran across a simple comparison method used by the bloggers over at Cracked Sidewalks to quickly analyze Marquette's Big East slate for next season (thereby declaring the Warriors had a favorable schedule...). The Big East Report Blog used a similar method which created a conference-wide average for the schedules, and individual "scores" for each team...
1. Divide the Big East into quartiles (1^st -- top, 2^nd -- with 1^st that makes "upper division", 3^rd -- 9 through 12 finishers and 4^th -- 13 through 16 finishers).
2. The quartiles (with your own team in place) will give a hint as to how to triage the conference. There are those teams you won't beat, either at home or on the road (top level); those teams you won't lose to (bottom level -- ha ha, we saw how that worked last season...); those teams which you can most likely beat if you played them at home.
3. Factor the mirror (home-n-away) series games/opponents. The conference generally tries to allocate a top, middle and bottom level opponent to each team. While considering historical rivalries, television revenues and strength of schedule for both top and bottom level teams. In past off seasons the conference coaches were asked about where they thought their teams would place in the conference and where other teams might place.
While a number of very early preseason conference rankings have been published at virtually all of the electronic media sites covering D1 basketball, the consensus appears to be that few (any?) teams will move up or down even a quartile over where they finished, going into next season.

Based on last season's regular season finish, the quartile breakdown looks like this (alpha order within quartile):
First Quartile -- Connecticut, Georgetown, Louisville, Notre Dame.
Second Quartile -- Marquette, Pittsburgh, Villanova, West Virginia.
Third Quartile -- Cincinnati, Providence, Seton Hall, Syracuse.
Fourth Quartile -- DePaul, Rutgers, St. John's, South Florida.

Of course an injury, late hire/fire staff shakeup and academics, etc. can significantly change any team's roster or staff, and that in turn can affect how they perform next season. The latest round (post NBA draft) of preseason &quor;rankings" project between 1 - 3 Big East teams in the Top 10, 4 - 6 in the Top 20 and something like 7 - 9 in the Top 25 to Top 50. It is always the same 7 - 9 teams, and individually each appears to be a contender for a spot in the NCAA field next March.

Villanova's match-ups are:
Home -- Cincinnati, Georgetown, Louisville, Pittsburgh, Rutgers and St. John's.
Away -- Connecticut, DePaul, Notre Dame, Seton Hall, South Florida and West Virginia.
Mirror Series -- Marquette, Providence and Syracuse.

The Triage Philosophy posits that a team wants the games it is virtually certain to win, and those it is certain to lose to be played on the road, thereby applying home court advantage to those games where it may tip the balance.

Applying that theory to the home match-ups suggests the home court advantage may be best applied to 4 of the 6 opponents. Villanova has had home success recently against Georgetown, Louisville and Pittsburgh, while losing on the road to each of them last season. The Georgetown game was evenly matched until (literally) the last second of regulation. Villanova split last season with Pittsburgh (taking the home game, losing on the road) and has held serve against some challenging Cardinal teams over the past 4 seasons. Rutgers and St. John's appear to be misplaced. I suspect most Nova fans would feel confident (no doubt this will bring a snicker or two from Scarlet Knight fans) about beating those teams on the road. Cincinnati beat Villanova in Fifth Third last season, in one of the most physical games the Wildcats played. With another brutally large front court contingent ready to play this season, the Bearcats may be a good team for the Pavilion. So it appears 4 of the 6 games are "advantageously placed".

If certain/virtually certain wins and losses belong on the road, then the news is good, as it appears that 4 of the 6 match-ups fall into that category. Villanova, despite a good deal of fan confidence, has won only 1 of the last 3 Joyce Center meetings (that one on a last possession tip in by Kyle Lowry, during Notre Dame's "unlucky season"). Notre Dame beat Villanova in Philadelphia last season. The Irish lost Rob Kurz from a team that finished 4^th in the conference and won 14 conference games. On paper at this point in the preseason, that looks like a loss to me no matter where it is played. Wins appear to be in the offering at South Florida and Seton Hall, again irrespective of location. That leaves Connecticut, DePaul and West Virginia to resolve. UConn appears to be a consensus #1/#2 finisher in the conference, which I weigh against Villanova's 3-1 record versus the Huskies in the past 3 seasons. While definitely blessed with a very athletic front court, they did not quite know how to use that advantage effectively against the Wildcats. I don't know if the UConn squad has progressed enough to take a game in Philadelphia, though I suspect if the preseason evaluations are true, they could take a Storrs-based (or possibly even a Hartford) game versus the Wildcats. DePaul has struggled with conference opponents in 3 of the 4 seasons they have been members of the Big East. But Villanova has lost twice in the Rosemont Arena over the past 2 seasons. They should win this season, but then I said that last season too. Villanova beat West Virginia in the Pavilion late last season. But historically they have struggled in the Coliseum, even against some mediocre Catlett and Beilein teams. Huggins will field a pretty solid team next season. I'd rather see them in Philadelphia.

Villanova's mirror series will be with Marquette, Providence and Syracuse. Marquette finished in the second quartile last season (5^th seed going into the Big East Tournament), while Syracuse finished tied with Villanova and took 9^th seed in the BET. Providence was the last team "invited" to New York City, finishing with a 12^th seed. The schedule-maker seems to have put aside the top-middle-bottom approach in favor of 3 possibly very even matchups. Another oddity (and I am not sure, should it truly be an advantage...) is that the head coaches at Marquette and Providence are new. Buzz Williams over at Marquette has had, as an Assistant Coach, a single run through the Big East. Let's hope he was taking notes. Williams also has a single season of experience as a head coach (at New Orleans in CUSA). Though on the job less than 2 monthes he has completed his staff and signed several recruits for the class of 2009. MU will bring a senior-led squad to play next season. Taken together, Marquette may not step back far (if at all? -- a number of major web sites list Marquette above Villanova in preseason rankings, often placing Marquett in the Top 15) with the changing of the guard. Providence also brings a senior-heavy squad under the direction of Keno Davis. Davis has a single season of head coaching experience at Drake, and appears to have a position or two still open at Providence. Recruiting has also lagged a bit as Davis has adjusted to the conference and change of scene. The transition may not go as smoothly for the Friars. And the third oddity of the mirror series is Syracuse. The Orange and Wildcats have been "mirror-matched" every season since the Big East expanded. If someone paired me 3 times in a row with someone I would have to think they are match-making. I wonder if there is not some thinking along those lines going on in the conference offices (and maybe at Syracuse and Villanova too?). The record for those three series is 4 - 2 Wildcats. Villanova swept Syracuse in 2005-06, holding serve in 2006-07 and having a somewhat crazy reversal last season. The Wildcats won in the Dome, but dropped the home game. Gaining a sweep with any of those 3 will be very important, though not, on paper, likely if the opponent were truly drawn from the group that will pretty clearly finish in the same quartile (or close to it) as Villanova. Any sweep would be important not only to establish more conference wins and enhance the team's place in the standings, but a sweeep against an opponent most likely to finish in the same quartile would also gain a clear tie-breaker advantage over that team...and work to push the team (downward) out of the quartile.

On the whole I believe these match-ups will work to Villanova's advantage come Big East play.

Villanova Efficiency Usage -- Thoughts on 2008-09

Looking at the Villanova roster for next season and reading through Pomeroy's "Effective Usage" article naturally brought me to the queston of which players will assume what roles next season. The roster and circumstances offer a rare opportunity, because while the slate of opponents will change (a good deal less than one might imagine, given the 18 game Big East schedule on top of the 4 game round robin format for the City Series...that's 22 games out of 29 that don't change from seaon to season), the team (Malcolm Grant excepted and Acts of God, unknown at this point) will return except where noted, intact. Maurice Sutton will most likely not (exigent circumstances excepted) have much (if any) opportunity to demonstrate he is indeed the Nova Nation's much pined for "man in the middle". The Wildcats will most likely play the hand they were dealt -- from last season.

I have reproduced the rotation (players who logged 10% or more of the available minutes at their position) from last season, along with each player's playing time as a percentage of the available minutes at their position (Min%), the percentage of possessions they had (Poss%) and Offensive Rating (ORtg). The last set of columns ("Avg.", "25-75", "5-95", "Diff?" all ordered under the heading "2008-09 Projected") is drawn from the charts Pomeroy reproduced for his article, "Effective Usage" over at The Basketball Prospectus last October:

	2007-08			|----2008-09 Projected----|
Name	Min%	Poss%	ORtg	Avg.	25-75	5-95	Diff?
Dante Cunningham	75.0	19.0	104.1	19.5	17-21	15-24	No
Shane Clark	50.9	16.8	111.1	17.5	15-18	13.5-22	No
Reggie Redding	51.6	14.2	99.7	15.0	13-16	11-19	Yes (+)
Casiem Drummond	19.9	21.4	104.4	21.5	19-24	16-26.5	Yes (+)
Dwayne Anderson	45.1	15.3	114.9	16.0	15-18	13-22	Yes (-)
Scottie Reynolds	81.4	25.1	105.6	24.5	23-26.5	19.5-30	Yes (-)
Corey Fisher	53.1	25.5	94.1	24.5	23-26.5	19.5-30	NA
Malcolm Grant	26.2	21.6	113.2	NA	NA	NA	NA
Antonio Pena	48.4	21.5	95.4	21.5	19-24	16-26.5	NA
Corey Stokes	45.9	17.6	101.6	18.5	14.5-20.5	14-24	NA

Pomeroy correlated the percentage of possessions a player used in year N by the percentage of possessions he used in year N-1. Determining an "average" percentage of possessions, he was then able to develop the range of possessions (to 50% certainty 25-75, and 90% certainty 5-95) for the same class of players. Assuming his calculations hold for this next season (they seemed to hold for last season, exceptions noted below), then we can anticipate the approximate percentage of possessions for the rotation next season. For Dante Cunningham, who used 19% of the possessions (earning him a "regular/starter" role on last season's Wildcat squad...), we find by looking under the heading "2008-09 Projected" that players using a similar percentage of possessions on average used 19.5% of the possessions the next season. To a 50% certainty they consumed in the range of 17% to 21% of the team's possessions (a marginal starter to solid starter role) and to a 90% certainty they used in the range of 15% to 24% (role player role to borderline go-to guy role) of their team's possessions (when they were in the game...). Pomeroy also detmined possessions for year N+1, based on year N as well. My last column indicates whether the 1 and 2 year charts agree with respect to the particular player (juniors and seniors only), and if not, I indicate whether the average from 3 year chart showed an average higher or lower than the 2 year chart.

Growth in the percentage of possessions (and Shot%, hence the enlarged role within the offense...) is largely due to minutes that become available when players leave (graduate, exhaust eligibility, become injured and move on to the NBA). Few opportunities spontaneously open because someone suddenly got very good in practice. As Pomeroy put it:

...Players do jump from being decoys to go-to guys in one season, and some even regress...Those are the exceptions. By and large, a player's role on his team in one season is a good indicator of his role the following season. It's largely on that principle that the team previews are based...
- Ken Pomeroy, Effective Usage, 10/10/07

The only block of "unclaimed" possessions (and minutes) come from Malcolm Grant's transfer. And while his transfer is disappointing to this Villanova fan, the actual loss quantified is something on the order of about 157 (credited) points scored (out of a total of 2,547 team points) on 137 possessions (out of 2,428 total possessions). Not a lot of opportunities to prove yourself, especially if you are trying to impress the coaching staff and earn more playing time. If the team is to improve on the 2007-08 results, the players who make the rotation will have to use their possessions more efficiently than last season. They will have to score more consistently on the possessions they get.

Rising Juniors
I believe the trendline will hold for Dante Cunningham and Shane Clark. Their Poss% may grow modestly (see the Avg. column above) and they should, as seniors, outperform their ORtgs from this past season.
While Dante Cunningham's growth in possessions from his freshman to sophomore years was just outside (high side) of the 5-95 confidence interval, his growth slowed measurably last sesson, staying in the predicted range. The loss in ORtg was modest, and frankly common when the growth in any player's Poss% is factored in. There is no reason, based on his performance last season, to believe Dante has yet one more "growth spurt" in him.
Shane Clark has garnered about 16% of the possessions in each of the past 2 seasons. As with Cunningham, the trend appears to be pretty much set. The fans have seen the same player start relatively strong in the beginning and fade as the season moves into conference play. Unlike his sophomore season, we did see a resurgence at the end of the 2007-08 season, but the mid-season malaise set the tone for the entire team's slump. A borderline breakout, 95^th percentile growth in role would put Shane at 22% for Poss%, about at Pena's level of involvement, and nowhere near the "go-to guy" role of Scottie or Corey Fisher. If he could maintain his scoring efficiency (with a Poss% of 22), that would constitute a very good season. But given the potential of the other members of the squad, I don't expect it. Last season, with the loss of Sumpter and competition from a freshman and historical bench players, Shane had an opening to make the job (and role) his, but he didn't. He may, like Mike Nardi who had similar durability problems, finish out his career with the numbers of a solid contributor, but his most valuable contribution may be the intangibles, leadership and scrappy play, that don't translate easily into numbers.
Predicting the function Dwayne Anderson will have on next season's team is, given the dramatic change in his function this season, difficult to say the least. Injuries to Clark and Drummond in the days before the start of Big East play took the edge off the team's performance. Early Big East losses added to the (then) growing uncertainty about the team's prospects for the regular and post seasons. The staff turned to the leader of the scout team to inject some energy (and optimism?) into the starting line up and shore up the team's confidence. And Anderson stepped up with a series of steady performances that featured not only timely buckets, but (much needed) rebounding on both ends of the floor and a body-on-the-floor example for the freshmen to rally to. Paradoxically, while his playing time increased (nearly doubled actually) over 2006-07, his Poss% rate declined by just over 5%, even as his efficiency (ORtg) increased. The paradox is resolved by understanding that as a player becomes more selective in his scoring opportunities (and less conspicuous in the consumption of possessions) he greater success at scoring (and thereby more efficient)..The increase in playing time may have arisen frome exigent circumstances, but the staff must have liked what they saw. As for next season, the charts are contradictory. The one year chart indicates that on average, players who had Anderson's Poss% will decline very slightly, dipping to just under 15 (14.5 possibly?), while the two year chart indicates that players who had Anderson's sophomore possession rate (20.1), will, on average, have a Poss% of about 22. The "overlap" in the 2 charts falls in the 18-22 range, or borderline "regular" type possessions. Anderson provided solid rebounding support and a terrific on court example, I would not be surprisd to see him continue gathering 15 or so percent of the possessions ("role player" function) while playing in the 35%-40% range for that quality alone. He will remain a 4^th/5^th scoring option when he is teamed with some combination of the go-to guys (Reynolds & Fisher) and the obvious regulars (Cunningham, Stokes, Pena & Drummond).

Rising Sophomores
The prognosis for The Two Rs (Reggie Redding and Scottie Reynolds) ought to be "more of the same". When Pomeroy wrote about continuity of roles (function) he could have cited Reggie and Scottie as examples. From freshman to sophomore years both progressed almost exactly as averages of the cohort.
Scottie Reynold's Poss% declined slightly just as predicted. His minutes increased even as his possessions declined. The decline in Scottie's ORtg was unexpected and brought a down note to his season. For most players a decline in Poss% usually brings an increase (however slight) in the player's ORtg (see Dwayne Anderson below). The 2007-08 team lacked a second, very efficient "go-to guy" like Curt Sumpter, someone who could draw enough defensive attention to give Reynolds more room to operate.
Reggie Redding took 14.2% of the possessions when he played, in both his freshman and sophomore years. Projections (see the "Avg." column) suggest he will get a few more possessions next year, but I would not be at all surprised to see his Poss% stay very close to 14. When Redding is teamed with Reynolds and Fisher, the 'Cats need someone to set the screen, box out on the shot, or grab the ball...and pass it out to the open man. Reggie did improve his ORtg this season, coming very close to crossing that "one point per possession" threshold that brings credibility to a college player. Experience alone accounts for some of his progress (learning to deal with the nerves; when to shoot, when to pass...), but Reggie's ORtg climbed by 13%, suggesting the improvement may be more than just a year of experience. Another 13% increase next season and Redding will be an efficient role player. While his ORtg though improved, was still below the team ORtg, given the modest number of possessions, any impulse to assign blame for underperformance on Reggie is probably misplaced.
Casiem Drummond, like Dwayne, is an enigma. How he works into the team next season will depend not only on successful rehabilitation (and continued good health) of his foot, but also whether he comes back with the conditioning necessary to play Villanova basketball and if the staff maintains a commitment to utilize post passing and an inside offensive presence. Arriving out of shape his freshman year Cas was able to play only 10% of the available minutes while functioning as a role player within the offense. This past season, properly conditioned from an off season program, he posted break-out type numbers before injuring his foot in early December (late November?). And while he appeared in 13 games after leaving the Temple game (and started three of them), his effectiveness was clearly diminished. If healthy and effective Drum may not take a higher percentage of the possessions than he did this season (see table above), but he will get minutes because he can score consistently near the basket (or get to the line...) and also put up some very nice rebounding numbers, an area of weakness for the Wildcats last season.

Rising Freshmen
Attrition through transfer has already taken two of the five members of the class which started using it's eligibility in 2007. The three remaining were very well regarded coming in, and last season provided opportunities to see each at his best. Recognition and consistency are the key, but this is the group which can have the greatest positive impact on Villanova's ORtg...mostly because their individual ORtgs were low.
Many predicted Corey Fisher would start at #1 from the first game. Ironically Fisher missed the first game (versus Stony Brook), but he was the first in the group to start (versus Penn in early December). He went on to start 21 of the next 25 games before being brought off the bench for the 2 games in the BET and the 3 NCAA games. While Corey F. had some very good games (Penn -- 13 points on 4-9 shooting, DePaul - 23 points on 8-15 shooting; second Syracuse game -- 21 points on 7-15 shooting, etc.), he also had some stinkers (Georgetown -- 4 points on 1-16 shooting, though he did have 4 assists; Connecticut -- 5 points on 1-6 shooting with 4 assists, etc.). And he (along with Scottie Reynolds) showed a lack of maturity by being late to the team bus in Louisville. Coach Wright started neither, and kept Fisher on the bench for all but garbage time, as the 'Cats lost by 14. Fisher, a natural leader and creative passer took too many shots last season and hit far too few of them. Despite some decent assist numbers Fisher logged a 94.1 ORtg while consuming just about 1 in 4 possessions while he played. Scoring about 3.75 points for every 4 possessions while the team as a whole scored 4.16 points every 4 possessions suggests Fisher was a drag offensively on the unit. That he was able to log only 53% of the minutes available in a situation that was tailor-made for his talent and skills, suggests there was a good deal that was not working for Fisher last season. Assuming Corey F. will again be paired with Reynolds next season, look for him to repeat this past season's pattern and function as one of the (2) go-to guys when the team is on offense. Whether his Poss% declines slightly (anticipated outcome -- see Avg above) or grows to 30% (within the 95% certainty range) Fisher must convert his opportunities at a more efficient rate.
Antonio Pena was drafted into the starting rotation because of injuries to Casiem Drummond. Pena performed well, but unevenly as he managed to work himself into the offense as a regular type player, a 2^nd/3^rd option on offense. Whether he starts over Drummond or comes off the bench, Antonio will (if Drum is healthy) have to split the minutes at #5 with Cas and any minutes at #4 with Cunningham (and Anderson?). Like Fisher Pena took too many possessions to score that inefficiently (ORtg was 95.4). Unlike Anderson, Cunningham and Drummond, Pena has not exhibited especially strong rebounding skills for a big man either. Unlike the Villanova teams of 2004-06, last season's version of Wildcats did not feature a strong rebounding backcourt. Don't look for it to suddenly develop next season either. That makes this an essential skill for anyone playing in Villanova's frontcourt.
Corey Stokes looked dazed and confused for a good part of the season. The staff plugged him in and let him play through some pretty terrible games (1-6 versus George Mason, 1-8 versus Penn, 3-9 versus Louisville, 2-9 versus South Florida...), Stokes logged minutes in the mid and upper 30s (percent of the available playing time at his position), rising to about 45 by season's end with Poss% numbers which indicated he was functioning as a borderline regular in the offense. His shooting efficiency and ORtg also rose through the end of the season, both positive signs for this season. If Corey S. can start strong next season (ORtg >120) he may be able to push hard on the 95% confidence limit of 24% possessions. The average growth for a player with Corey's Poss% (17.6) is 18.5. Stokes has the fundamental skills to do better.

With no impact players due on the Mainline in the fall, the burden for growing the offense falls squarely on the shoulders of the returning rotation. Reallocating the playing time will most likely happen as the season progresses, but redistribution of possessions tends to occur only when one (or more) players becomes a lot better, ie the player scores more efficiently than his teammates. Candidates for biggest improvement (most improved ORtg) include Cas Drummond, Corey Fisher and Corey Stokes. One or all of them will have to improve if the Wildcats are to finish higher in the Big East and move deeper in the NCAA tournament.

Villanova Efficiency and Usage 2007-08

When I wrote the "Villanova Efficiency and Usage" post last October I knew I would eventually revisit those ideas with this past season's data in hand. Ken Pomeroy's ideas on how to use efficiency and usage stats (from "Effective Usage" over at the Basketball Prospectus) to understand an individual team member's roles during play helped me understand why some of my player evaluations and projections from the 2006-07 season were wrong. With time in the off season to work with last season's data I was curious to see how accurate his analysis of player usage was (with respect to the Wildcats) and I have reproduced the rotation from last season in the table below, providing percentage of playing time (Min%), percentage of possessions (Pos%) and offensive rating (ORtg) for all seasons in which the squad member has played (again >10% of the minutes available at their position -- Min% >10.0). To provide some context I have also included the 2005-07 Min%/Poss%/ORtg statistics for the three seniors, Mike Nardi, Bump Sheridan and Curtis Sumpter, who graduated in the spring of 2007.

	2005-06			2006-07			2007-08
Name	Min%	Poss%	ORtg	Min%	Poss%	ORtg	Min%	Poss%	ORtg
Sumpter				73.9	27.1	113.2
Sheridan	69.8	12.8	102.7	64.4	12.2	107.9
Nardi	65.6	18.3	110.2	74.3	20.6	114.7
Cunningham	47.3	9.9	98.4	68.3	17.3	110.2	75.0	19.0	104.1
S. Clark	21.5	12.2	104.0	60.1	16.3	106.8	50.9	16.8	111.1
Redding				33.6	14.2	86.5	51.6	14.2	99.7
Drummond				10.7	15.0	99.1	19.9	21.4	104.4
Anderson				16.0	20.1	109.7	45.1	15.3	114.9
Reynolds				74.1	27.7	107.0	81.4	25.1	105.6
Fisher							53.1	25.5	94.1
Grant							26.2	21.6	113.2
Pena							48.4	21.5	95.4
Stokes							45.9	17.6	101.6

Three questions going into last season...
1. How would the team fill the roles of Curtis Sumpters & Mike Nardi? Sumpter had, by graduation, evolved from a back-to-the-basket HS center to a versatile wing. As demonstrated throughout the 2004-05 season he was an inside-out player, as adept at backing down a defender in the paint as he was stepping out and dropping a 3. The 2006-07 (rehabilited from 2 knee surgeries within a 7 month period...) version was a less mobile (and still rehabbing) outside-in version of the 2004-05 edition, a bit more inclined to take the jumper and avoid the contact of the low post. He had attained borderline star status as a go-to guy who was highly efficient (ORtg of 113.2). And a challenge to replace. Mike was a highly efficient regular type player. His role (as represented by Poss%) had grown consistently in his last 2 seasons with the team. Arriving on the Mainline as a point guard with an ok outside shot who often preferred to set up his teammates, Nardi re-engineered his game to become a "take-em" style player who could attack off the dribble or (as he was eventually used most often) a catch and shoot sniper.
2. How would the freshmen (true and redshirt) fit with the returning players? The entering players did not fit the mold of the players leaving. Pena had more range than Bump (but not the technique); Fisher a classic point guard with a decent jump shot, given to distribute and penetrate, was not a Mike Nardi; Corey Stokes had a much better jump shot, but few inside skills to match Sumpter. Malcolm Grant benefited from a year of prep school, but was in the same 6-0 to 6-2 height range as Reynolds and Fisher. These four had to fit with a collection of players who, aside from Reynolds (himself a sophomore), had more experience as role players and setting up their teammates than scoring (and leading) themselves.
3. How "up" is Scottie Reynold's upside? He came to the Mainline, an MDAA guard with the reputation as a scoring point guard, the latest in a growing line of guards. As the season progressed Scottie's role in the offense expanded, in part because he was indeed a creative scorer; in part because a series of in season injuries to Nardi and Shane Clark eliminated established scoring options. Reynolds stepped up, and with a number of outstanding individual efforts, managed to earn the Big East Newcomer of the Year award.

Most attempts to classify a player's role use the individual's playing time as the yardstick for classification. Players who start are "regulars/starters"; regular members of the rotation, but tend to see limited minutes (or are used in specific situations) are "role players"; players who start, get lots of minutes and are the primary scoring options for the team are "go-to guys/stars"; players who rarely see the court except at garbage time are "deep-bench (marginal?) players". While role can be the nexus between performance and playing time (team members who play well get more playing time, etc.), that does not (as explained by Pomeroy) necessarily describe a player's function with respect to the entire team. Shifting focus (temporarily) from playing time (or Min%) to possessions allows for a similar classification system, but one that provides a little different insight into how the players function with respect to the team. Using the possessions from the table above, the 2007-08 Wildcats could be classified this way:

Go-to Guys/Stars	Poss% >25 or so...
Scottie Reynolds	Carry on from 2006-07
Corey Fisher	Poss% tailed off at season's end
Regulars/Starters	Poss% About 20%, range 17 - 23
Malcolm Grant	Declining minutes by season's end
Antonio Pena	Freshman stepped in for injured Drum
Casiem Drummond	Greatest improvement...until injury
Dante Cunningham	Poss% improved again
Corey Stokes	Big improvement in Feb & Mar
Role Players	Poss% About 15, range 13 - 17
Dwayne Anderson	Reshaped his role, and got more PT
Reggie Redding	Another season as a complement player
Marginal Players	Poss% About 10, range <13...
None	Unusual development...for Villanova

Not having a marginal player jumped out at me right away. The Wildcats have always had a Bump Sheridan-type player -- someone who sets the screens, gets the rebounds and takes the charge (but doesn't shoot or handle the ball much, beyond passing to someone else). There was no "glue guy" type on the squad last season. Every player who logged more than 10% of the available minutes at their position also had at least "role player" type Poss%. More than half of the Big East teams had at least one "Marginal" player. Then I noticed that there was no "star" on the squad either, another unusual development for Villanova. The "star" has >25 (or so) of the possessions and a high ORtg. The star gets and converts possessions points at an efficient rate. In 2006-07 it was Curt Sumpter -- note his ORtg of 113.7, a top 500 level performance. The 2007-08 squad had 2 go-to type players (Reynolds and Fisher), but neither was extremely efficient at converting those possessions into points. Villanova's team offensive rating was 104.9, historically a bit low. Note that Reynolds was just a bit better (105.7) than the team's raw rating. Fisher's relatively inefficient rating would (94.1) not have been a problem per se (note Reggie Redding's 99.7 rating), but he took 1 in 4 available possessions when he was on the floor, but converted at a rate less than 1 point per possession, creating a drag on the overall team efficiency. An offensive rating <100 is not unusual for a freshman (see Pena in 2007-08, Redding in 2006-07, etc.), and given Fisher's gained experience, look for his efficiency to be an area of noticible improvement next season.

Casiem Drummond's numbers illustrate the problem with statistics for season-long performances. Comparing his numbers from 2006-07 with those from 2007-08 it appears he made very good (but not impact-type) progress this past season. Improving his ORtg by just over 5% even as he nearly doubled his playing time and pushed his Poss% to the 95^th percentile (15.0 up to 21.4). Yet at the time of the first snapshot (12-3-07) Drum's Min% was 53.8, his Poss% 21.0, his Shot% 18.6, his ORtg 112.8 very impressive even accounting for the out of conference level of the competition. As the chart below shows, Casiem's Min% declined consistently (he appeared in only 3 of the 12 games that followed Penn on 12-1-07) until early February. While he appeared in 2 games following Penn, and no one officially acknowledged his injury until he missed the Hartford game on 12-14, the crashing minutes and steady eFG%, Poss% and Shot% suggest he was injured (most likely) right around the time of the Penn game just before or possibly just after. He returned to action in the second Pitt game (1-30-08) and appeared in every game during that 7 game stretch. His Min% rose during that period, but note also his eFG% declined even as his minutes were rising. Clearly still hampered by the foot injury he logged double digit minutes in only of the last 10 games of the season, appearing not at all in 5 of them.

This season held some promise as Clark's ORtg from 2006-07 showed he could be efficient as a scorer. At 16.2, his Poss% rate held a 50% probability of growing to nearly 20, regular/starter level involvement in the offense. Like 2006-07, he again started quickly, but faded in the middle portion of the season as he battled a malaise described by the staff as "fatigue" -- see chart below. Clark's playing time peaked in the 1-21 snapshot (just after the first Syracuse game) and his minutes declined steadily through February. His minutes bottomed out in early March, while his possessions (Poss%) and effective field goal percentage (eFG%) bottomed out in the 2-25 snapshot. The 10% overall decline in Min% from 2006-07 to 2007-08 is troubling, as it raises concerns (for me) about his durability. I suspect that next season, if the same trend holds, Stokes and Anderson will be able to pick up that time. Boosting his ORtg by 4%, even as his Poss% remained (more or less) steady is a good sign, but he did not grow into either Nardi's nor Sumpter's vacated roles.

The 2007-08 team did not feature a "star" level player, the most efficient scorers (Grant, Clark and Anderson) were either role players (Anderson and Clark) or very limited by minutes. Note there is no indication that Grant, had he received additional playing time, would have been able to maintain his level of efficiency. Though it will be interesting to see how he does in the ACC. The largest (proportionately) possession consumers did not either significantly outperform the rest of the team (Reynolds, Cunningham) or constituted a drag (of varying degrees) on the team's overall performance (Fisher and Pena). Given Fisher and Pena (and Stokes...) were freshmen, their efficiencies are not surprising. They will get better with time. The best thing about freshmen is that they become sophomores.

Coaching Carousal: And Then There Were None...For Now

The list of D1 job openings remained at one for all of 10 days, more or less, but after bumping back up to 2 with the resignation of Ohio University head coach Tim O'Shea (to bring Bryant University's D2 program up to D1), it looked as if the job market was opening up again. The Bobcats wasted little time though, and just days after Mississippi Valley State closed out the 3^rd longest search this offseason, Ohio U. named John Groce, an Ohio State assistant coach, as their new head coach...

Homecoming...
Ohio University head coach Tim O'Shea has made no secret about his desire to coach at a northeastern college. A native of Rhode Island, O'Shea broke into the coaching ranks even as his collegiate career (at Boston College) ended. Holding 5 jobs at 4 different New England colleges over the 16 years after his graduation in 1986, O'Shea took the Ohio job in 2001. After winning the MAC championship in 2004-05 and taking the Bobcats to the NCAA (first round loss to Florida), O'Shea bagan to agressively pursue openings at Providence, Seton Hall, Bucknell, among others, over the past three offseasons. O'Shea accepted Bryant University's offer to lead their program up from D2 to D1 status. With a contract whose length should be enough to see him through the first few years, Bryant is slated to join the NEC after a "probation" period. It will be interesting to see if O'Shea can replicate the success of his Ohio Bobcat teams at Bryant.

Delta Devils Catch a Wildcat...
Mississippi Valley State ended it's 64 day search for a head coach to replace recently departed James Green (bound for Jacksonville State) by appointing former Kentucky basketball player Sean Woods as their new head coach. Woods started as an assistant at TAMU Corpus Christi and then moved over to Texas Christian, moving up from positions in Basketball Operations to an assistant coaching position.

The Bobcats pounced...
And grabbed Ohio State assistant John Groce as their new head coach. Given Tim O'Shea's very public pursuit of open coaching positions throughout the Northeast the past 3 off seasons, it comes as no surprise that Ohio University AD Jim Schaus had a short list ready. Groce, like his predecessor, had a long career as an assistant with stops at a number of top calibre programs.

By the Numbers...
Ohio University marked the 43^rd vacancy this off season. Closing out in just under a week has to be a plus for keeping the Bobcats on track this off season. If the last cycle of the coaching carousal is a good guide, there will be another 6 openings over the next 6-7 weeks. After pauses in June, July and August of 2007, vacancies opened, about half arising from exigent circumstances, the balance due to malice (employer or employee originating...). Two coaches left to pursue NBA jobs, Skip Prosser passed away unexpectedly and a fourth coach was terminated after a judicial process. I suspect however, that number may be a bit high. The most active period for openings was again during the 5-6 weeks between the end of the regular season and the Championship Game. In each of the past 2 seasons 60% of the jobs came onto the market during that period.

The Four Freshmen

I have been looking at the possession-based stats (Min%, Poss%, Shot%, eFG% and ORtg) I collected from Ken Pomeroy's Villanova Scouting Report during the course of the season. Pomeroy updated the possession-based stats for the players at roughly (very roughly mind you...) one week intervals from early December through the end of the season. I had looked at the player usage-efficiency on Villanova teams from 2004 through 2007 in a post from last October and wanted to follow that with another that looked at this year's team. Since I collected the data (more or less weekly) for all the Big East teams, I thought it would be another resource to draw on when considering the state of the conference going into next season. I was curious to see how the much discussed four Villanova freshmen ("true" freshmen Corey Fisher, Malcolm Grant and Corey Stokes, along with redshirt freshman Antonio Pena) who, as it turned out played for most of this season, would track as the team moved from the out of conference schedule, through the Big East regular season, and (though not always a given...) into the post season (the Big East Tournament and it turned out, the NCAAs).

During the Labor Day Canadian Tour reports had Malcolm Grant, an under recognized recruit who had prepped a year before joining the squad, as probably the most "big East ready" of the cohort. Corey Stokes, a sharpshooting wing, had a rough start in the first game (versus tournament host Carleton University), but had, by the end of the tour, righted the ship with several impressive outings. Stokes emerged, behind sophomore Scottie Reynolds and junior Dante Cunningham as the team's third scoring option, according to per game average for points scored. Both Corey Fisher and Antonio Pena had proven to be sufficiently productive to stake legitimate claims to turns in the rotation. I plotted their possession-based offensive ratings, as Pomeroy updated them on his Villanova Scouting Report during the season. A chart of their Offensive Ratings (ORtg) over the course of the season...

For Malcolm Grant the Canadian Tour was a good predictor. Alternately dubbed "Malky" and "Microwave" the freshman had a number of scintellating outings and established himself as a Nova Nation favorite. But (like Canada) his time on the court was limited by circumstances. With Scottie Reynolds and Corey Fisher ahead of him on the depth chart, Grant's playing time (and ORtg, as suggested by the chart above) peaked in mid December, in the midst of the out of conference part of the season. While speculation over "the proper mix" of guards on the floor mushroomed into an unexpected firestorm (about the time the team was hitting bottom in late January/early February) within the Nova Nation, Grant's PT dwindled. After appearing in 21 consecutive games, Grant saw the court in only 8 of the last 14 games, averaging about 4.5 minutes per game (or 8 in those 8 games in which he actually logged time...). Grant maintained "regular" status within the offense (even in the face of diminishing PT), taking about 21.5% of the available possessions and 22% of the shots when he was on the floor. His ORtg's gentle decline parallells the decline in his eFG%, from a high of 73.9% to start the season down to 53.4 at season's end. Grant did not appear in the Kansas game, and logged only 2 minutes in Villanova's win over Siena in the second round. Grant applied and received a release from his scholarship and transferred to Miami (Fla) in the off season.

The Canadian Tour also suggested Corey Stokes' path this past season. Similar to his performance over the Labor Day weekend, Stokes started out very slow, posting an underwhelming ORtg of 68.9 in the first of December. 100.0 is something of a dividing line, as it suggests that a player "scores" about one point per possession taken (rating is expressed as "per 100 possessions"). Since most (but not all...) teams post composite ORtgs of 100 or greater, an individual ORtg <100 suggests an inefficient offensive player. While many well regarded freshman post first year ORtgs in the 90s, at 68.9 Stokes was a very inefficient player. Stokes' Poss% and Shot% through 12/3 were 15.3% and 19.2% respectively, suggesting he was taking role player-level possessions, but regular player-level quantities of shots when he was playing. Though he was getting < 40% of the available minutes at his position, his inefficiency (when he was playing) was a drag on team performance. His offensive rating never returned to that level, throughout December and most of January it fluxuated from the mid 80s to the low 90s. While not obvious at the time, Stokes' week-over-week ORtgs began a steady improvement (no declines) from January 21 onward. He was finally getting it. His playing time through this period gradually increased from 38.2% to 48.5%, with a single week-over-week decline in early February. With the momentum of increased playing time and improving efficiency, Stokes crossed over the 100.0 barrier in early March (see chart above) and did not fall back below through season end. A most promising development was Stokes' ability to also build his poss% and shot% even as his ORtg increased. By season end Corey's Poss% and Shot% were 17.6% (between role player and regular) and 21.9% (regular) respectively. For an individual player to build his ORtg even has his role with the team is expanding suggests there even more room for growth, ie the player has not reached his optimum level of offensive efficiency. Kyle Lowry went from a 97.8/20.7/19.1 (ORtg/Poss%/Shot%) his freshman year to 113.4/22.1/17.8 (ORtg/Poss%/Shot%) his sophomore year. A similar jump (about 15%) in efficiency for Stokes next season would place him among the top 200 or so players in D1.

Antonio Pena's season was a bit more typical of a high major D1 freshman. After a very strong start, a 115.8 ORtg, based largely on an eFG% of 50.0% through 12/3, rising to 122.4 and 66.7 (ORtg & eFG% respectively), his numbers began a season-long decline, settling down to 95.4 and 47.5 (ORtg & eFG% respectively) by 4/3. Pangs of disappointment ought to be mitigated by knowing that Pena's Poss% and Shot% rose (along with his PT) even as his efficiencies declined. And that is the more frequent trend for D1 and NBA players, as their role in the offense increases, their efficiencies decline. Pena's percentage of the available minutes rose from 27.5 (12/16) to 48.4 (4/3), even as his Poss% also rose, from 18.7 to 21.5 and his Shot% rose from 13.9 to 18.8. Other teams discovered that Antonio Pena could put the ball in the bucket and began to give him more attention on defense.

As the NJ HS Player of the Year while point guard for the State Champion St. Patrick's Celts Corey Fisher might have been the most anticipated recruit of the young Wildcat cohort. His start indeed suggested the ink was true, posting an eFG% in the low 50s for much of December and January. His Poss% and Shot% stats suggested Fisher was assuming a starter/star level role in Villanova's offense. His 12/3 Poss% and Shot% numbers were 24.2 and 28.5 respectively, suggesting he would assume an offensive role second only to Reynolds as the season progressed. Unfortunately his scoring did not keep up with his possessions. While his minutes (percentage of minutes avilable) climbed from the high 40s in early December to the high 50s in early January, he maintained his Poss% (24%-26%) and Shot% (27%), even as his ORtg declined from 115 to 108. By mid-February his ORtg and eFG% had declined to 96 and 45 respectively, despite the fact that he maintained Poss% and Shot% at early January levels. He maintained a high level of involvement in the offense, even as he was becoming considerably less efficient. Fisher's slump mirrored the team's through that part of the season and given his major role in the offense, he had therefore, a role in the Wildcat's loss of momentum during that period. The staff began to cut his minutes, distributing them to other members of the rotation. The loss of time was approximately 7 minutes (24 to 17), but point production declined considerably, as Fisher's average for points scored per game dropped from 10.9 to 5.7. Fisher maintained his playing level for assists (2.0 per game for both periods) and defense (as measured by steals and rebounding), but cut back considerably on his shooting. As Corey F works to bring his contribution more in line with the team, I expect fans will see his Poss% and Shot% numbers decrease to regular/starter levels, which should (assuming he better recognizes his scoring opportunities and becomes more selective in his choice of shots...) improve his efficiencies (eFG and ORtg) considerably.

Pace: The Beilein-McCullum Effect

I decided to track pace for all Division 1 programs for the 2007-08 season, much as I had done for the 2006-07 season. The source, as last it was for the 2006-07 season, was Ken Pomeroy's website, a terrific source for statistical data for all D1 teams. At the beginning of each of the past 2 or so seasons, Ken Pomeroy has written something like "...it looks like the pace may be higher this season...", this season it actually happened. The raw pace at the end of the 2006-07 season was 66.85, while this season the average pace for D1 was 67.09, a very modest increase of 0.24 possessions per game. I suppose if you consider there are nearly 5,000 games played in the season, that translates into about 1,250 additional possessions this past season. I plotted Pomeroy's efficiencies over the entire season to see the "shape" of the curve. I noted in a post last August that the 2006-07 chart showed a "hump" (a rise in the average possessions) during the middle of December, the product of seasonal invitational tournaments possibly. I looked for the hump this season and while not a prominent as last season (see the blue line on the chart below), it nevertheless appears again in late December. This time it appears to be due more to a "lessening in the rate of decline" of the average possessions, rather than an actual reversal in the trend of decline. Not as pronounced, it nevertheless disappears during the early conference play, right on schedule. From late December (early January?) the decline, though gentle, is clearly present and persistent through the end of the season. The end of season average raw pace for 2007-08, 67.09, reverses a three or four season trend, but does not quite match the pace from the 2005-06 season, 67.61. In truth I think the changes are largely at the edges, and most likely do not represent a long term trend...a 1 to 2 possession, year-over-year change would be a trend. The three year numbers suggest the college is settled in at (overall) a game in which both teams play for about 67 or so possessions over 40 minutes.

The mid-December hump aside, another interesting pattern is the gentle decline that settles in as conference play begins. This is another year-over-year fixture. The large pattern may be in place, but the devil is in the details. The causes, late season player fatigue, a familiarity with playing styles that leads to more deliberate approaches to game planning and strategy, etc. -- I have nothing more than anecdotal evidence and conjecture to offer at this point. If the reasons are not clear, the trend is persistent. There was a very small rise in pace at the very end of the 2007-08 season (March 23 to April 7), surprising given that at most only 20% of the D1 teams were still playing at that point (between single round eliminations in the NCAAs & NITs, and maybe the 4 or so teams playing out the CBI series?), a number I thought would be too small to move the average even the very small bump of 2/1000 of a possession.

When I compared the average pace for Big East conference teams with the average for all D1 teams, I got a surprise (at first). Comparing those averages for the past three season:

Average Pace
	2006	2007	2008
All D1	67.61	66.85	67.02
Big East	66.61	66.17	67.84

The 2007-08 season was the first (in a long while) since the Big East teams played for more possessions than D1 as a whole. When I considered the coaching changes during the last off season, the reason for the "faster" pace was obvious...though Georgetown's Princeton Offense is absolutely the most deliberate system used in the Big East, both John Beilein and Robert McCullum implemented systems that were notoriously deliberate. A comparison of the Mountaineers' and Bull's season-ending pace for the past three seasons shows the difference:

Raw Pace
	2006	2007	2008
South Florida	64.81	66.39	66.80
West Virginia	64.09	63.64	66.14
Arkansas	69.34	67.61	-----
Kansas State	-----	67.56	-----

Note the jump from the 2007 to the 2008 seasons, especially for West Virginia. I also included the pace for Arkansas (Stan Heath's last stop) and Kansas State (Bob Huggin's previous stop) for comparison. Both teams were above the D1 average for those respective seasons. While each team was below the D1 average, they were above (well above in West Virginia's case...) their previous season's average. Those Arkansas and K State pace stats suggest both the 'Eers and Bulls will "play faster" as Huggins and Heath fully install their systems with their own players.