Monday, September 24, 2007

Pace Notes: Is There a Comfort Zone?

I came across the notion of a comfort zone while reading pace threads over at the APBRmetrics Board a few years ago. The idea is very commonsensical -- a team will play better at a pace (number of possessions) which matches its rhythm of play. A team will get into a flow for example, that will complement the style of play of the coach and players. Remembering the Rollie Massamino teams of the 1970s through early 1990s, the idea seemed plausible. Game announcers were fond of quoting a won-loss stat for Villanova teams when opponents were held to less than 70 points in the game. The stat was pretty impressive, a winning percentage in the .700s as I recall. And it made sense. The pace would allow the team to operate at optimum efficiency while the opponent may well feel uncomfortable playing at a more deliberate pace. I decided to start by looking at all the Big East teams, and calculating a "zone" that would be within 5% + and -, of each team's raw pace. I used Ken Pomeroy's Efficiency Stats Page to determine each team's raw pace. I then gathered the pace of each game each team played from his "Game Plan Page" for each team. A link to each team's Game Plan Page is given in the Overall Record table below:

Overall Record
West Virginia362790.75063.1
Notre Dame322480.75070.1
St. John's3116150.51664.2
Seton Hall2913160.44870.0
South Florida3012180.40065.8

The table is sorted by winning percentage. Teams which played 50% or more of their games within 5% (+ or -) of their raw pace are highlighted with lime green. It turned out to be about ½ of the conference (8 teams). The average raw pace for all D1 teams was about 66.9, so the fact that Louisville, Villanova, South Florida and Connecticut all had raw paces within 5% of the average may explain why they played 50% or more of their schedule within their comfort zone...but that begs the question of why Marquette, with a pace of 67.3 (the closest of all Big East teams to that D1 average), played only 41.2% of their schedule within that 5% zone. The table confirms there is no trend between a team's ability to "hold it's pace" and win its games. The table below marks the boundaries of each team's 5% zone, the number of games played within the zone, the percentage of that team's schedule and each teams won-loss record (with winning percentage) for games in that 5% zone

Comfort Zone
Range# of SchedRecord
South Florida65.869.162.5150.500870.533
Seton Hall70.073.566.5140.483860.571
West Virginia63.166.259.9160.4441150.688
Notre Dame70.173.666.6140.4381040.714
St. John's64.267.461.0100.323640.600

The winning percentages for Providence and Pittsburgh were virtually the same within the 5% zone as they were overall. Those teams highlighted in burnt orange (Georgetown, Villanova, West Virginia and Notre Dame) had 5% zone winning percentages less than their overall winning percentages. For those teams the comfort zone was not, apparently, very comfortable. For Georgetown, the team with the largest percentage difference, this would have translated into a -3 margin in wins. Those teams (Connecticut, Syracuse and St. John's) whose 5% zone winning percentage was greater than their overall winning percentage are highlighted in yellow. Those teams whose difference exceeded .100 percentage points (a 10% or greater margin), are highlighted in aqua. This > 10% group is the largest subset in the conference, including not only teams with winning overall percentages, but teams like Cincinnati, Seton Hall and South Florida, whose winning percentages were less than 0.500. The team who seemed "most comfortable" in their zones (as measured by the difference between their 5% zone winning percentage and their overall winning percentage) were Louisville (+17.4%), Marquette (+15.1%) and South Florida (+13.3%). For Louisville and South Florida, teams who played ½ their schedule within the 5% zone, the effect is pretty clear. That 10 of 16 schools had a higher winning percentage of games within their 5% zone suggests there may be some value to the notion.

Tuesday, September 18, 2007

Oh Canada! Part 4: Player Efficiencies

A four game run may be bit light to look at (possession-based) efficiency statistics, but I decided to crunch the numbers and take a look. The offensive & scoring numbers look like this:

D Anderson32.520.618.6134.370.81.5058.350.0
S Clark45.618.418.780.044.10.9635.323.5
J Colenda9.
D Cunningham49.425.024.4104.450.01.1570.854.2
C Drummond43.118.111.7111.060.01.22170.0100.0
C Fisher51.321.010.8114.077.31.63118.2100.0
M Grant50.615.119.9150.675.01.5310.010.0
A Ott18.125.630.559.640.90.7036.40.0
A Pena51.321.322.678.656.
R Redding37.516.517.4101.746.20.9315.47.7
S Reynolds55.024.924.7176.298.11.9051.940.7
C Stokes56.318.725.9134.867.21.376.96.9

My offensive rating calculation biases in favor of assists (and against offensive rebounders...), so Grant, Fisher, Reynolds and Redding's ORtgs may be a little high while Cunningham, Pena, Drummond and Clark's may be a little low (compare their ORtgs & PPWSs to get a sense of the differences...). But for estimating and getting a general sense of their play, these numbers will suffice. A few quick impressions:

  1. Drummond & Fisher got to the line a great deal relative to the number of FGAs they took (check their FTRates). Note they did not (comparatively speaking...) take a lot of shots when they were on the floor (in the range of 10% - 12%, definitely marginal -- less than even a role player), but the relatively high number of fouls suggests the Canadians had problems coping with their respective offensive games...that could be a good sign for the season.
  2. Their (relatively low) FTRates (and rebounding percentages -- see below) suggest that Grant, Pena, Redding and Stokes are back court/wing players. On site spectators say Pena was playing a bit tentatively. Maybe that explains his lack of contact and free throw opportunities.
  3. Eight players logged 40.0% or more of the available time at their positions. Two other (Redding and Anderson) logged >30.0% (but <40.0%) of the available time at their positions. I suspect those 10 will see time early in the season. The schedule running up to the Old Spice Classic is a little abbreviated -- Bloomsburg (exhibition), Stoney Brook and Bucknell -- so the fans may see 10 or more players in each of those games as the staff settles on a rotation.
  4. Given the level of competition, PPWSs <1.0 (note those instances where the PPWS is <1.0 the field goal efficiency -- eFG% -- is also <50.0) should be a little disconcerting -- those men just did not shoot well. Clark's, Redding's and Ott's shooting problems did not jeopardize the team's chances for victories because each had a (comparatively) small role in the team's offense. Redding took "role player" level possessions and shots (<17.0 or so), while Clark was a borderline "regular". Ott consumed comparatively large proportions of both possessions and shots, but played relatively few minutes.

The rebounding and defensive efficiency stats:

Dwayne Anderson32.513.820.132.316.2
Shane Clark45.69.917.58.621.6
Jason Colenda9.
Dante Cunningham49.415.910.33.012.1
Casiem Drummond43.17.820.223.213.9
Corey Fisher51.30.08.534.337.7
Malcolm Grant50.62.24.337.318.7
Andy Ott18.112.432.00.015.4
Antonio Pena51.38.812.73.430.4
Reggie Redding37.53.017.451.846.0
Scottie Reynolds55.
Corey Stokes56.32.012.913.920.9

OR% and DR% represent the percentage of available offensive (OR%) and defensive (DR%) rebounds the player was able to collect while he was on the floor. For defensive rebounding 20.0 is a very good number (the average in the Big East last season among players with 40.0 or more of the minutes in their respective positions was around 14.0). For offensive rebounding 10.0 is a very good number (the average in the Big East last season among players with 40.0 or more of the minutes in their respective positions was around 6.0). Turnover rates (TORate) and assists (ARate) are per 100 possessions (to establish a comparative standard for all players). A few observations about the freshmen:

  1. Corey Fisher had a typical freshman tour. He did a few things well (assists -- ARate was 34.3, shooting, getting to the line), a few things poorly (TORate was 37.7, no offensive rebounding) and somethings very tentatively (despite good shooting he took comparatively few shots relative to the number of possessions -- 10.8). The rate at which he got to the line suggests he gave his defenders a very difficult time.
  2. Malcolm Grant had a very good tour. Those who suggested he appeared "more ready" at this point have a good deal of support in the numbers. His shooting was accurate (eFG, PPWS) and he was comfortable taking the shot (compare Grant's Shot% to Fisher's) when it was available. Grant also had a terrific ARate (37.3 -- according to Ken Pomeroy's Villanova Scout Page the best ARate in D1 last season was 47.0; Scottie, with a 30.6 was ranked 69th), as did Redding (51.8!), Anderson (! 32.3) and Fisher.
  3. Corey Stokes played well within the offense. Note his Poss% and Shot% are around 20.0, suggesting he functioned within the offense and took "his share" of the possessions/shots available. He shot the ball well (see his eFG & PPWS), though given the relatively few free throws to FGAs, it appears he took jumpers and most likely stayed out around the 3 point line, rather than putting the ball on the deck and penetrating, to score his points. His rebounding was less than I hoped and he might also work to limit turnovers. His shooting was terrific; had Reynolds not shot the lights out in Ottawa and the Raven's Nest, the talk around the 'Nova Nation would be Corey Stoke's shooting.
  4. Antonio Pena had a pretty good tour. He was comfortable taking his his shots and possessions (see his Poss% and Shot%). His shot making was accurate, but the turnovers killed him (TORate was 30.4 -- he lost about 3 in 10 of his possessions).

Overall I suspect the staff had to be pleased with those performances. There is much to work on over the next 6 weeks, but there is an excellent talent-level with which to work. Among the returning players there were few surprises, Scottie shot out the lights in Ottawa; Dante pulled down offensive rebounds like Bump did while shooting very well; Shane Clark, hobbled by knee problems last season, showed some offense (though with a bit of rust) and rebounded well, suggesting he had regained his mobility. Of the other returning players:

  1. Dwayne Anderson posted some strong numbers in the time he played. Granted about half of his PT was taken against Queens University (10 points & 10 rebounds in 25 minutes), he showed comparable stats in the other three games (13 points & 3 rebounds in 27 minutes). Anderson will face much stronger competition for playing time this season.
  2. Reggie Redding's tour was, aside from his ARate (and defensive rebounding), largely a struggle. He continues from last season to struggle with his (see his eFG% & PPWS) shot. And his strong assist performance was largely negated by a high turnover rate. Reggie performed well In the structure offense the staff used in the Carleton University game (the first game of the tour). Redding dished out 5 assists (though shooting 0-4, 1-2, 1). In the less structured games against Ottawa and the Carleton All-Stars Redding rediscovered his shot (5-9, 0-0, 12), but had diffuculty involving the rest of the offense (4 assists to 5 turnovers & he fouled out versus the Carleton All-Stars).
  3. Casiem Drummond showed a good deal of consistency in the middle. His shot dropped as he played within the offense. He rebounded well (could do a bit better on the offensive boards) and showed some assist capabilities (ARate was 23.2, strong for a front court player).

Of the ten players who played the most minutes in Canada, Drummond, Redding and Anderson received the fewest minutes. All three should be in the rotation at least through the early part of the season (the run-up to Orlando? through December?). Who remains in the rotation, and how much time he/they get after that may well determine the look of, and prospects for this season's team.

Saturday, September 15, 2007

Oh Canada! Part 3: Player Summaries

I am posting the summary stats for each of the team members who saw minutes in Ottawa. The player stats (given the blog margins...) will be presented in three groups -- a general background (Games played, minutes), shooting (offense?) and a third table with rebounds (offensive, defensive and total), assists, turnovers, blocks, steals and personal fouls. Given the questions & discussions about the roles several players will have on this year's team, I was very curious to see how the minutes were allocated among the individual team members.

Dwayne Anderson524
Shane Clark733
Jason Colenda152
Dante Cunningham793
Casiem Drummond694
Corey Fisher824
Malcolm Grant814
Andy Ott292
Antonio Pena824
Reggie Redding603
Scottie Reynolds883
Corey Stokes904

For those wondering about Scottie's minutes next season, know he logged an average of 29.3 minutes per game (mpg). He sat for the second game in the set (against the Queens Golden Gaels). The stats confirm what those fans who made the trip observed -- Fisher and Grant did log about the same number of minutes. Antonio Pena also logged minutes in the low 80s, no doubt the staff wanted to see these freshmen (redshirts and trues) as much as the average Villanova fan. The surprise is Corey Stokes. Honored by McDonalds as an HS All-American, invited to the U19 tryouts and generally recognized by the scouting services as the highest ranked HS baller in Villanova's entering class, nevertheless, most Nova fans had shown greater anticipation about Corey Fisher's debut as a Wildcat than Corey Stoke's. Stokes, despite what seemed to be first game jitters against Carleton (5 minutes, 2 fouls, no points on 0-2 shooting), put together three very nice efforts in the games that followed. And he in fact finished with the most minutes played out of the entire squad. He played practically the entire Queens game (38 minutes) and averaged about 28.3 mpg over the last 3 games -- almost as much as Scottie (see above). I was not surprised that Jason Colenda (the Wildcat's latest walk-on) logged only 15 minutes in two games, but I was surprised that Andy Ott, despite rumors of a few dings did see only 29 minutes in just two games. I understand the staff may have seen little need to risk further injury on an exhibition tour. Reggie pulled a solid 20 mpg for the 3 games in which he appeared. Like the other starters/significant returning players from last season, he was held out of the Queens University game to keep him fresh for the Ottawa game later that day. As for offense:

Overall3 Pt2 PtFree Throw
D Anderson71235476723
S Clark717156124619
J Colenda020101000
D Cunningham1224011223131737
C Drummond61000610101722
C Fisher7113546111328
M Grant1320489122232
A Ott4111437049
A Pena13230213210226
R Redding51326371213
S Reynolds20271316711111464
C Stokes13291327022241

That Scottie led all scorers is hardly shocking (though his average of 21.3 caught my eye). That the next two scorers were the freshman Stokes and Junior Dante Cunningham is a bit of a surprise. Cunningham appears to have picked up from his mid-season form, though he may have backslid a bit on his free throws. Stokes, a wing guard/forward who started slow, showed very good point production in the last three games. Malcolm Grant, the third guard brought in this year, showed a greater inclination to shoot than fellow freshman Fisher, and Grant also had a slightly higher accuracy (at least from inside the 3 point arc). Pena, Drummond and Clark's scoring numbers (all averaged about 5 - 7 ppg) suggest there may be more offense in the frontcourt than in prior seasons. If this tour is a predictor for next season, it appears that Scottie, working off the ball more than last season, will serve as a very effective first option on offense. Cunningham and Stokes (when they are available) will be effective 2nd (inside) and 3rd (outside) options. Grant, Fisher, Clark, Pena and Drum virtually guarantee that the Wildcats will have 4 - 5 realistic scoring options on the floor at any one time. As for the rebounding, steals, blocks, etc.:

D Anderson491363107
S Clark4111525612
J Colenda00000100
D Cunningham7714141023
C Drummond31215531032
C Fisher0661011913
M Grant13484703
A Ott281002700
A Pena491319916
R Redding1910981106
S Reynolds01010106607
C Stokes1101146501

Frankly it is a bit disquieting that no one averaged a double-double against this competition. Given Cunningham's scoring he would have been the most likely candidate. But on average, he was outrebounded by both Shane Clark and (little used) Andy Ott. Overall rebounding stats were lower than I anticipated, Ott and Clark averaged about 5 rebounds per game while Cunningham had an even more anemic 4.7 average. The 'Cats have been effective at rebounding over the past 4 - 5 seasons, but the tour results suggest this may be an area that requite attention and additional work before the start of the season. A few other observations:

  1. Dwayne Anderson had a very productive run with comparatively few minutes. Dwayne tied Scottie for most steals, though he did it in 3/5 the amount of time alloted for Scottie. Those steals were distributed across all 4 games (1-2-3-1), not concentrated against the relatively weaker Queens University squad. In addition to steals, he was 5th for defensive rebounds, but showed a very good DR% (about 20%) for the amount of time he played. His distribution of 2 point to 3 point FGAs was fairly even, suggesting he is diversifying his game a bit more than last season. The question, given the amount of time he played, is whether this progress comes a bit late.
  2. Andy Ott played very little, due at least initially, to a slight injury. Ott was one of the red shirted front court players that many have been anxious to see play this season. He was held out of 2 games and saw extensive time in only the Queens game. That the Queens game is the only one in which both he and walk-on Jason Colenda saw double digit minutes, Unless something happens in practice I suspect Ott will see only very occasional minutes this season.
  3. Junior Frank Tschuisi was held out of all four games. He is rumored to have an injury.
  4. Of all the 'Cats, Reggie Redding had the most uneven performance. His 5 assist, 3 steal, 2 rebound effort in the first game (against the Carleton Ravens) was marred by an 0-4 shooting night. He scored a single point on a 1-2 night at the line. He followed his 12 point (5-6; 2-2; 3-4) scoring outburst against the Ottawa GeeGees (which also included 5 rebounds, 3 assists and 2 steals) with an 0-3 0 point, 1 assist and 2 turnover effort against the Carleton All-Stars, cut short when he fouled out in 19 minutes of play.

Tuesday, September 11, 2007

Oh Canada! Part 2: Team Defense

Villanova's team defense has fallen down a bit over the past two seasons. From a high ranking of #4 (per Ken Pomeroy's possession-based stats) in the 2004-05 season the Wildcat's defensive rating has drifted downward to #18 last season. The culprit (as many fans who watched the games know...) has been field goal defense, principally on the perimeter. Numbers for rebounding (OR% -- opponent's offensive rebounding rate) and turnovers (TORate) have been fair to good, but fouling (FTRate) has been weak -- the 'Cats have been known to foul a good deal, a byproduct of the aggressive style of defensive play. Like perimeter field goal defense, Villanova's defensive free throw rate (free throws per 100 field goal attempts) has regressed over the past 3 seasons. Last year's 42.0 earned the 'Cats a rank of #262 (out of 338 in D1). The overall stats look very good for the defense on this tour:


Of particular interest is the eFG%, which at 36.3, is extremely low (the Canadians weren't hitting), well below Connecticut's #1 defensive eFG% of 41.9 last season. The FTRate (41.4) appears to be a bit high, possibly reflecting the style of Canadian (international) officiating, but along with an uncharacteristically low DR% of 30.4, something to watch. Individual game breakdowns, given the different levels of competition and team composition, should provide better insight on some of these numbers:

Carleton All-Stars38.320.036.634.488.60.83

Keeping all 4 Canadian teams' eFG% in the 30s is a good sign the shot defense might be back. That the Golden Gaels registered the lowest eFG% is not surprising. The percentages given by the two Carleton squads however are especially encouraging. The Wildcats were able to force a turnover rate of about 23.1 last season, so the overall exhibition result (22.8), while statistically insignificant, is not particularly good. This may rise as the rotation is settled. The opponent's OR% of 30.4 was an improvement over last year's 32.1, which was surprising given the larger front line offered by the 'Cats. The poorer showing against the Carleton All-Stars (36.6) however, raises concerns. Losing Sumpter and Sheridan, combined with less than dominating rebounding on this tour suggests this may be an area of concern. Ottawa's surprisingly high FTRate (71.2) is obviously a anomaly, something which might be said about their OR% and eFG% from that game as well. The higher paced game may have developed the characteristics of an AAU type game -- note the DRtg% of 96.1 for example. Not high as a result of a D1 game, but about 15.1% higher than the next nearest game (the Carleton All-Stars) and suggesting an inconsistency in the defensive effort game-to-game (note the Queens game). Is the higher pace responsible for the apparent degraded defense? The overall raw DRtg% of 81.1 may again be consistent with the level of the competition, but is nevertheless encouraging. Hopefully these numbers will serve as predictors for a strong defensive presence next season.

Monday, September 10, 2007

Oh Canada! Part 1: Team Offense

Villanova's 4 game 3 day sojourn to Ottawa afforded fans a taste of the team Coach Wright will take into next season. The extended exhibition tour gave me 2 - 3 more games to look over than I might see during the "regular" exhibition part of the season. While there is little doubt this was exhibition basketball, some of the numbers should encourage Wildcat fans.

Some background about the tour and competition. Villanova faced four opponents over the course of a three day tour. The 'Cats faced Carleton University, the five time CIS champion (the Canadian equivalent of the NCAA) on Saturday evening, September 1, Queens University at noon on Sunday September 2, the University of Ottawa a semifinalist in last year's CIS championship (defeated by Carleton) at 7pm (central time) on Sunday September 2 and a team of ex-Carleton players who have pursued pro careers in Europe and Australia (the Carleton All-Stars) on Monday September 3. The overall offensive (and pace) stats:


Those are very strong numbers, even for an exhibition series. The eFG% (field goal efficiency) is much higher than last season, suggesting the Wildcats, even against tougher competition, will most likely see a greater shooting accuracy than they have in a few years. The OR% appears to be right around where it has been historically under the Wright regime, while the team got to the line pretty consistently (FTRate), suggesting they showed an appropriate amount of aggression on the ball & backboards. The (raw) Offensive Rating (ORtg) is encouragingly high, even for an exhibition series. The lack of practice/preparation time, even with new players to integrate into the system did not depress the ORtg, which was better than last season (107.7) by about 5% has to be, with all caveats factored (competition, lack of prep, inexperienced players, no starters for one game, etc.) very encouraging. The overall numbers however, do mask a game-to-game variation which, if examined closely, suggests there are still a number of things to resolve. I have put together the pace and offensive and defensive efficiency statistics for each of the four games, and they appear below:

Carleton All-Stars70.073.828.621.442.9110.01.52

The Wildcat's field goal efficiency caught my eye first. The 'Cats converted in the 48-49 range (48.8) last season. Pushing that number above 50 would be very encouraging. But the playing the toughest team first (Carleton) the 'Cats registered a rather disappointing 43.3. The efficiency for Ottawa game, 82.1, was the product of a 9-11 3 point shooting night, complements of Scottie Reynolds (5-5) and Corey Stokes (4-6). Not likely they will hit 81.2% of their 3s every night. But the eFG% for the trailing 3 games (Queens, Ottawa and the All-Stars) showed considerable improvement over the first game (Carleton U.). Maybe the 'Cats are indeed on their way. The eFG% (effective field goal percentage) against the all-pro All-Star squad, well into the mid 70s, was an excellent number considering the competition. Turnovers on the other hand, showed a downward trend. As the pace quickened the 'Cats lost more (and a higher percentage) of their possessions. This may well be due to the new players (more on individual performances in a later post) and therefore naturally correctable with more playing time. The offensive rebounding however is a bit more troubling, and may in some ways represent the dilemma the staff will face this season. Under Coach Wright Villanova has had a number of very good rebounding guards (Lowry and Foye stand out, but Ray, Nardi and Benn contributed on rebounding as well), but this squad's guard contingent (wings included) seemed less inclined to grab missed shots. The OR% versus the two Carleton squads in particular is bothersome. The Ravens are a well coached/schooled program headed up by Dave Smart, an assistant coach for the Canadian National Team. And Villanova's OR% against both of those squads hovered around 20, well below their tour average and well below their OR% from last season (39.1). The OR% from the Queens game is nearly the opposite of the 2 Carleton games, but consider that Ott, Pena and Drummond played significant minutes against the Golden Gaels, while the starting lineup sat for the entire game. That Drum, Ott and company were able to snag 41.1% of the Wildcat's own missed shots should not be a surprise. That it should be only marginally better than last season's ending OR% (and produced against a not very competitive university program) should raise some modest concerns at this point.

The average raw pace for D1 ball last season was about 67.5, while exhibition games tend to be in the 70 or so range. The Carleton game came in a bit below those two paces and while eyewitnesses reported that Carleton initiated the deliberate pace, I was curious to know if the Ravens played that slowly for two other American teams, the University of Illinois (Urbana-Champaign) and the University of Alabama, who alos toured the Capital Region that weekend. The answer is no. The pace for the Illinois game was about 65.2, while the Ravens victory over Alabama took about 68.5 possessions. Clearly Villanova's game was the lowest paced of the three. If, as suggested by fans who were present to watch the game, the Ravens slowed the tempo, it might suggest a deliberate strategy to limit possessions as a means of increasing their chances for a win. Villanova's raw pace last season was 66.5, about 3 possessions lower than the overall (raw) possessions for the Tour.