Draft Study: A Prospect's Level of Competition
A quick look back at the sophomore season of Derrick Williams reveals an interesting case study in numbers. Williams was not an unknown going into his sophomore season as he won the Pac 10 Freshman of the Year award. He entered the year as the fifth ranked power forward and fourth ranked sophomore on this site’s preseason rankings. By the start of the Pac 10 season, he was sixth on the Big Board, and was regarded as a potential top pick by mid-January and has been in the top three ever since. It was a remarkable rise driven by jaw dropping numbers and efficiency. Before Arizona started conference play, Williams was averaging 19.3 points per game over just 27 minutes, pulled in 2.9 offensive rebounds per game, and shot 63 percent from the field as well as 68 percent from three. He averaged 2.1 points per shot. It really was an awe-inspiring rise, but were the extenuating circumstances for why it happened? In those first thirteen games where Williams rose to prominence, only three of those opponents went on to make the NCAA tournament, two games of which Arizona lost, and only four were against opponents who finished the year in the RPI Top 100. At the end of the season, Arizona finished the season with a 30-8 record. Twenty of their thirty-eight games were played against teams who finished the season in the RPI Top 100. Arizona was a very respectable 13-7 in those games. Williams averaged 20.4 points per game on 53.9 percent shooting and 48 percent from behind the arc to go with 9.5 rebounds per game. Those numbers are undoubtedly impressive. The low assist total is not ideal for someone who projects as a first or second option on the next level, but even beyond that it is not at the Basketball Jesus level he had otherwise put up. After all, in the 18 games against teams outside the RPI Top 100, Williams averaged 18.6 points per game while shooting 68 percent from the field and 75 percent from three. It was a performance out of Playstation. The issue is not for Williams or Arizona alone. Connecticut played 30 of its 40 games against RPI Top 100 opponents. North Carolina and Kansas played 27 and 26 games against RPI Top 100 opponents respectively. Conversely, Morehead State and Oakland played 7 and 9 games against RPI Top 100 opponents. It raises an interesting question. How much weight is given to performances of NBA prospects against lesser opponents, and do those games matter to a player’s NBA prospects?
The rhythm of a college season is not unknown. Major college programs, where most NBA prospects play, use the non-conference to find which plays and schemes will work with their personnel, get the players on the same page with one another, take a snow bird trip to help sell recruits, get through the cross-conference made-for-tv showcase games, and do their best to limit the damage for any growing pains that take place. They want to have any issues settled by the time the conference season begins. Obviously, the conference slate is outside of the team’s control as is the postseason tournament, but the part of the schedule that can be controlled accounts for between a quarter and half of the season. Weaker programs try to pack on as many wins as possible early knowing that the conference season will be a struggle. Stronger programs will have their fair share of body bag games, but also mix in higher caliber mid-majors for low risk resume builders. It creates a sizable scheduling disparity. Even within these guarantee games, there is a disparity in how teams and players treat them. Some teams experiment with rotations and extend their bench to see what they have, others go for the jugular and put up comical numbers.
The disparity in the level of competition is heightened because the cyclical element of college programs and conferences, and subsequent vast swings in the overall strength or weakness of a schedule. The conference season is typically the best time to evaluate players because of the high level of preparation that is involved. The familiarity that exists between the programs and of the players creates an environment where flaws are more readily exposed. The ACC carries arguably the strongest brand name in college basketball. In the 2008-09 season, the ACC was rated by the RPI system as being the strongest conference in the nation. The league received seven NCAA tournament bids, and North Carolina won the NCAA title. That summer, the league had seven players get selected in the first round of the NBA Draft. It was a year where the conference product lived up to the brand name. This past season, the ACC was ranked fifth by the RPI behind the Big East, Big Ten, Big Twelve, and Mountain West Conferences. The league was weak, but was it perceived that way? While teams like Wake Forest, Georgia Tech, Maryland, and NC State are historically strong programs. They suffered through rebuilding seasons, but did that translate to perception?
Perception Is Not Reality
The problem is that the brand of a college program or a conference outweighs the actual performance for that year. The Mountain West Conference is not regarded as a “Big Six” league, but for 2010-11 the RPI had the MWC ranked fourth strongest league in the country ahead of the ACC, SEC, and Pac 10. In fact, it was the second straight year the Mountain West finished the college season as a stronger conference than the Pac 10. It does not mean that the Mountain West will be an elite conference next year, certainly conference realignments are going to hurt the league, but it should go some way towards respecting the strength of the league in 2010-11. It was a year where San Diego State and BYU played in the Sweet Sixteen, UNLV made the NCAA tournament, and both Colorado State and UNLV played in the NIT. This is not always the case, and players coming out of the league – namely Jimmer Fredette, Kawhi Leonard, Malcolm Thomas, and Billy White have to fight the stigma of playing a weaker level of competition than they actually did. This past season, San Diego State played eighteen games against opponents who finished the season in the RPI Top 100, ten of those games were against opponents who finished the season in the RPI Top 50. Similarly, BYU played eighteen games against opponents who finished the season in the RPI Top 100, and twelve of those games were against those who finished the season in the RPI Top 50. By comparison, USC and UCLA each only played seventeen games against opponents who finished the season in the RPI Top 100. Due in part to a soft non-conference schedule and a down ACC, Virginia only played sixteen RPI Top 100 games.
Mid-Majors and Low-Majors
The annual fluctuations in mid-major conferences are even greater than in the traditional power conferences. On a name basis, the WAC is not perceived to be a weak league, but this year the conference was down considerably. Not only did Utah State roll through the league with just one loss to receive the league’s only NCAA bid, but the team to finish with the second best RPI ranking was Boise State at 114. This followed a year where they had two bids and four draft picks. What does it say about Greg Smith that he regressed in a year where his conference was as weak as it has been in recent memory? Conversely, the Colonial Athletic Association teams was stronger than it is typically perceived. Even site President Aran Smith fell into this trap when expressing doubt about the numbers amassed by Charles Jenkins. The Colonial Athletic Association was a tale of haves and have nots. The league produced three NCAA tournament teams in George Mason, Old Dominion, and the Final Four appearing VCU. Hofstra and James Madison also went to a postseason tournament after finishing 86th and 88th in the RPI rankings. Drexel also finished in the RPI Top 100 at 74th. This means that six of the Colonial Athletic Association teams finished in the Top 100. The other six teams in the league had losing records and were generally best referred to as awful, but that does not take away from the caliber of the top six who not only were good but also with the exception of Hofstra and James Madison excellent defensive units by just about any metric. If the hypothesis about placing a greater premium on performance against good teams, should it matter how horrendous the bad teams are?
Looking at the numbers of players from low major conferences is even more difficult. The opportunity to play against the type of competition players at Connecticut, Kentucky, or North Carolina see is rare. This leads to issues of sample size where a low major program might get one or two games against high quality opposition. Even in the rare instance a low major school gets a chance to big a high level opponent outside of a tournament, it is almost always on the road. The pro prospects from those programs also often do have conference opponents capable of challenging an NBA caliber talent. Even in non-conference competition, opponents will use a gimmick defense knowing that there is only one player who can possibly beat them. For example, Stephen Curry had very different splits in his sophomore and junior seasons. In 2007-08, Davidson played four RPI Top 100 teams prior to mid-December where they went 0-4, and then four more in the NCAA tournament where he had a coming out party. In the eight games, Curry averaged 27.4 PPG on 43.7 percent shooting and 41.1 percent from three. Following the graduation of three senior starters, teams were able to gear up defensively on Curry. His shooting percentages went down across all splits and Davidson was not as good of a team, but how much of his decline was the way in which defenses approached him after losing several key teammates?
|2007-08 Total (29-7)||36||33.1||8.8||18.2||48.3||4.5||10.3||43.9||4.3||8.0||54.0||3.8||4.2||89.4||0.8||3.8||4.6||2.9||2.6||2.0||0.4||2.4||25.9|
|2007-08 Non-Conference Reg (3-6)||9||32.3||7.8||17.8||43.8||4.1||10.4||39.4||3.7||7.3||50.0||2.6||2.7||95.8||0.9||3.8||4.7||2.6||2.9||1.7||0.7||3.1||22.2|
|2007-08 Big Twelve (23-0)||23||32.6||8.9||17.7||50.5||4.4||10.4||45.7||4.5||8.0||56.3||4.0||4.5||88.3||0.8||4.0||4.8||2.9||2.7||2.0||0.3||2.1||26.2|
|2007-08 RPI 100 (3-5)||8||37.4||9.1||20.9||43.7||4.9||11.9||41.1||4.3||9.0||47.2||4.3||4.8||89.5||0.5||3.4||3.9||2.9||2.1||2.6||0.4||2.8||27.4|
|2007-08 Outside RPI 100 (26-2)||28||31.9||8.7||17.5||49.9||4.4||9.8||44.9||4.3||7.7||56.3||3.6||4.0||89.4||0.9||3.9||4.8||2.9||2.7||1.9||0.4||2.3||25.4|
|2008-09 Total (27-7)||34||33.7||9.2||20.2||45.4||3.8||9.9||38.7||5.4||10.3||51.9||6.5||7.4||87.6||0.6||3.8||4.4||5.6||3.7||2.5||0.2||2.4||28.6|
|2008-09 Non-Conference Reg (7-4)||11||35.4||9.2||21.7||42.3||3.3||10.5||31.3||5.9||11.3||52.4||6.4||7.4||86.4||0.8||3.2||4.0||6.6||4.3||2.9||0.4||2.5||28.0|
|2008-09 Big Twelve (19-2)||21||32.6||9.1||19.1||47.5||4.0||9.6||42.1||5.0||9.5||53.0||6.7||7.6||88.1||0.6||4.0||4.6||5.1||3.1||2.4||0.1||2.2||29.0|
|2008-09 RPI 100 (2-5)||7||37.0||8.9||24.7||35.8||3.4||11.9||28.9||5.4||12.9||42.2||6.1||6.9||89.6||0.6||5.1||5.7||5.6||6.3||2.4||0.4||3.0||27.3|
|2008-09 Outside RPI 100 (25-2)||27||32.8||9.3||19.0||48.6||3.9||9.4||41.9||5.3||9.7||55.2||6.6||7.5||87.2||0.6||3.5||4.1||5.6||3.0||2.6||0.2||2.2||29.0|
This season, the highest rated prospect from the lower level of Division I is Kenneth Faried. He led the nation in rebounding this past season pulling down 14.5 rebounds per game at Morehead State. His four year body of work put him among the elite rebounders of all time. He now holds the NCAA modern-era (post-1973) career rebounding leader, and ranks eleventh in NCAA Division I history for all players. It is commonly stated that rebounding translates across levels, but that is really only a half truth. Rebounding is as much a statistic of effort, skill, and size, but also of opportunity. The more quality rebounds a team has alter the rebounding statistics for each of the individuals who make up that team. No matter how many players on a team go for a rebound, there is still only one ball coming off the rim to be had. The opposite is also true. The tremendous numbers Faried amassed are not only a product of his athleticism, motor, and feel for rebounding, but also situation. Faried only had one other teammate in the Eagle rotation who was listed as being taller than 6’ 4,” and that player, freshman Drew Kelly, played 18.5 minutes per game. In previous college seasons, Jason Thompson, Larry Sanders, and Paul Millsap have been elite rebounders on mid-to-low major college teams and have been less than elite level rebounders in the NBA. The question with really no solid answer is what weight should be given to numbers, and what should be given to the situation that helped make them historic?
Splitting the Stats
The attached file at the bottom of article includes the splits of the players who eligible for the 2011 NBA Draft based off their performance in games against opponents who finished in the RPI Top 100 and then against all others. A more detailed breakdown was created for more highly projected players where their numbers are broken down by how they played overall, in regular season non-conference games, conference play including the conference tournament, games against the RPI Top 100, and games against opponents outside the RPI Top 100. The advantage of isolating the games against the best teams in college basketball is that it allows a comparison of players over a comparable level of competition, the best 100 teams the NCAA has in a given year, that can not otherwise be seen due to the disparate level of play across the NCAA Division I ranks. It is not intended to fault prospects for playing and performing against the level of competition that they did, but rather to simply remove the games on the schedule where the results are almost predetermined so as to make comparing players a bit more standardized.
A Few Noticeable Findings
* Reggie Jackson of Boston College had a breakout junior season. He finished the year shooting 50 percent from the field, better than 40 percent from three, and had 1.9:1 assist-to-turnover ratio. This combined with his size and athleticism is why Jackson has been a riser in this draft. The concern is that there is a huge divide in his efficiency in games against the RPI Top 100 and those outside. He was dominant against lesser opponents, shooting 60 percent from the field and 50 percent from behind the arc, but his numbers against RPI Top 100 competition were not quite as good. His three point shooting was still at a respectable 35 percent but only shot 42.6 percent overall. His assist total was also lower by 1.2 per game- 5.1 and 3.9 per.
* A number of power forwards and centers in the 2011 draft class have below 50 percent field goal percentages on two-point attempts, which is a statistic that has not boded well for players being able to finish in the NBA. In past years, Jermareo Davidson (39.1 percent in his final college season against RPI Top 100, 40.8 percent as a rookie), Craig Brackins (41.1 percent, 36.4 percent in 3 total games), Daniel Orton (43.2 percent, DNP), Earl Clark (44.8 percent, 37.0 percent), Brook Lopez (45.6 percent, 53.2 percent), Ekpe Udoh (45.7 percent, 43.7 percent), Hassan Whiteside (46.7 percent, No shots taken in 1 game), Glen Davis (47.7 percent, 48.4 percent), Solomon Alabi (47.7 percent, 20.0 percent over only 12 games), Tyler Hansbrough (48.4 percent, 36.7 percent), Luke Harangody (49.0 percent, 45.0 percent), JaVale McGee (44.7 percent, 49.4 percent), and Kosta Koufos (49.2 percent, 50.8 percent) were below 50 percent in their final collegiate season. The front court draft prospects who were under 50 percent include: Lavoy Allen (42.3 percent), Troy Gillenwater (44.3 percent), Sam Muldrow (44.4 percent), Keith Benson (47.1 percent), Greg Smith (47.8 percent), Matt Howard (48.0 percent), Malcolm Thomas (48.1 percent), Alex Tyus (48.6 percent), Jerai Grant (48.8 percent), [Player: Matthew Bryan-Amaning], and Justin Brownlee (both 49.1 percent).
* Chris Singleton has been pegged as an NBA-ready defensive player whose offensive game needs to come a long way. He certainly has ideal length and athleticism, but his rate of committing fouls, 4.8 per 40 minutes, against RPI Top 100 opponents is a concern. Most notably, Singleton battled through early game foul trouble in games against Notre Dame, Texas A&M, Duke, and North Carolina, four of Florida State’s most difficult games this past year, but it was not just limited to those games. The impact of being foul prone often goes overlooked when evaluating a prospect, but the propensity to commit fouls is greatly correlated to how much a rookie is on the floor.
The tables above use the data of NBA rookies who came from NCAA Division I programs between the 2006-07 and 2010-11 seasons and got on the court in at least ten games in their first year out of college. The data shows that collegiate foul rates against top opposition is strongly connected to NBA rookie season foul rates, and that the kind of foul rate Singleton had at Florida State have largely been limited in their minutes as NBA rookies. This is not to say there are not exceptions. DeMarcus Cousins was fourth among all qualified rookies in foul rate this past season and still played 28.5 minutes per game, though his specialty was not his defense. Part of being a defensive stopper in the NBA is guarding an elite scorer without fouling, and his rate of committing fouls is significantly greater than that of recent defensive aces Ronnie Brewer, Sam Young, and Luc Richard Mbah a Moute in the year’s they entered the draft as well as for their NBA careers. It is also higher than that of other players whose defensive abilities are their calling cards, DeAndre Liggins, David Lighty, Gilbert Brown, Tyler Honeycutt, and Damian Saunders.
* Part of the intrigue surrounding Malcolm Lee comes from the fact that Russell Westbrook and Jrue Holiday have found a good amount of NBA success after not having standout careers at UCLA. Like Westbrook and Holiday, Lee is a big and athletic for the point guard position, but the difference that needs to be pointed out is that Russell Westbrook and Jrue Holiday were in less prominent roles because of Darren Collison. From an efficiency standpoint, Westbrook and Holiday performed well for underclassmen. Both had assist-to-turnover rates of better than 1.5:1 and shot around 50 percent on attempts taken inside the arc against the RPI Top 100. Also, despite the presence of Collison both averaged better than 3.5 assists per game. Lee has had the opportunity to be in charge of the UCLA offense, yet the results have not been there. He has taken more three pointers than the previous guards despite the fact he is not particularly good at them, and his assist totals this year were down this season from his high mark of 3.1 as a sophomore. The other marked difference is that UCLA is 13-20 over the past two years against opponents who finished in the RPI Top 100. They were 22-3 in Wesbrook’s sophomore season and 11-9 in Holiday’s only year.
* Last year, Avery Bradley declared for the draft following a freshman season at Texas that was underwhelming given his prep accolades. Not only did he struggle to run the point, but also had trouble scoring against better opposition. He ended up getting taken in the middle of the first round off his prep laurels and strong workouts, but proved incapable of contributing as a rookie. This year, there are two more players with the same M.O. have made the same decision, Josh Selby and Cory Joseph. The freshmen seasons of Joseph and Bradley are especially similar given that they both played at Texas and had similar splits in their performance against Top 100 competition and then lesser opponents.
|Cory Joseph, Texas||Fr.||6' 3 1/4"||186 lbs.|
|2010-11 Total (28-8)||36||32.4||3.8||8.9||42.2||1.4||3.5||41.3||2.3||5.4||42.9||1.4||2.0||69.9||0.6||2.9||3.6||3.0||1.5||1.0||0.3||1.8||10.4|
|2010-11 Non-Conference Reg (12-3)||15||31.8||4.0||9.4||42.6||1.5||3.9||39.7||2.5||5.5||44.6||1.6||2.4||66.7||0.6||3.2||3.8||3.2||1.7||1.0||0.3||2.1||11.1|
|2010-11 Big Twelve (15-4)||19||32.6||3.7||8.7||42.2||1.4||3.4||41.5||2.3||5.3||42.6||1.3||1.7||72.7||0.7||2.8||3.6||2.8||1.3||1.1||0.3||1.6||10.1|
|2010-11 RPI 100 (13-8)||21||34.1||3.8||9.8||38.5||1.1||3.0||35.9||2.7||6.7||39.7||1.4||2.1||66.7||0.7||2.5||3.2||2.7||1.2||1.0||0.4||1.8||10.0|
|2010-11 Outside RPI 100 (15-0)||15||29.9||3.8||7.8||48.7||1.9||4.1||46.8||1.9||3.7||50.9||1.4||1.9||75.0||0.5||3.5||4.1||3.5||1.9||1.1||0.2||1.8||10.9|
|Avery Bradley, Texas||Fr.||6' 3 1/4"||180 lbs.|
|2009-10 Total (24-10)||34||29.5||4.7||10.8||43.2||1.2||3.3||37.5||2.3||5.4||42.9||1.1||1.9||54.5||1.0||1.9||2.9||2.1||1.5||1.3||0.5||2.4||11.6|
|2010-11 Non-Conference Reg (14-1)||15||25.6||4.3||9.9||43.6||1.3||2.5||51.4||2.5||5.5||44.6||1.3||2.5||51.4||0.7||1.2||1.9||2.5||1.4||1.6||0.6||2.2||11.0|
|2010-11 Big Twelve (10-8)||18||32.1||5.0||11.3||44.1||1.4||3.6||40.6||2.3||5.3||42.6||0.9||1.5||59.3||1.3||2.5||3.8||1.8||1.7||0.9||0.4||2.6||12.3|
|2010-11 RPI 100 (8-9)||17||32.8||4.1||11.4||36.3||0.9||3.6||24.6||2.7||6.7||39.7||1.1||1.8||61.3||1.2||2.2||3.4||1.7||1.9||1.2||0.6||2.5||10.2|
|2010-11 Outside RPI 100 (16-1)||17||26.2||5.2||10.3||50.9||1.6||3.0||52.9||1.9||3.7||50.9||1.0||2.1||48.6||0.7||1.6||2.4||2.5||1.2||1.4||0.4||2.2||13.1|
* Selby is similar, however, in that his underwhelming college season seems to be getting dismissed. Of the sixty players currently in the nbadraft.net mock draft, Selby had the worst turnover rate – 4.3 per 40 minutes - and field goal percentage – 34.8 percent – of that group in games against opponents who finished in the RPI Top 100. While Selby was forced to sit out the first nine games of the season, and there is a theoretical case that missing the start of the season could have been an explanation for his struggling to adjust, the fact of the matter is that he started out with a number of solid scoring outings. If the ankle injury he suffered in late January zapped him of his quickness, then what was the explanation for his performances against Michigan, Nebraska, Baylor, and Texas?
* One of the great conversations when looking at any draft is finding a sleeper, the next Wesley Matthews. It is an enjoyable topic, but it is quite difficult to do so when one considers the set of circumstances that resulted in Wesley Matthews making the Utah Jazz. If not for Matt Harpring’s late decision to retire as well as preseason injuries to Kyle Korver and C.J. Miles, Utah would not have been compelled to carry another wing on their roster. Wesley Matthews would have not gotten the chance at that point to prove himself the player he is today. Players who get selected to successful teams that return almost everyone in their rotation are going to find themselves waiting for an opportunity even if they are ready to contribute. Coaches also have their own tendencies with regard to rookies. The last time Larry Brown had a rookie play more than 30 minutes per game was David Robinson in 1989-90. Phil Jackson, whose teams consistently had more overall talent than Brown, played all of two rookies – Stacey King and pseudo-rookie Toni Kukoc - more than 20 minutes per game in his entire coaching career. It is difficult to maneuver through all these variables before the draft, but while situation plays an undeniable there have been trends among 2nd round picks and undrafted players who have found success. Among all rookies this past year, Landry Fields had the lowest foul rate (1.7 per 40 minutes). He also had the second lowest turnover rate (1.7 per 40 minutes) among qualified rookies, only Greg Monroe was better (1.5 per 40 minutes). The lowest foul rate among all rookies was Landry Fields at 1.8 per 40 minutes. The year prior, high performing second round picks/undrafted players Chase Budinger (2.3 PFs and 1.3 TOs per 40 minutes), Wesley Matthews (3.0 PFs and 1.8 TOs per 40 minutes), and Jonas Jerebko (4.1 PFs and 1.5 TOs per 40 minutes) were three of the best six rookies in turnover rate, and among the nine best in foul rate. In 2008-09, Luc Richard Mbah a Moute was not only a tremendous defensive wing and offensive rebounder from day one, but also seventh among rookies in turnover rate (1.9 TOs per 40 minutes) and had a respectable foul rate of 3.9 per 40 minutes. Essentially, the trend is that players who do not make mistakes are the best candidates. It might not seem like much, but the ability to make plays is what will get a player’s feet into the team facility, but the ability to make plays while avoiding mistakes (turnovers, fouls, and obviously bad/missed shots) will earn the trust of the coach and allow them to stay on the floor.
I hope the Wolves have have this info so they don't f$#%@k up another draft.
This is seriously great stuff.
That is some seriously in depth research, the stuff you will find in the war rooms of GM's and their scouting team.
I did notice that nbadraftnet's mock leaves Charles Jenkins out completely - undrafted. I presume this is because of the allegedly weak competition Hofstra faced, but it is hard to believe he is not in the top 60 prospects. Your observation about the success of players who avoid mistakes should also work in Jenkins' favor, as he is very efficient in all types of shooting and has a good assist/turnover ratio.
I also see you rightly cast doubt on the notion that rebounding translates well from college to NBA. If Kenneth Faried turns out to be a good, not great, NBA rebounder, how will he compensate? His offense is awful, he's a lousy FT shooter and he measured out an inch shorter than expected. Do you risk a first-round pick on Reggie Evans?
Once again, my compliments for the fine number-crunching.
It's mentioned that there's a high correlation between personal fouls and minutes played. With R^2= 0.178 and 0.443, respectively; you can say there is "certain" correlation. Of course, it's noticeable, but it's not THAT high. I'm sorry if I'm being too harsh on this =)
Besides, you all know statistics can be a interesting way of drawing conclusions. You just should be aware of the scope of the research and also distinguish between objective and subjective information. What I'm trying to say is that analysis can differ so much from one person to another. Another author could have drawn the opposite conclusions from these statistics.
"the propensity to commit fouls is greatly correlated to how much a rookie is on the floor."
"Just be careful with the conclusions, It's mentioned that there's a high correlation between personal fouls and minutes played. With R^2= 0.178 and 0.443, respectively; you can say there is "certain" correlation."
The statement was that players who have a high tendency to commit fouls play fewer minutes, not that players who avoid fouling necessarily get minutes simply because of the avoidance.
great article. Now this is something impressive.