The NBA’s assistant coaches have made their choices for the NBA’s Rookie Challenge. And missing from the roster – as Henry Abbott at TrueHoop noted yesterday — is Ty Lawson.
The argument that Lawson is a very good player has been made by many people (John Hollinger, David Thorpe, etc…). But apparently these arguments didn’t impress the assistant coaches.
In an attempt to understand what did impress the assistant coaches, let’s first examine the selections for the sophomore team. And that examination begins with the following list [PPG=points per game, WP48 = Wins Produced per 48 minutes, numbers after 41 games played]:
- Derrick Rose: 19.4 PPG, 0.053 WP48
- Brook Lopez: 19.0 PPG, 0.141 WP48
- O.J. Mayo: 18.4 PPG, 0.100 WP48
- Eric Gordon: 17.0 PPG, 0.085 WP48
- Michael Beasley: 16.1 PPG, 0.061 WP48
- Russell Westbrook: 16.1 PPG, 0.109 WP48
- Kevin Love: 15.2 PPG, 0.431 WP48
- Marc Gasol: 14.8 PPG, 0.199 WP48
- Danilo Gallinari: 14.8 PPG, 0.117 WP48
The above is a list of the nine leading scorers from the sophomore class. And all nine were selected by the assistant coaches to the Rookie Challenge roster. Yes, it appears that scoring was the dominant factor coaches considered in making these selections.
When we turn to the rookies it’s a similar story, although the coaches faced a significant constraint. Here are the top six scorers in the rookie class:
- Tyreke Evans: 20.9 PPG, 0.145 WP48
- Brandon Jennings: 17.9 PPG, 0.063 WP48
- Jonny Flynn: 13.9 PPG, -0.051 WP48
- Stephen Curry: 13.4 PPG, 0.099 WP48
- Omri Casspi: 12.6 PPG, 0.145 WP48
- James Harden: 9.8 PPG, 0.136 WP48
Each of these players was selected by the assistant coaches. The next two players on the scoring list – Ty Lawson [9.4 PPG, 0.173 WP48] and Marcus Thornton [9.3 PPG, 0.022 WP48] — are both guards. Of the top six scoring rookies, though, five are also guards. Consequently the assistant coaches ran into a significant problem. The rookie challenge roster cannot consist of just the nine “best” (however that term is defined) players. The players selected have to play in an actual game. That means the team needs someone for each position on the court. And consequently, Lawson and Thornton – if points per game is the deciding factor – had to be skipped over.
Moving further down the scoring list we see the following three names.
- Tyler Hansbrough: 8.5 PPG, 0.018 WP48
- Jonas Jerebko: 8.4 PPG, 0.116 WP48
- Taj Gibson: 8.3 PPG, 0.065 WP48
These three players are quite close in scoring per game. But Hansbrough missed 12 of the first 41 games and hence the assistant coaches turned to Jerebko and Gibson.
Of the eight players noted so far, none have played significant time at center. The top scorer at that position only ranks 16th among all rookies in points per game. And that player is DeJuan Blair [7.1 PPG, 0.273 WP48]. One could argue that Blair is really a power forward (although he has played some center this year). But if you move past Blair, the next center on the list is David Anderson, and he only averages 6.0 points per game (and after Anderson you have Hasheem Thabeet at 2.8 points per game).
A Familiar Story
So it appears that we have a familiar story. Scoring dominates the evaluation of players in the NBA. Obviously scoring can increase if a player improves his shooting efficiency. But scoring also increases with shot attempts, and hence players have an incentive to shoot as much as their coach will allow.
Many examples of this observation have been provided before. The latest is the case of Jonny Flynn. Al Jefferson is generally considered the “star” of the Minnesota Timberwolves (if a 9-38 team can have a star). And Flynn — the team’s point guard – does make sure Jefferson gets his shots. But after Jefferson, the player with the second most shot attempts is Flynn. Flynn, though, is well below average with respect to shooting efficiency. So why does he so often call his own number? Because he probably knows that the more an NBA player scores, the more accolades an NBA player will receive.
And the assistant coaches choices for the Rookie Challenge are not the only evidence in support of this proposition. You see the same story when you look at free agent salaries, voting for the All-Rookie team (by NBA coaches), where a player is selected in the draft, and the coaches’ allocation of minutes (we discuss all of this in our next book).
Other Snubs
If we move past scoring to WP48, what other snubs do we observe? Here are the top five forwards and centers among the sophomores in WP48 (minimum 15 minute per game, 20 games played):
- Kevin Love: 0.431 WP48
- Greg Oden: 0.316 WP48
- Marc Gasol: 0.199 WP48
- Ersan Ilyasova: 0.146 WP48
- Ryan Anderson: 0.146 WP48
And here are the top five sophomore guards:
- Rudy Fernandez: 0.165 WP48
- Russell Westbrook: 0.109 WP48
- Brandon Rush: 0.101 WP48
- O.J. Mayo: 0.100 WP48
- Eric Gordon: 0.085 WP48
Greg Oden is hurt, so he can’t play. So it appears the snubs – if WP48 is the measuring stick – are Ilyasova, Anderson (maybe if the coaches wanted five frontcourt players), Fernandez, and Rush. Despite these omissions, the average WP48 on the sophomore roster is 0.144. So the sophomore team – as it’s assembled – is pretty good.
Turning to the rookies, here are the top five forwards and centers (same minimums as above):
- DeJuan Blair: 0.273 WP48
- Serge Ibaka: 0.145 WP48
- Omri Casspi: 0.145 WP48
- Jonas Jerebko: 0.116 WP48
- Chase Budinger: 0.090 WP48
And the top five rookie guards:
- Ty Lawson: 0.173 WP48
- Tyreke Evans: 0.145 WP48
- James Harden: 0.136 WP48
- Stephen Curry: 0.099 WP48
- Brandon Jennings: 0.063 WP48
So the “snubs” for the rookies are Ibaka, Budinger (maybe if the coaches wanted five front court players) and Lawson. As assembled, the rookies have an average WP48 of 0.110. Replacing Flynn with Lawson boosts that average to 0.135. So adding Lawson would have made this game quite competitive.
As it stands, though, the sophomores look to be the better team (once again). This is not surprising, since second-year players are generally better than rookies. This year, though, the gap could have been mostly closed with one different choice. Unfortunately, the choices Flynn has made (he gets to shoot) means Lawson gets to stay home.
– DJ
The WoW Journal Comments Policy
Our research on the NBA was summarized HERE.
The Technical Notes at wagesofwins.com provides substantially more information on the published research behind Wins Produced and Win Score
Wins Produced, Win Score, and PAWSmin are also discussed in the following posts:
Simple Models of Player Performance
What Wins Produced Says and What It Does Not Say
Introducing PAWSmin — and a Defense of Box Score Statistics
Finally, A Guide to Evaluating Models contains useful hints on how to interpret and evaluate statistical models.
mrparker
January 28, 2010
This being especially true since exhibitions seem to be played mostly outside of the paint.
Italian Stallion
January 28, 2010
One thing I believe absolutely goes into both all star choices and rookie of the year choices is who is going to be better over the long haul.
IMO, the selection process contains an element that’s similar to the draft where if two players are similar statistically, but one has vastly superior athleticism, physical attributes, and/or is younger he will get the nod.
In the draft, the consideration is long term value.
In the all star and ROY selections, the consideration is marketing future mega stars to the public as early as possible.
So while a stats guy would argue that guys like Lawson are superior to some of those that were selected ahead of them (something I would agree with), the selection committee probably agrees with that but sees guys like Flynn and others as eventually becoming the better player. So just as with the draft, they take Flynn higher.
It’s the kind of politics we see elsewhere in life where merit is only partially the determining factor for who wins (IMHO a horrible thing in any circumstance).
Daniel
January 28, 2010
Players that perform as Flynn did in college and their rookie seasons almost NEVER become good players. If he takes 4 more shots per game next season, he’ll be regarded as one of the top PGs in the game (Rose is similar), which is ridiculous because Flynn doesn’t do anything well.
The T-wolves picked three point guards in the first round, and somehow ended up with the only dud of the three playing for them this season.
Lior
January 28, 2010
Well, All-star games are about showmanship, not about efficiency. There is a correlation between flashiness, a license to shoot, and scoring many points per game. For the fans, the game is 100% about entertainment and 0% about winning. Moreover, most fans are not entertained by setting good picks and blocking out for rebounds. They are entertained by high-flying dunks and acrobatic layups.
The all-star game is thus about showcasing the most entertaining players, not the most efficient ones. Rodman was for a while the best power forward in the league; he’s a serious contender for the all-time power forward position. Nevertheless but has appeared in only 2 all-star games. His ratio of 5 championships to 2 all-star games is quite proper given his style of play.
workingsforsuckers
January 28, 2010
How about a campaign to convince the league to let hoops stat geeks pick the rookie team to go up against a media selected sophomore team?
The game would be less TV friendly but it would be great for the blog-o-sphere.
Statement
January 28, 2010
Berri,
What was your prediction for the Raptors for this year. If memory serves it was 39 wins. Well they are 4-0 in the last 4 with a huge win over the Lakers.
Long and short of it is….your model sucks, Rosenbaum was right.
Matt
January 28, 2010
Do you really think Flynn takes so many shots because he’s trying to get more accolades? I think there are two reasons a) as inefficient as he is, he is still the best perimeter option on his team and b) he’s very good at creating good looks, he’s just not very good at cashing in on them yet.
That said, the argument for Lawson over Flynn in this game is valid. I just don’t think Flynn is a particularly selfish player when I watch him, just a confused player.
KMils
January 28, 2010
Statement
What was your prediction for the Raptors for this year. If memory serves it was 39 wins. Well they are 4-0 in the last 4 with a huge win over the Lakers.
Long and short of it is….your model sucks.
————–
Raptors are on pace for 43 wins.
Flynn isn’t shooting because he wants accolades; he’s shooting because that’s what he’s good at – shooting. Not a great distributor (mediocre at best), not a great defender (5’11”); what else is he supposed to do? He dribbles too much and he shoots too much.
brian
January 28, 2010
I agree that Flynn isn’t really being selfish, but that he just kind of sucks.
user Statement: the Raptors currently have a -1.0 point difference per game. This corresponds to a sub-42 win season. So I don’t see how 39 wins is a bad prediction. Yes, the Raptors currently have a winning record, based on looking at the outcomes of their 47 games. But the outcome of each of these 47 data points is only ever going to be a W or an L–a lot of data is lost by simply looking at the overall w/l record. Point difference per possession is more telling because a) more data points are used (every possession utilized throughout the season) and b) the goal of every single basketball game is to end the game with more points than the other team. What better way to predict the likelihood of achieving the b) goal than by looking at a stat that tells you just that?
Lior
January 28, 2010
Just for kicks I computed the probability that a team with a given winning percentage will have a 4-0 streak at some point during the season. It turns out that the naive guess is right: a team that wins 1/3 of its games (and hence expects one 4-game winning streak) has almost a 50-50 change of having a 4-game winning streak at some point during the season.
Since even weak teams will quite often have 4-game winning streaks, just knowing that the team had such a streak at some point during the season tells us little.
Lior
January 28, 2010
[by “expects one streak” I meant that the expectation of the number of games at which a 4-0 or better streak begins is equal to one; this happens when the winning probability is the inverse of he fourth root of 82]
Man of Steele
January 29, 2010
One thing I’ve noticed several times lately is that players who come out of college early seem to bend the developmental curve. Kevin Durant has been much discussed on this blog, but I don’t think anyone has pointed out that a lot of his struggles in his rookie year and subsequent improvement wee to due to his youth; he would just now be a senior at Texas. When he was 19, it was not necessarily a terrible sign that he was a little below rookie average, since the rookie average over the years includes many nore 22 year olds than 18 or 19 year olds. I remember seeing in the first book that LeBron James was .099 his rookie year, which was technically a tad below average but was really quite good for an 18 year old. Anyway, O.J. Mayo got me thinking about this, because he wasn’t any good last year but has become average this year. I think he’s like 21, so he may be a good player yet. Perhpas Dr. Berri could create a post examining various developmental patterns among young players who enter the league at different ages.
Myabe the title of the post could be “The Ages of Wins”
dberri
January 29, 2010
“The Ages of Wins”
Now that is a great title.
Italian Stallion
January 29, 2010
The Raptors have been playing a lot better lately after a slow start. It’s more than just a random 4 game winning streak. Someone might want to check the point differential over the first part of the season relative to last few weeks. They have a few new pieces that are now starting to mesh. They are not in the top tier because they still lack some rebounding and other things, but they are good.
Zach K
January 30, 2010
Someone give Statement his own blog. That guy is a genius.
Steave Thomason
February 15, 2010
Rather nice blog you’ve got here. Thanks for it. I like such themes and anything that is connected to this matter. I definitely want to read more on that blog soon.
Truly yours