Imagine you are a resident of New York who wishes to pledge allegiance to an NBA team. Naturally your thoughts turn to the home town Knicks. Unfortunately, the Knicks don’t exactly inspire allegiance. Currently their record stands at 13-28, one of the worst marks in the league. And if you look over their recent history you see a team that hasn’t had a winning record since 2000-01.
The Brooklyn Nets
Unfortunately, unlike baseball, football, and hockey fans in the Big Apple, basketball fans who wish to follow a hometown team are stuck with the woeful Knicks. Soon, though, this might all change. The Nets look to be departing New Jersey for Brooklyn, and in 2010 (maybe) the Brooklyn Nets – or New York Nets – could be challenging the Knicks for supremacy in New York (assuming this whole move works out).
Assuming the Nets ultimately make this move, they look to be an attractive alternative to the Knicks. The Nets have not had a losing season since 2000-01. Jason Kidd, the team’s point guard, led the NBA in Wins Produced last season. Although the Nets are currently below 0.500, their 18-24 mark clearly tops the Knicks this season. So clearly, if you could choose, the Nets would be your choice (and if you are lucky enough to also root for the Mets and Jets, you can take great pride in knowing that the names of all your sports teams in 2010 will rhyme).
The Knicks vs. Nets in 2007-08
Before any Knicks fans consider switching allegiance (as if that were possible, a point I will make in a moment), they might want to consider a few more numbers. When we look at efficiency differential – offensive efficiency minus defensive efficiency – the Nets and Knicks this season don’t look much different. At the midpoint of the 2007-08 season, the Nets have a differential of -6.5. The Knicks differential after 41 games is -6.7. Each mark is consistent with a team that wins 24 games over a full season. This means that although the Nets currently have a better won-loss mark, the Nets have not performed significantly better this season.
When we turn to Wins Produced – which simply connects efficiency differential to the players – we see a similar story.
Table One: The Nets and Knicks at the midpoint of 2007-08
From Table One we see that the summation of Wins Produced for each team is basically the same. How we get to each team’s summation, though, is not quite identical.
Summing the Nets
The Nets in 2007-08 are once again led by Jason Kidd. As noted last summer, since the days when this team went to the Finals, Kidd’s supporting cast has consistently declined. And in 2007-08, the decline has gone a bit further. Vince Carter, Sean Williams, and Jason Boone are the only other above average performers in the rotation. Richard Jefferson, who was generally an above average player in the past, has dropped off considerably the past two years (primarily because of injury).
The Nets don’t have many above-average players. But it’s the quality – not the quantity — of below average performers that’s really causing problems. If all the Nets after Kidd, Carter, S. Williams, Boone, and Jefferson offered zero wins, the Nets would be on pace to win 42 games. Unfortunately, every other player on the roster currently has a WP48 [Win Produced per 48 minutes] in the negative range. These nine negative performers basically act as a huge anchor dragging down the overall productivity of this team.
The good news is that by 2010 (according to HoopsHype), all of these negative players (except Marcus Williams), will no longer be under contract. Unfortunately, the same can also be said for Jason Kidd. Kidd’s contract expires after next season. And given his age, it’s more than likely his productivity has to drop off soon, making signing him to a new contract a risky proposition.
Without Kidd, though, where will the Nets get wins in 2010? The players who might be under contract (depending on various options) in 2010 include Carter, Jefferson, S. Williams, Boone, and M. Williams. Currently these players are on pace to produce 20 wins in 2007-08. For the Nets to be contenders, it will have to find more wins someplace. Given the ability of the Nets to find players to complement Kidd these past few years, though, one can seriously question the ability of the Nets to find production to replace Kidd.
The Future of the Knicks
Turning to the Knicks we see a somewhat different story. The Knicks will probably be without the services of Stephon Marbury by 2010 (it’s possible they are without him now). But much of the remaining roster is still under contract in 2010 (again according to HoopsHype).
The big decision between now and then is the re-signing of David Lee. Last year Lee led the Knicks with 13.6 Wins Produced and a 0.378 WP48. This year his minutes, per-minute productivity, and Wins Produced have all declined. One should note that there is a link between declining minutes and per-minute performance. And I think it’s reasonable to think (although I have not studied this) that increasing the variability in minutes also adversely impacts performance.
Last year, before Lee was hurt in February, he was averaging 33 minutes a contest and only once had he played fewer than 20 minutes. This year Lee is averaging 26 minutes a game, and already he has played fewer than 20 minutes seven times. Given the yo-yo nature of his minutes, I think we shouldn’t be surprised to see Lee’s productivity slip.
Although Lee has slipped, he’s still the most productive player on this roster. And if the Knicks can re-sign Lee, there is some hope for this team.
So here is the choice facing our hypothetical fan.
Should you pledge allegiance to a team that is sort of on top now, but looks to be in clear decline once Kidd departs the scene?
Or should you turn to team that could build around a very productive player, but first has to realize that Lee is indeed the most productive player on the roster?
Or should you just move someplace else?
Can You Switch Allegiance?
An unanswered question in all of this analysis is whether it’s even possible to choose between the Nets and Knicks. Certainly if you are not aligned with either team, you could choose one or the other. But I think New York fans are just like me. I grew up in Detroit, and although I have not lived in the Motor City for 27 years, I still follow the Lions, Tigers, and Pistons. And although the Lions consistently disappoint, I am unable to shake my addiction.
My sense is the same is true for fans of the Knicks. Even if the Brooklyn Nets were a better team, Knicks fans cannot shake their addiction. Consequently, the only people the Nets will be able to win over in Brooklyn are those that are currently non-affiliated. These would include primarily children (who haven’t formed an addiction yet) and recent transplants to the Big Apple. On the surface this is not a very big group.
Let me close by noting that there is a literature that discusses sports allegiance in terms of addiction. Perhaps I can get one of the writers in this area to offer a guest post in the future.
– DJ
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.
Wages of Wins Journal 
antonio
January 25, 2008
This is what I think is wrong with this statistic. Jason Kidd is certainly not the best player in the league. While he grabs an extremely high amount of rebounds, it is due more to his other players incompetence than his own greatness. I don’t believe that as he has gotten older his rebounding skills have increased. He is a clear case of a player, in my opinion, who would not add the 8.7 rebounds he is averaging to another team. While I can’t prove this, I don’t really think there are many other teams where he would average these numbers. Also, his turnovers and inability to shoot the ball are just killing the team. Boone and Williams have almost no offensive game, and carter and Jefferson need help and Kidd is just unable to provide it. While his rebounding is great, what the team needs out of their point guard is more of a scorer. Sadly, Kidd does not provide that
Jason
January 25, 2008
Kidd’s rebound numbers are probably helped by the rest of his team being lousy on the boards. However, it seems that there’s a feeling that the *only* reason that he grabs those boards is because of his lousy teammates. Historically, Kidd has always been a very, very, very good rebounder. He’s averaging about two rebounds and a half above his career average per 48. Since there’s been an upward trend in his rebounding for much of his career, I don’t think it’s unreasonable to believe that he’s better at this skill than he was early in his career. So how many rebounds would he be averaging with a more competent front line? It’s probably not *that* much lower.
For what it’s worth, the Nets are out-rebounding their opponents. Since Kidd is their leading rebounder and they’re winning on the glass (albeit by a small margin) it’s peculiar to suggest that Kidd’s ability to make up for a lousy front court somehow diminishes the accomplishment. Apparently, they *do* need rebounding from their point guard since the rest of the team isn’t providing it. Perhaps they’d be better off if they *also* got scoring, but they do need the rebounding and would suffer without it.
If Kidd was not there and an average point guard was inserted into the lineup, would someone else on the team be able to make up for the lousy rebounding by the rest of the team? If no, then he really *is* that valuable to the team as those rebounds really are important.
antonio
January 25, 2008
I recognize Kidd is a great rebounder, but I just think his numbers are certainly inflated by some of the terrible rebounding from their front court. Clearly he he has lost much athleticism, which has led to more turnovers, less steals, and less points. Thats why I question him improving his rebounding. Rebounding does take a lot of athleticism. If he is going down an other athletic indicators, why is his rebounding improving tremendously? I think Kidd is a great player, but this team was much better when he was able to drive to the basket and teams had to at least respect him from 3 (not give him WIDE open threes.) Obviously I recognize in the past other players have performed differently, but I think Kidd has certainly contributed to the poor record this year more than usual. Now he is unable to get to the basket, nor can he make any 3s, and its just killing them on offense. On other teams I think he would be much more valuable, but I don’t think that if this team were to replace Kidd with an average point guard they would just fall apart. I guess we will have to wait in two years for when he leaves.
Kent
January 25, 2008
“One should note that there is a link between declining minutes and per-minute performance. ”
Doesn’t this show selection biases? If a guy is good the coach will play him more. That coaching decision alone would motivate a correlation between minutes played and per minute performance in the data.
Todd
January 25, 2008
I have been following this website for some time now and I have read the Wages of Wins. I truly appreciate the use of statistical analysis and economic tools to demystify some of the greatest illusions of conventional wisdom. I thought this post would be as good a post as any to talk about the problem of overstating the importance of rebounding in the wages of wins metrics, as Jason Kidd is certainly a player whose win score benefits from the inflated importance placed on defensive rebounding.
The first, and biggest, issue is that the rebounder is getting credit in their win score for more than just their rebound. In creating the Win score metric, regression analysis was applied to all of the box score statistics to determine how much each contributed to a win. A defensive rebound and a steal both are counted equally in the win score because both statistics perfectly correlate with a change of possession in which your opponent fails to score. It is important to note that what is really being measured by each rebound in this analysis is not just the rebound, but the rebound and the missed field goal that preceeds it. Unfortunately there is no box score statistic that shows what defensive player most contributed to an opposing player’s missed shot. Since regression analysis showed that a blocked shot (which represents a forced missed field goal) is worth ½ a point and a defensive rebound is worth a full point in the Win Score metric, it would seem reasonable to assume that all defensive rebounds are actually a ½ point defensive stop and a ½ point rebound. So, lets pretend Richard Jefferson plays some really strong defense forcing his opponent to take a hard shot and miss, then Jason Kidd picks up the rebound. Jefferson’s contribution to ending the opponent’s possession has been given to Kidd.
The other issue is that once a shot is missed, it can either become an offensive or defensive rebound, so there are only as many rebounds to have as there are missed shots. In addition, close to 75% of all rebounds are defensive rebounds. Right now San Antonio tops the league, only allowing 9.3 offensive rebounds by their opponents each game, and Phoenix is last in the league, allowing 14.1. New Jersey gives up 10 offensive boards a night. This tells you that going from the worst defensive rebounding team to the best involves taking 5 boards a game from your opponent. All of the very best rebounders at each position, like Jason Kidd or Josh Howard, are going to take more rebounds from the other players on the floor with them than they will from players on the opposing team. Jason Kidd gets 6.5 more rebounds per 48 minutes than the average point guard. Clearly, replacing him with an average point guard will not result in the opponent grabbing 6.5 more offensive rebounds, making New Jersey by far the worst rebounding team in the league. Likewise, players like Eddie Curry’s futility on the boards is overrated in the metric because for every 4 rebounds he fails to grab, David Lee and Zach Randolph probably pick up 3.
Animal
January 25, 2008
Todd, very good post.
dberri
January 25, 2008
Todd,
What you said has been said already (many times). And it has been responded to already (many times). So thanks for the comment, but I think you can find a response on any number of posts (search for Jason).
Animal
January 25, 2008
Todd, Here’s a reply to your post …
(from Jason)
The issue isn’t just whether or not the formula is more complex, it’s whether it is more useful. Your division of a rebound and credit for a missed shot makes an assumption that this division will more accurately measure the individual contributions of players. It may be true; it will not change the team score, but if it doesn’t explain something else better, then the move seems more one made out of a sense of equity rather than one derived from necessity.
What would it need to explain to justify such a change? A change that could modify player ratings would have to better explain the individual contributions of a player such that if the player was removed from the team and placed on another team, his performance *and the performance of the new team* would be more accurately gauged based on the individual’s rating.
While it may appear more fair, it is an empirical question as to whether or not it makes the formula more *accurate* in terms of its predictive power. If you are rewarding players for rebounds they did not gather or giving less than full credit for a rebound that someone did pull down, you have to justify that making this adjustment will better explain future returns.
It still seems like the arguments against weighting a rebound as a rebound are philosophical rather than empirical.
Animal
January 25, 2008
Todd, more from Jason on the subject …
The issue with dividing up a rebound (in any form) vs. giving the whole credit to the guy who gets it is whether or not we get better predictions from one form or another. At the team level, the totals are the same, so it should not have influence on team wins, but it will change how players on the team are ranked relative to one another.
Whether or not this change in rankings is superior or not depends on how well past results predict future results. When I looked at it, dividing up rebounds in the .6vs4x.1 allocation was that this did not improve predictive ability and in fact, made things more variable. It was *not* more accurate at the level of predicting future results for the team based on the performance of players on the team.
It is possible that some other division of credit would be better (and entirely possible that the division of credit varies from team to team and with different player combinations). We could toss out various numbers, but there’s reason to believe that this isn’t necessary and isn’t likely to improve things given the limits of the data collection. There is a strong, strong, strong correlation between a player’s rebounds/min from season to season and this correlation is not diminished by changing teams and teammates at all. This suggests strongly that the biggest single influence on how a player rebounds is that player himself. The influence of teammates may be there–it probably is–but it’s just not variable enough to need to be accounted for in the metric.
Other players may contribute, but if the contributions beyond grabbing the rebound are rather consistent (e.g. there are very few superior boxer-outers) then giving the credit to the rebounder in whole. All players, for the most part, seem to receive the same benefit from teammates.
Please see the comments thread of https://dberri.wordpress.com/2008/01/06/two-from-the-sun-on-the-knicks/ for even more discussion.
Animal
January 25, 2008
Kent writes …
“One should note that there is a link between declining minutes and per-minute performance. ”
Doesn’t this show selection biases? If a guy is good the coach will play him more. That coaching decision alone would motivate a correlation between minutes played and per minute performance in the data.
This I have no answer for. Anyone?
Jason
January 25, 2008
I don’t remember the details, but I believe Dave mentioned that he had controlled for the notion that better players would receive more minutes because they were playing better. Off the top of my head, I don’t know how one controls for this, but I suspect it’s not an intractable problem.
Brendan
January 26, 2008
Leaving aside the states questions-
For whatever it’s worth as an anecdote, I’m a Brooklynite who’s followed both the Nets and Knicks equally for about 16 or 17 years now. The Nets moving to the city will definitely have a strong impact on my financial expenditures related to basketball if not my overall attention paid, since as a fan I would love to attend Nets games but hate the travel, and can’t stand the thought of contributing money to the current owners of the Knicks. So financially, the Nets moving here actually would increase my total hoops-related outlay.
Am I an odd example, or fairly common? Hard to say. I do suspect there’s a strong niche in this city for a “people’s team”, given the general glitzy corporate “Manhattan” image of the Knicks.
As for the futre trend of both teams, I think it’s fairly obvious that both teams are going to be pretty bad for a while; they’re frequently barely watchable right now. The difference going forward is: while the Knicks probably have more actual talent on the roster, the Nets have a huge advantage in terms of overall managerial competence. As a fan, I’d rather bet on Rod Thorn and Kiki to rebuild than I would on Jim Dolan and Zeke- and that effects my interest and willingness to spend and attend as well.
antonio
January 26, 2008
” When I looked at it, dividing up rebounds in the .6vs4x.1 allocation was that this did not improve predictive ability and in fact, made things more variable. It was *not* more accurate at the level of predicting future results for the team based on the performance of players on the team.”
The one thing I don’t understand is why are future results different? I thought at the team level they were both the same, so they would predict the same results? How can one be more predictive than the other? I think I am just slow and not understanding something…
Kent
January 26, 2008
“The one thing I don’t understand is why are future results different? I thought at the team level they were both the same, so they would predict the same results? How can one be more predictive than the other? I think I am just slow and not understanding something…”
Antonio, if players switch teams than your allocation of credit at the player level will influence the predictive accuracy.
Evan
January 26, 2008
Since regression analysis showed that a blocked shot (which represents a forced missed field goal) is worth ½ a point and a defensive rebound is worth a full point in the Win Score metric, it would seem reasonable to assume that all defensive rebounds are actually a ½ point defensive stop and a ½ point rebound.
The difference between a blocked shot and a defensive rebound is that a blocked shot does not necessarily result in a change of possession. That’s a pretty huge difference.
Todd
January 27, 2008
I recognize a blocked shot does not result in a change in possession. What I was saying is that a block shot is the equivalent of an opponent’s missed shot, and since their is not a box score stat for opponent’s missed shots, all other forced misses get lumped in with the rebound in formulating the win score. Therefore, the credit that is given for a blocked shot should be roughly equivalent to the score for the forced miss portion of each rebound, and the remainder should aproximate the contribution from preventing the opponent’s offensive board.
Jason
January 27, 2008
Todd, so here is part of the problem of evaluating a partial method. Win Score rounds the effect of a blocked shot to 1/2. Wins produced calculates is a bit differently *and* adjusts (in the “teammate adjustment”) for the effect that blocks have on a given team.
This latter part is important. Some teams (either by coaching design or simply because of personnel) try to block shots, others do not. Some may argue that in addition to the blocks, other shots not blocked are ‘altered.’ Others argue that blocked shots can result in easy buckets when a defender moves out of position and fails to make the block. I suspect both are, to some degree true. But there is a measurable, albeit not huge effect of adding to the probability of a win. But since a block that results in a change of possession is elsewhere accounted for in terms of the change of possession, the amount credited to the shot blocker has to come off of some other part. In WP, essentially a small portion of other changes in possession are deducted from everyone and credited to the shot blocker. This results, more or less, in the apportionment of the miss between the shot blocker and rebounder (who prevents the opposition from maintaining possession).
The opponents missed shots *are* tracked in the box score. They’re tracked as defensive rebounds and opponent offensive rebounds. This is factored into wins-produced but not win-score.
Again, this is where getting the values from regression and seeing how much win probability changes with the statistical values is important and *independent* of the ‘logical reasoning’ of what is going on. A block, on average, raises the probability of a win.
We are up against the limits of what is recorded in the box score. If there was a reliable “forced miss” stat, things would likely be more accurate. There isn’t, however, and as such, the question is “are things accurate enough to be useful.” I think that this answer is quite clearly “yes”, though your mileage may vary.