If we look at the standings one team stands alone. The Dallas Mavericks are the only team to win 60 games this season and currently – despite the loss to the Phoenix Suns yesterday – are still six games better than the Suns.
When we look at offensive and defensive efficiency, though, the Mavericks do not stand alone. After 72 games (which does not include Sunday’s action), the Mavericks had scored 108.9 points per 100 possessions while surrendering only 100.2 points. So the Mavericks efficiency differential stood at 8.7. Meanwhile, the Suns had an efficiency differential of 7.2, scoring 110.9 points per 100 possessions while surrendering 103.7.
And then there are the San Antonio Spurs. Despite having the worst record of the three, the Spurs have the best efficiency differential. Per 100 possessions the Spurs score 106.1 points while only surrendering 96.4. This results in an efficiency differential of 9.7.
As I noted a few weeks ago, no other NBA team comes close to this trio. Certainly we should expect the Western Conference champion, and eventual NBA champion, to come from these three.
Evaluating the Best
Looking at these teams one might wonder: Why are these three teams so good?
The answer is in the players. And to evaluate the players, we turn to the Wins Produced of the players on each team.
Table One: The San Antonio Spurs after 72 games
Table Two: The Dallas Mavericks after 72 games
Table Three: The Phoenix Suns after 72 games
The Wins Produced algorithm simply takes what we know about a team’s offensive and defensive efficiency and applies it to the evaluation of players. As a result, if offensive and defensive efficiency indicates that a team should win 57 games, this is what Wins Produced will indicate as well. In other words, it should not surprise us to see that as offensive and defensive efficiency indicates, Wins Produced tells us that the Spurs are currently the “best” team in the NBA (a point I will weasel out of at the end).
Looking at these teams via the lens of Wins Produced, we see that each of these teams has a few outstanding players, and a collection of less than outstanding players. The Spurs are led by Tim Duncan, Manu Ginobili, Tony Parker, and Brent Barry. These four players have combined to produce 48.5 of the 57.2 wins this team’s players have produced. In other words, once we get past these four, the Spurs suddenly look like a fairly bad team.
The same story can be told of the Mavericks and Suns. The Mavericks are led by Dirk Nowitzki, Jason Terry, Josh Howard, and Erick Dampier. These four have produced 45.1 wins, or all but 10 of this team’s Wins Produced. On the Suns we see Steve Nash, Shawn Marion, and Amare Stoudemire. These three players have produced 43.1 wins, or again, all but 10 wins on the Suns.
Each of the eleven players I listed has produced at least 0.180 wins per 48 minutes played [WP48]. Given that an average WP48 is 0.100, all of these players are very good and the key to the success these teams have achieved.
Once you get past these eleven, though, there isn’t much depth. The Spurs have no other above average players in their regular rotation. The Suns have only Leandro Barbosa and Kurt Thomas while the Mavericks can only offer Devin Harris and DeSagana Diop. Although these players are technically above average, none are far from the 0.100 mark.
Given that each team has a limited supply of “great players”, as is always the case in basketball, injuries can play a significant role in each team’s drive for a title. Injuries to any of the major producers identified can quickly derail each team’s playoff train.
The Impact of Regular Season Time
Another issue we might consider is minutes played. Erick Eschker and I, in a paper we published last year (which was cited in The Wages of Wins), found that the more minutes you played per game in the regular season the worse your playoff performance. The effect is small, yet statistically significant.
Let me ignore the “small” part of the story for a moment and pretend that all that mattered was statistical significance. Both the Mavericks and Suns are playing some of their major stars more than 35 minutes per night. Meanwhile, the Spurs are coasting in the regular season, playing only three players more than thirty minutes per contest (and none more than the 34.4 minutes played by Duncan).
This means that the Spurs will enter the playoffs as a lower seed than either Dallas or Phoenix, but will also be more rested. And come May and June, the well-rested Spurs might be in better shape to go the distance in a long playoff series.
At least, that is our story if we only consider statistical significance. Often I am seeing people post regression results on line that only consider whether the estimated coefficient is different from zero and ignoring the size of the coefficient. If a decision-maker is to use regression analysis to make choices, that person needs to consider both the statistical significance and the size of the impact. With respect to the link between minutes played and playoff performance, there is a significant relationship, but I do not think the size of the effect is so large that the Spurs truly have an advantage.
So Who is the Best?
In the end, the primary determinant of playoff performance is regular season productivity. And when it comes to these three teams we see that each employs several players that were very productive in the regular season.
Of course we want to know which team is the best. Unfortunately, I do not think there is an answer to that question (even if I said there was earlier in this column). All three are very good. At least two of these teams will not win an NBA championship this season. Although we might think the team that does win is indeed the “best” in the NBA, we know that the sample of games in the playoffs is insufficient to answer that question.
All that being said, if I had to pick a “best” team in the NBA I would have to pick…. the Detroit Pistons. And this is my pick because I was born in Detroit, and for me they are always the best team (as I said, I weaseled out of the question).
– DJ
Wages of Wins Journal 
Jerry
April 2, 2007
I have a comment about something you posted a week or two back. You were discussing the book by the sabernomics guy, and how he found how much each player baseball player should have been paid. He did this by finding out how many wins that each produced and then finding out how much a win was worth. I think you have done something similar for basketball. The rationale for this process is that economics tells us that a worker should be paid a wage equal to his marginal product of labor, that is he should paid equal to what he produces for the company. This implies zero profit for the employer. However this is not accounting profit, it is economic profit. It takes into account opportunity cost. Because a team owner could make a a ton of money investing their 500 million dollars elsewhere then we should expect that they should make a profit on their team. This means that we cannot in fact say that a player should be paid equal to the amount of money they produce for the team because that would create a negative economic profit. So, I think the process that Bradbury used to determine those “fair” player salaries was incorrect. Tell me where I am wrong.
Evan
April 2, 2007
(as I said, I weaseled out of the question).
Well, you are an economist.
Rashad
April 3, 2007
One topic I would love to be addressed is your thoughts on why Dallas is performing so far above their projected wins (based on offensive and defensive efficiency) and why San Antonio is so far below. Is it just random? Historically, are there any qualities that are common to teams that beat/miss their projections? How do this seasons deviations from projections measure up historically? Basically, is it common what we’re seeing this season or not?
Zebra
April 3, 2007
Might be interesting to see a study of what Rashad asks about based on the 4 Factors offense and defense.
Daniel
April 4, 2007
I can’t comment on the Suns and Spurs as I haven’t watched nearly as many of their games as my team, which is the Mavericks. According to your table, the Mavericks’ fourth-best player is Erick Dampier? I don’t think so — Jerry Stackhouse is clearly the fourth-best Maverick. His minutes are limited in the regular season because at this point in his career he’s injury-prone and is unable to play a full 82 games. When the playoffs arrive, his minutes will go up and he will be more effective having had his regular season playing time curtailed.
In addition, according to your table, Devean George is the second-worst player on the team, behind guys like Berea and Bonsu who don’t play. I’m not arguing that George is some sort of superstar, but he has clearly been a bonus as a role player for this team this year, if you actually bother to WATCH the games.
I certainly am a big supporter of sabermetrics in baseball and I always cringe when people like Joe Morgan slam it without any understanding of what they’re criticizing. But I think this kind of statistical analysis doesn’t translate as well to basketball. Most basketball statistics only measure what happens when a player touches the ball. Point guards handle the ball about 30% of the time and everybody else has the ball in their hands about 18% of the time. Your stats don’t measure screens set, loose balls corralled, boxing out, on the ball defensive pressure applied, etc., all of the intangible things that don’t show up in box scores.
I particular do not understand the overwhelming emphasis applied to point differential. The Spurs have not only been winning consistently lately, 19 out of 21 to be exact, but they’ve been whooping their competition in blowout after blowout. However, they haven’t beaten that many quality teams that they would potentially see in the playoffs, and I do believe that once a game gets out of hand, it really does not matter if the final margin of victory is 20 or 30 points. For all practical purposes, a win by 20 points is qualitatively no different from winning by 30. But what does it say about the Spurs that they’ve lost home games to the likes of Charlotte and Boston? Of course they would easily handle those teams in a playoff series, but those kinds of losses to inferior teams make me question their focus and ability to play up to their potential on a consistent basis.
I don’t really feel like regular season success is any kind of reliable guide for what will happen in the postseason. Miami had a 52-30 record last year, good for sure, but hardly dominant and nobody was picking them to go all the way with Detroit/San Antonio/Phoenix/Dallas also in the running. If a player has a ball in his hands only 20% of the time, then statistics do not measure what he is doing the other 80% of the time he is on the court.
Playoff basketball is also fundamentally different from regular season basketball. Rotations shorten, rookies and second-year players tend to get benched in favor of veterans with more experience, the games are more physical and are called differently by the refs, and each team has the same amount of time to prepare for games and should be equally susceptible to fatigue. In the regular season, one team might be at home and was idle the previous night, whereas the opposing team is in the midst of a road trip and is playing its fourth game in five nights.
As to why Dallas might be winning more than what your statistical analysis predicts, having watched a lot of their games, they don’t blow teams out very much, but they tend to execute better in the fourth quarter and crunchtime, game after game. They’re a good defensive team, and they get stops when they need them. They’re the only team in the league to have a winning record when trailing after three quarters. They have the best record in the league in games decided by three points or less. They’re the best free-throw shooting team in the league, so they tend to preserve leads, the mid-March chokejob against Phoenix notwithstanding.
Also according to your analysis, Devin Harris, Greg Buckner and Jerry Stackhouse are “average” players that do not significantly affect this team’s success. False. Harris’ ability to take guys off the dribble and Stackhouse’s ability to create shot opportunities for himself in iso situations makes it hard for opposing teams to double Dirk. If your statistical analysis fails to discern this, then I think it leaves a lot to be desired.
I certainly do not feel that Dallas’ regular season dominance guarantees it a title (look at Detroit last year), but up to this point they have outperformed both San Antonio and Phoenix, and any kind of analysis should reflect that. Maybe you need to tweak your formula some.
Huey
April 4, 2007
Daniel, you should probably read some of the other articles on this blog.
dberri
April 6, 2007
Sorry for the delay on posting comments.
Let me try and answer some questions.
Jerry,
In a competitive labor market, workers are paid a wage equal to their marginal revenue product. Owners derive their return from the input of capital and entrepreneurial ability. Better labor, by the way, should enhance the return to capital (just like more capital makes labor more productive).
Rashad and Zebra,
Nestled in Daniel’s lengthy comments is part of your answer. Dallas has done better in close games this year. San Antonio has blown some people out. We are looking at the link between wins and aggregate stats, and that relationship will be somewhat skewed because of close games and blowouts. I am skeptical that teams have an ability to win close games. My sense is that this is more about luck.
Daniel,
I did post on the merits of Jerry Stackhouse during last year’s playoffs. I know Stackhouse looks like he is helping the Mavericks win games. But when you look at everything (not just scoring) it is clear that is not the case.
Flagstaff Home Auto Insurance
June 26, 2007
Detroit kind of let down against Cleveland. Too bad, I would rather Detroit face San Antonio. Even though Cleveland beat Detroit, the Pistons would have given Duncan and Parker a better game.