Home Unfiltered Articles Players Baseball Prospectus
Basketball Prospectus home
Click here to log in Click here for forgotten password Click here to subscribe

Click here for Important Basketball Prospectus Premium Information!

<< Previous Article
Everyone's Back (03/06)
Next Article >>
Around the Rim (03/06)

March 6, 2009
The Spears Factor
Measuring Inconsistency in the NBA

by Kevin Pelton


Time to open up the old mailbag. Last week, Justin wrote me to pose the following question:

My question is how can I find stats to back up my hunch that this year's Chicago Bulls are the most inconsistent team in NBA. It would really quell my frustrations to find out that I'm wrong and that in fact, our ups and downs this year are typical of a mediocre, middle of the pack team. Just to give you some quick background information, this season we have LOST to Memphis, Milwaukee, Minnesota twice, Oklahoma City, and now the Wizards. Meanwhile, we have BEATEN Phoenix twice, Utah twice, Dallas, Cleveland, Denver and Orlando.

Has there been a more Jekyll and Hyde team this season?

This gives me a chance to take a look at team consistency, a subject my colleague John Gasaway has considered extensively in the NCAA--coining the term "Winehouse Factor" to describe the teams that are most volatile from game to game. In the NBA, consistency hasn't gotten the same attention, making this a worthwhile endeavor.

We can start by looking at the issue along the same lines as Gasaway's methodology, by considering the deviation in team performance. Using game-by-game data from Basketball-Reference.com (through Monday), I calculated each team's average point differential and the standard deviation of those differentials, measuring how much they vary from average each game. Here are the most and least consistent teams by that measure:

Inconsistent     Dev   Consistent        Dev

New Jersey      15.5   Indiana           9.8
Phoenix         15.1   L.A. Lakers      10.0
L.A. Clippers   14.9   Detroit          10.5
Denver          14.6   Houston          11.2
Dallas          14.2   Oklahoma City    11.4

The Nets have been the NBA's most Winehouse-esque team in this regard, with plenty of blowouts in both wins and losses. They've lost nine games by 20-plus points, incredible given that New Jersey is still in the playoff race. Yet the Nets also have one of the season's most lopsided wins, by 44 points over another inconsistent squad, the Denver Nuggets.

On the other end, there are few measures where you would find the Lakers just barely ahead of the Oklahoma City Thunder. I like that this measure is in practice independent of team quality. The Lakers' outcomes and the Thunder's are both centered around the same range. It's just that in the Lakers' case they're centered around wins by 8-10 points, while Oklahoma City has played an unthinkable number of close games (and has lost most of them, as we considered earlier this week).

While this measure is interesting, it doesn't really serve to answer Justin's question. With a standard deviation of 12.7, the Bulls are basically exactly average. The thing is, the way this method works doesn't take into account what Justin considered subjectively, whether Chicago should have won or lost based on the quality of the opponent. A team that was perfectly consistent, winning or losing each game by the same margin, would in some sense be offering a different caliber of performance each night to offset changes in the opposition.

My next step, then, was to utilize the differential for each team as well as the location to create a projection of what the outcome would have been if a .500 team had played each game. For example, a completely average team going on the road (-3.5 points) to play the Cleveland Cavaliers (+9.6 points) would reasonably be expected to lose by 13.1 points. To determine how well the team actually played in this circumstance, we have to add 13.1 points to the actual outcome. So a nine-point loss at Quicken Loans Arena actually indicates a team has played better than average.

Recalculating everything using adjusted differentials, here is the new list of most and least consistent teams.

Inconsistent     Dev   Consistent        Dev

New Jersey      14.6   Indiana           8.8
L.A. Clippers   14.2   Charlotte        10.0
Denver          12.9   Cleveland        10.2
Dallas          12.6   Houston          11.2
Sacramento      12.5   Philadelphia     10.3

In most cases, the changes are not dramatic. The Nets and the Pacers still show up at opposite extremes, and Indiana comes out as positively metronomic once the quality of opposition is taken into account, at least relatively speaking.

What might be most interesting here is the differences between inconsistency and schedule-adjusted inconsistency. Take a look at the teams that saw their standard deviation decline the most and the handful that saw it increase (as well as the smallest decreases).

Decrease         Dev   Small/No Decrease  Dev

Phoenix         -2.7   Detroit           +0.4
Utah            -2.5   L.A. Lakers       +0.3
Portland        -2.3   Minnesota         -0.1
Philadelphia    -2.2   Memphis           -0.3
Toronto         -2.0   L.A. Clippers     -0.7

What are we to make of these lists? My interpretation is it shows us which teams play up and down to the level of their opponent and which simply play their game night in and night out. Take the Lakers, for example. They're winning fairly consistently, but because they are so good, when we adjust for schedule we'd expect the margins to be much larger. The Lakers have a relatively paltry six wins by 20-plus points, and just two of those since the start of December. By contrast, the Orlando Magic--with a worse overall point differential as well as record--has ten 20-point wins. The Lakers are toying with their opponents, allowing them to play close before ultimately putting them away.

At the other end of the extreme, having followed the Portland Trail Blazers more closely than any other team this season, it does not surprise me to see them as one of the teams that plays to the level of its competition the least. Portland has a strong record against sub-.500 teams, but has struggled to defeat playoff teams on the road.

The question then becomes, is consistency a good thing or a bad thing? The answer to that, as APBRmetrics pioneer Dean Oliver has repeatedly and persuasively argued, is "It depends." For an elite team like the Lakers, consistency means a higher winning percentage than we would predict based on their point differential alone. An inconsistent elite team tends to drop games it should win. On the other hand, the fact that a sub-.500 team like the Clippers shows up as inconsistent is a good thing. The Clippers are stealing some wins while getting blown out at other times.

None of this is intended to minimize the importance of point differential. Differential is still the best predictor of performance the following season. If like me, however, you believe the Lakers have the ability to find another gear in the postseason above and beyond what is indicated by their point differential, this is evidence in your favor.

What of Justin's beloved Bulls? Even when we adjust for quality of competition, Chicago still doesn't rank as one of the league's most inconsistent teams. Their schedule-adjusted deviation, 11.8, ranks the Bulls as more inconsistent than average but still just ninth in the league. Is Justin crazy? Actually, the evidence backs up his theory when you take a closer look.

For whatever reason, in terms of consistency, there have been two different Chicago squads this season. From the start of the season through Dec. 27, the Bulls were very consistent. 26 of their 30 games were within 10 points either direction of what we would expect based on the opposition and location. Over the next 30 games, Chicago went the other direction. Just 13 of their subsequent 30 games have been within 10 points of expectation, half the number as before. If we look strictly at this portion of the schedule, the Bulls' schedule-adjusted deviation is 13.8, which would rank them third in the league behind New Jersey and the L.A. Clippers.

Chicago's schedule-adjusted point differential in the last 30 games (-0.6) is only slightly better than over the first 30 games (-1.2). With a near-.500 team, consistency isn't a big factor in determining a team's record. So the Bulls haven't really benefited or been hurt from their tendency to defy expectations. All they've really done is drive Justin and his fellow fans crazy. So maybe the corresponding term for inconsistency in the NBA ought to be the "Spears Factor." Catchy, huh?

Kevin Pelton is an author of Basketball Prospectus. You can contact Kevin by clicking here or click here to see Kevin's other articles.

0 comments have been left for this article.

<< Previous Article
Everyone's Back (03/06)
Next Article >>
Around the Rim (03/06)

State of Basketball Prospectus: A Brief Anno...
Tuesday Truths: March-at-Last Edition
Easy Bubble Solver: The Triumphant Return
Premium Article Bubbles of their Own Making: Villanova, Temp...
Tuesday Truths: Crunch Time Edition

Around the Rim: Memphis Defends Itself
Everyone's Back: So What Happened?
From 344 to 65: Friday Update

2009-03-13 - Heating Up: Dwayne Wade Carries Miami
2009-03-12 - Jazzing It Up: Breaking Down the Streak
2009-03-10 - The Sloan Conference: An APBRMetrics Gatheri...
2009-03-06 - The Spears Factor: Measuring Inconsistency i...
2009-03-04 - Mythbusting: Experience and Close Games
2009-03-03 - Age and Success in the NBA: Doing the Math
2009-03-02 - Defying Expectation: The Spurs Keep Winning

Basketball Prospectus Home  |  Terms of Service  |  Privacy Policy  |  Contact Us
Copyright © 1996-2017 Prospectus Entertainment Ventures, LLC.