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
Playoff Prospectus (04/27)
Next Article >>
Playoff Prospectus (04/28)

April 27, 2009
Prospectus Hoops List
There Is No Secret Sauce

by Bradford Doolittle

Printer-
friendly
Contact
Author

When the playoffs began last spring, I decided to try on the big shoes at Prospectus headquarters belonging to Nate Silver. Specifically, I decided to find some magical combination of statistical indicators that would identify the prime playoff contenders, a formula that would be the hoops version of Nate's "secret sauce".

Don't remember the results of that work? There's a good reason for that. After crunching the numbers in my post-ABA database in ever conceivable way I could imagine, I could not come up with any combination of metrics that correlated better with playoff success than just plain old efficiency margin. Even that relatively advanced measure barely outpaced regular-season wins or unadjusted point differential. I tried to find something more interesting. I looked at late-season performance. I looked at road success. I looked at every mashup of metrics under the sun and nothing emerged as a "magic formula." Alas, the hoops secret sauce (hot sauce?) was never born.

When this season rolled around, I decided to take on another proven Prospectus convention when I produced the weekly Prospectus Hoops List. When I developed the formula by which the teams were ranked, I had two aims. First, I wanted to filter out any noise in a team's won-loss record to as much as possible identify the true talent level of the team. Second, I wanted that talent level to be expressed in terms of wins. Thus when I say that Cleveland has a 65.2 power rating, I'm saying that the Cavaliers have the true talent level of a 65-win team.

In the end, I think the formula worked pretty well. The Hoops List solidified for the most part by mid-December and some teams, like Portland, were slotted according to their eventual won-loss record well before they actually got on pace to achieve that record. The formula wasn't perfect. For one, it was slow to catch onto how good Orlando was this season because of the Magic's relatively weak early schedule. By the end of the season, however, the power ratings looked very much like the final standings, with no team deviating more than 3-4 wins from its final record and most teams finishing just about spot on.

With the regular-season rankings behind us, the next question was whether the system had any predictive value for the playoff bracket. After last spring's ordeal, I wasn't optimistic. Indeed, once I back-calculated my current formula for power rating for the teams in my historical database, I again found that I had yet to discover a magic bullet for forecasting the NBA playoffs. Maybe someday.

The good news is that the system did hold up as well as efficiency margin, which means that I can employ my power ratings in all sorts of fun ways. So that's what we're going to do today.

First off, let's record the final power rankings list for the sake of posterity.

FINAL 2008-09 POWER RANKINGS

Rank. (Last week) Team (Power rating) [Win pace / Pythagorean win pace / Preseason projecion ]

1. (1) Cleveland Cavaliers (65.3) [ 66 / 64 / 44 ]
Rankings: NET: 1; OFF: 4; DEF: 3; PACE: 25

2. (2) Boston Celtics (61.8) [ 62 / 61 / 63 ]
Rankings: NET: 2; OFF: 5; DEF: 2; PACE: 18

3. (3) Los Angeles Lakers (61.4) [ 65 / 60 / 49 ]
Rankings: NET: 4; OFF: 3; DEF: 5; PACE: 6

4. (4) Orlando Magic (58.3) [ 59 / 59 / 43 ]
Rankings: NET: 3; OFF: 9; DEF: 1; PACE: 12

5. (5) Portland Trail Blazers (55.2) [ 54 / 56 / 38 ]
Rankings: NET: 5; OFF: 2; DEF: 12; PACE: 29

6. (7) Houston Rockets (53.1) [ 53 / 52 / 61 ]
Rankings: NET: 7; OFF: 16; DEF: 4; PACE: 19

7. (6) Denver Nuggets (52.2) [ 54 / 50 / 43 ]
Rankings: NET: 8; OFF: 7; DEF: 8; PACE: 5

8. (8) San Antonio Spurs (51.8) [ 54 / 52 / 55 ]
Rankings: NET: 6; OFF: 12; DEF: 6; PACE: 27

9. (11) Dallas Mavericks (48.2) [ 50 / 46 / 51 ]
Rankings: NET: 9; OFF: 6; DEF: 16; PACE: 16

10. (9) Utah Jazz (48.0) [ 48 / 48 / 50 ]
Rankings: NET: 11; OFF: 8; DEF: 11; PACE: 10

11. (10) New Orleans Hornets (47.4) [ 49 / 45 / 45 ]
Rankings: NET: 12; OFF: 13; DEF: 9; PACE: 28

12. (12) Atlanta Hawks (46.8) [ 47 / 45 / 38 ]
Rankings: NET: 13; OFF: 10; DEF: 10; PACE: 24

13. (13) Phoenix Suns (45.6) [ 46 / 46 / 49 ]
Rankings: NET: 10; OFF: 1; DEF: 25; PACE: 4

14. (14) Miami Heat (42.9) [ 43 / 41 / 35 ]
Rankings: NET: 14; OFF: 18; DEF: 13; PACE: 22

15. (15) Philadelphia 76ers (41.1) [ 41 / 41 / 38 ]
Rankings: NET: 15; OFF: 20; DEF: 14; PACE: 20

16. (16) Chicago Bulls (40.6) [ 41 / 40 / 43 ]
Rankings: NET: 16; OFF: 19; DEF: 18; PACE: 9

17. (17) Detroit Pistons (39.3) [ 39 / 39 / 55 ]
Rankings: NET: 17; OFF: 21; DEF: 17; PACE: 30

18. (19) Indiana Pacers (37.4) [ 36 / 37 / 41 ]
Rankings: NET: 19; OFF: 15; DEF: 19; PACE: 3

19. (20) Milwaukee Bucks (36.1) [ 35 / 38 / 31 ]
Rankings: NET: 18; OFF: 23; DEF: 15; PACE: 11

20. (18) Charlotte Bobcats (35.7) [ 35 / 37 / 38 ]
Rankings: NET: 20; OFF: 27; DEF: 7; PACE: 26

21. (21) New Jersey Nets (34.0) [ 34 / 34 / 38 ]
Rankings: NET: 21; OFF: 14; DEF: 23; PACE: 23

22. (22) Toronto Raptors (33.3) [ 33 / 33 / 42 ]
Rankings: NET: 22; OFF: 22; DEF: 22; PACE: 14

23. (23) New York Knicks (33.1) [ 32 / 34 / 33 ]
Rankings: NET: 23; OFF: 17; DEF: 24; PACE: 2

24. (24) Golden State Warriors (29.5) [ 29 / 31 / 35 ]
Rankings: NET: 24; OFF: 11; DEF: 28; PACE: 1

25. (25) Minnesota Timberwolves (26.5) [ 24 / 27 / 26 ]
Rankings: NET: 26; OFF: 25; DEF: 27; PACE: 15

26. (26) Memphis Grizzlies (24.7) [ 23 / 25 / 19 ]
Rankings: NET: 25; OFF: 28; DEF: 20; PACE: 21

27. (27) Oklahoma City Thunder (23.3) [ 23 / 24 / 23 ]
Rankings: NET: 27; OFF: 29; DEF: 21; PACE: 8

28. (29) Washington Wizards (19.9) [ 19 / 21 / 32 ]
Rankings: NET: 28; OFF: 26; DEF: 29; PACE: 17

29. (28) Los Angeles Clippers (18.9) [ 19 / 18 / 33 ]
Rankings: NET: 29; OFF: 30; DEF: 26; PACE: 13

30. (30) Sacramento Kings (17.9) [ 17 / 19 / 39 ]
Rankings: NET: 30; OFF: 24; DEF: 30; PACE: 7

By the end of the season, there was very little movement in the rankings. Between the last two lists, eight teams moved one spot, while two teams moved two. The Mavericks jumped up a couple of notches with their late-season surge. The Bobcats fell two places by getting hammered down the stretch. Let's look at the progression of the ratings by looking at the rankings at the end of each month.

Team                    11/27  12/25  1/25  2/22  3/29  4/15
Cleveland Cavaliers 2 2 1 1 1 1
Boston Celtics 3 1 2 2 2 2
Los Angeles Lakers 1 3 3 3 3 3
Orlando Magic 11 4 4 4 4 4
Portland Trail Blazers 4 6 7 7 5 5
Houston Rockets 8 7 8 8 8 6
Denver Nuggets 5 5 5 5 6 7
San Antonio Spurs 21 11 10 6 7 8
Dallas Mavericks 12 10 13 11 12 9
Utah Jazz 14 13 11 10 9 10
New Orleans Hornets 10 8 6 9 10 11
Atlanta Hawks 9 9 9 12 11 12
Phoenix Suns 7 12 12 13 13 13
Miami Heat 18 17 14 15 14 14
Philadelphia 76ers 19 20 15 14 15 15
Chicago Bulls 15 16 23 18 16 16
Detroit Pistons 6 15 16 17 17 17
Indiana Pacers 13 18 18 19 19 18
Milwaukee Bucks 17 14 17 16 20 19
Charlotte Bobcats 26 23 20 22 18 20
New Jersey Nets 20 21 22 23 21 21
Toronto Raptors 16 19 19 21 23 22
New York Knicks 23 22 21 20 22 23
Golden State Warriors 24 25 25 24 24 24
Minnesota Timberwolves 25 29 24 25 25 25
Memphis Grizzlies 27 26 28 26 27 26
Oklahoma City Thunder 30 30 29 27 26 27
Washington Wizards 28 28 27 29 29 28
Los Angeles Clippers 29 24 26 28 28 29
Sacramento Kings 22 27 30 30 30 30

corr w/ final >> .86 .96 .97 .99 .99 1.00

Mobility was limited by the time the calendar turned over to the new year. No team moved more than five spots in either direction. There was some movement within that range, but we could have started the playoffs on Christmas Day and the brackets wouldn't look much different than they do now.

Now let's get to the fun stuff. To start, let's look at where the Cavs' 65.3 final power ranking slots them historically, which in NBAPET means since the ABA-NBA merger.

Team         Year   POWER    Postseason
Bulls 1995-96 73.2 won NBA finals
Bulls 1996-97 72.5 won NBA finals
Bulls 1991-92 68.2 won NBA finals
Mavericks 2006-07 66.1 lost first round
Celtics 1985-86 66.1 won NBA finals
Lakers 1999-00 65.8 won NBA finals
Bulls 1997-98 65.5 won NBA finals
Lakers 1986-87 65.3 won NBA finals
Cavaliers 2008-09 65.3 ????
Celtics 2007-08 65.2 won NBA finals

Suffice to say, Cleveland's regular season bodes well for LeBron James' chances for his first championship ring. At the same time, the one playoff bust on this list, the Mavericks of a couple of seasons ago, is fresh enough in NBA fans' collective memory that no one can take it for granted that the Cavs will fly to the title unchallenged.

As for the other teams, there is always hope. In the 32 seasons that for which I calculated power ratings, the top team won 15 titles. The second-ranked teams (six) and third-ranked (five) combined to account for most of the rest. The outliers were the 1995 Rockets, which entered the postseason ranked No. 11 in the power rankings, and the 1978 Bullets, who were 10th. I guess that offers a glimmer of hope for this year's Jazz and Hornets, however in the case of the Bullets, that title run came during an unparalleled era of parity in the NBA.

To figure out the exact probabilities for this year's playoff bracket, I employed a log5-based simulater that I adapted from some work I did prior to the NCAA tournament. Armed only with this year's power ratings, I figured out the percentage chances each team win every possible matchup they could face in this year's playoffs. There are 120 possible matchups. The method for figuring those percentages is the same as I used last year.

After I had the percentages for every matchup, I was able to plug them into a simulation model and run the playoff bracket 1,000 times. After doing so, this what I ended up with:

w/ Garnett Seed R8% R4% R2%   R1%
Cleveland E1 99% 93% 64% 48.6%
LA Lakers W1 86% 65% 52% 23.0%
Boston E2 95% 64% 25% 15.0%
Orlando E3 88% 35% 10% 5.1%
Portland W4 56% 18% 10% 2.4%
Denver W2 64% 36% 13% 1.9%
San Antonio W3 60% 33% 10% 1.7%
Houston W5 44% 12% 7% 1.3%
Dallas W6 40% 17% 4% 0.3%
New Orleans W7 36% 14% 3% 0.3%
Atlanta E4 61% 4% 1% 0.2%
Miami E5 39% 2% 0% 0.1%
Utah W8 14% 5% 2% 0.1%
Philadelphia E6 12% 1% 0% 0.0%
Chicago E7 5% 0% 0% 0.0%
Detroit E8 1% 1% 0% 0.0%

What you're looking at is the percentage chance for each team to reach a given round. So the simulation suggests Cleveland is a near-even bet to win the title (48.6%). This, of course, underscores what I suspect is a split between hoops statheads and the mainstream. Everyone and their dog (including my dog Hunter) thinks the Cavs and Lakers will meet in the finals. Indeed, given the records of the respective teams in question, it would be kind of crazy to anticipate any other matchup. However, I think most "mainstreamers" favor the Lakers in that eventual clash. Stat-based power rankings, on the other hand, have favored Cleveland all season and, as such, the people behind those systems probably pick the Cavs to win. This is my suspicion, anyway.

You might notice that the Cavaliers also have a higher probability of reaching the finals than the Lakers do. What's interesting about that is the common perception is that Cleveland has the tougher road. The Lakers have been head and shoulders above the West all season, while the Cavs have the Celtics and Cavaliers with which to contend. Here's the thing: Cleveland's road is easier. The reason is simple. The Magic and Celtics do half of Cleveland's dirty work in their East semifinal matchup. In the first round, the Cavaliers have to contend with the Pistons, No. 17 in the power rankings and with only a 1.7% chance of beating Cleveland in a seven-game series. (Now down to zero.) In the second round, Cleveland will face either Atlanta (6.2% chance of beating the Cavs) or Miami (3.4%).

Meanwhile, the Lakers have Utah (13.9%) and then either Portland (28.9%) or Houston (23.5%). Cleveland only has one difficult series in their path to the finals; the Lakers have two. In the conference finals, however, the Cavs do face the bigger challenge.

The other thing you might notice is that the chart has that "w/ Garnett" designation. I ran the simulation based on regular-season power rankings, nothing more. I didn't account for Kevin Garnett's absence in Boston or Manu Ginobili's injury in San Antonio. I think that's the best way to handle things. Garnett could return. Another major player could get hurt. It's better to go with the bigger picture. However, if you re-figure Boston's power ranking sans Garnett, the overall probabilities take on a dramatically different look.

wo/ Garnett Seed R8% R4% R2%    R1%
Cleveland E1 99% 95% 80% 62.3%
LA Lakers W1 86% 63% 52% 21.7%
Orlando E3 90% 39% 9% 5.0%
Boston E2 81% 56% 9% 3.3%
Denver W2 65% 35% 13% 2.6%
Portland W4 60% 21% 10% 2.1%
San Antonio W3 62% 35% 12% 2.0%
Houston W5 40% 11% 5% 0.3%
Dallas W6 38% 16% 3% 0.2%
New Orleans W7 35% 14% 2% 0.2%
Utah W8 14% 5% 2% 0.2%
Atlanta E4 63% 3% 1% 0.1%
Miami E5 37% 1% 1% 0.0%
Philadelphia E6 10% 3% 0% 0.0%
Chicago E7 20% 2% 0% 0.0%
Detroit E8 1% 1% 0% 0.0%

I based my adjusted Celtics power ranking on Boston's efficiencies with and without Garnett. Basically, Boston has scored at about the same rate without KG, but is giving up about 4.3 more points per 100 possessions. That's enough to drop them from a 61-62 win team to a 51-52 win squad. It's also enough to further ensure that the Cavaliers and Lakers will eventually meet in the finals--a clash that feels inevitable. The inevitable doesn't always happen in sports and, in fact, it usually doesn't. (Think Giants/Patriots Super Bowl.) However, in this case, if Cleveland and L.A. do hook up for the championship, a Kobe-LeBron duel would be a nice way to top off the season.

Definitions:

NBAPET = stands for National Basketball Association Projection, Evaluation and Tracking = A database and system of metrics for analyzing professional basketball.

gRATE = a one-game metric that measures a player's offensive and defensive contribution and expresses it as a net point total. The sum of a team's gRATE figures for a game will equal its actual point differential for that game.

Adjusted winning percentage (AWP) = ((home wins x 0.6)+(road wins x 1.4)) / (((home wins x 0.6)+(road wins x 1.4)) + ((home losses x .1.4)+(road losses x 0.6)))

LUCK = the difference between a team's 82-game win pace and its 82-game Pythagorean win pace.

Opponents winning percentage (OWP) = aggregate percentage of games won for each team's opponents, based on the number of times the team has faced that opponent.

Pythagorean winning percentage (PYTH) = uses the basketball-reference formula of Games x (Points scored^14) / ((Points scored^14) + (Points allowed^14))

Power rating = (((PYTH + AWP)/2)*(OWP/.500)) x 82

WP82 = wins produced per 82 games, adjusted for playing time

WP3K = wins produced per 3,000 minutes

RANKINGS: NET = net efficiency ratio; OFF - offensive efficiency; DEF - defensive efficiency; PACE: average possessions per game

NOTE: Permanent credits are due to 82games.com and basketball-reference.com, both sites which are used liberally when compiling the Hoops List.

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

0 comments have been left for this article.

<< Previous Article
Playoff Prospectus (04/27)
Next Article >>
Playoff Prospectus (04/28)

RECENTLY AT BASKETBALL PROSPECTUS
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

MORE FROM APRIL 27, 2009
Playoff Prospectus: Another Classic

MORE BY BRADFORD DOOLITTLE
2009-06-08 - Playoff Prospectus: So Close
2009-06-04 - Back and Forth: The Finals
2009-05-27 - Matchup Problems: Why the Magic Gives the Ca...
2009-04-27 - : There Is No Secret Sauce
2009-04-12 - : Week of...Next to Last Edition
2009-03-31 - Stock Watching: The Impact of March Madness
2009-03-26 - Bracket Breakdown: Memphis Is Under a Micros...
More...

MORE PROSPECTUS HOOPS LIST
2010-02-01 - : Here come the Cavaliers
2010-01-20 - : Week ending Jan. 17, 2010
2010-01-12 - : 2009-10 Premier Edition
2009-04-27 - : There Is No Secret Sauce
2009-04-12 - : Week of...Next to Last Edition
2009-03-18 - : Week of March 16, 2009
2009-03-11 - : Week of March 8, 2009
More...

Basketball Prospectus Home  |  Terms of Service  |  Privacy Policy  |  Contact Us
Copyright © 1996-2014 Prospectus Entertainment Ventures, LLC.