After a banner week in which the Mavs defeated OKC, SAN, and MIA they have risen to the top of our power ratings. The Mavs have the league’s best defense while the Wolves have the worst. The Suns have the best offense and the Kings the worst. Again these ratings are based on game scores, not directly on win-loss records.
| Rank | Team | Off | Def | Overall |
| 1 | Dallas Mavericks | -0.05 | -8.16 | 8.11 |
| 2 | New Orleans Hornets | 0.10 | -7.21 | 7.32 |
| 3 | Miami Heat | 1.98 | -5.18 | 7.15 |
| 4 | Boston Celtics | 0.47 | -6.60 | 7.07 |
| 5 | Los Angeles Lakers | 7.02 | 0.60 | 6.43 |
| 6 | San Antonio Spurs | 5.37 | -1.05 | 6.42 |
| 7 | Denver Nuggets | 5.19 | -0.05 | 5.25 |
| 8 | Utah Jazz | 0.48 | -3.68 | 4.16 |
| 9 | Orlando Magic | -3.37 | -6.73 | 3.36 |
| 10 | Chicago Bulls | -1.18 | -3.72 | 2.53 |
| 11 | Phoenix Suns | 9.60 | 8.74 | 0.86 |
| 12 | Portland Trail Blazers | -3.66 | -4.16 | 0.50 |
| 13 | Houston Rockets | 3.84 | 3.63 | 0.21 |
| 14 | Atlanta Hawks | -0.08 | -0.19 | 0.11 |
| 15 | Oklahoma City Thunder | 5.92 | 6.18 | -0.25 |
| 16 | Memphis Grizzlies | 0.92 | 1.26 | -0.34 |
| 17 | Milwaukee Bucks | -8.44 | -7.73 | -0.71 |
| 18 | Indiana Pacers | -1.39 | -0.57 | -0.82 |
| 19 | Charlotte Bobcats | -4.08 | -2.84 | -1.24 |
| 20 | Toronto Raptors | 1.56 | 2.95 | -1.39 |
| 21 | Philadelphia 76ers | -1.41 | 0.32 | -1.73 |
| 22 | New Jersey Nets | -5.08 | -1.85 | -3.23 |
| 23 | New York Knickerbockers | 5.39 | 8.89 | -3.50 |
| 24 | Detroit Pistons | -3.52 | 0.84 | -4.36 |
| 25 | Golden State Warriors | -0.87 | 4.18 | -5.04 |
| 26 | Cleveland Cavaliers | -3.39 | 1.84 | -5.23 |
| 27 | Los Angeles Clippers | -2.18 | 4.46 | -6.65 |
| 28 | Washington Wizards | -2.04 | 5.27 | -7.32 |
| 29 | Minnesota Timberwolves | 1.77 | 9.73 | -7.97 |
| 30 | Sacramento Kings | -8.88 | 0.84 | -9.72 |
was just wondering, but have you ever released your methodology for your rankings? I know you’re a big plus minus guy, but i want to see what else your rankings are based on. just want to try and compare it to other formulas out there (to peltons, hollingers etc)
Comment by garron — November 28, 2010 @ 6:38 pm
Wayne,
I’ve been comparing various 2009 ratings to 2010 ratings and I noticed a few things.
In a few instances, I was able to accurately project changes in the 2010 ratings as time passed based on the 2009 rating plus various insights into whether the team was likely to improve, stay the same, or get worse in 2010 because of trades and other issues.
In short, I guess what I am saying is that there are probably some sample size issues with any point differential method during the beginning of a new season and that insights gleamed from the prior year can be helpful. It may not be easy to automate them, but it might be worth thinking about.
In any event, at what point is the sample size for a year large enough that the point differential starts to accurately reflect the actual ability of the team instead of a lot short term randomness?
I lack the statistical background to answer that simple question. (unfortunately)
Thanks
Comment by Italian Stallion — December 1, 2010 @ 1:06 am
I think after 20 games or so things should be reliable.
Comment by wwinston — December 1, 2010 @ 9:36 pm