[Power Rankings] Are the Rockets in trouble? (UPDATED) Every year since the 2005-2006 NBA season, I have computed my own NBA power rankings to see where the Rockets stand among the cream of the crop. Given the current status of the Rockets in the 2010-2011 NBA season, I have decided to share my findings. There are many power rankings floating out there in the tubes. Each power ranking has its own mechanism for computation. For example, let’s take a look at ESPN because it has two ranking mechanisms: the (1) Marc Stein method and the (2) John Hollinger method. Marc Stein’s method is a human-based weekly power ranking that includes the opinion of one person. In fact, Stein’s method focuses mainly on how well a team is performing lately. In John Hollinger’s method, the power ranking is a little more elaborate. The power rankings are obtained by performing a numerical analysis on various factors such as: scoring margin, strength of schedule, recent performance, and home and road games. NBA.com has power rankings that are based on the opinion of one (John Schuhmann). In these rankings, the pace, offensive efficiency, and defensive efficiency are listed but I am not sure if they use this in the physical computation of the rankings. How do I compute my power rankings? I compute the power rankings using the offensive efficiency (OE) and defensive efficiency (DE). Why use offensive efficiency and defensive efficiency? Most NBA fans would probably state that a team that averages the most points in a season is the team with the best offense. However, that is not the case. For example, a fast paced team that is coached by Mike D’Antoni is pushed to score the ball in seven seconds or less (SSOL). For the sake of this example let’s say that the SSOL team scores only 30% of the time. Now say we have a slow paced team coached by Jeff Van Gundy that is pushed to score the ball in 20 seconds or more (TSOM) but scores 90% of the time. Now let’s say that we have a middle of the road (MOR) team that scores at a rate of 60%. If the SSOL and TSOM teams play against the MOR team (independently of course), which team will end up victorious? In this scenario, the TSOM team ends up victorious against the MOR team, while the SSOL team would lose. Obviously, although the SSOL team lost to the MOR team, they would score more points than what the TSOM team scored when they beat the MOR team. Therefore… Would you argue that the SSOL team has a better offense? Or would you say that the TSOM offense is more efficient? That is what offensive efficiency tries to mitigate. Offensive efficiency is calculated as the number of points a team score per 100 possessions. That leads to the questions on what constitutes a possession. A possession ends when a field goal attempt is made, lack of offensive rebound when a shot is missed, a turnover is made, or the free throw scenario. In the free throw scenario, you have to either be shooting for two or three shots and either make your last attempt or fail to obtain the offensive rebound. There are several ways to compute possessions out there but all of them take into account the aforementioned scenarios to count possessions. This season the Atlanta Hawks (ranked 13th in the league) and the Minnesota Timberwolves (ranked 9th in the league) are scoring an average of 100.5 and 101.5 points per game, respectively. However, the offensive efficiency of the Atlanta Hawks is 106.8 (ranked 4th in the league) and the offensive efficiency of the Minnesota Timberwolves is 98.5. Consequently, the Atlanta Hawks are more efficient at scoring than the Minnesota Timberwolves although the Timberwolves score more than the Hawks. Of course, having a great offensive efficiency is worthless if you have the worst defensive efficiency. Therefore, my metric tries to encompass both of these numbers to come up with a ranking. Efficiency Chart Here is a plot of normalized OE versus normalized DE. <table style="width:auto;"><tr><td><a href="http://picasaweb.google.com/lh/photo/Lb9G4LsT7f-hqpYR31i5UA?feat=embedwebsite"><img src="http://lh6.ggpht.com/_8u4A05PwTfw/TPAzy5HKs8I/AAAAAAAACOE/nnTN6-m7dNs/s640/nba10_pr_norm_chart13.jpg" height="480" width="640" /></a></td></tr><tr><td style="font-family:arial,sans-serif; font-size:11px; text-align:right">From <a href="http://picasaweb.google.com/jsmee2000/PowerRankings?feat=embedwebsite">Power Rankings</a></td></tr></table> This plot is normalized to the league average; therefore, an OE greater than zero is considered to be good while a DE smaller than zero is considered to be good. Consequently, teams that are in the 2nd quadrant are the best teams in the league. Conversely, the teams in the 4th quadrant are the worst teams in the league. This season is a little weird compared to other seasons as there is a clustering in the data centered around the (0,0). Power Rankings It is pretty well known that to be a contender in the NBA, you must have a great offense and great defense. As a result, to compute the power rankings, I basically added the perfect team with an OE of 30 and a DE of -30. So the question is… What is the distance from this perfect team to all the teams in the league? Mathematically, this can be computed using the norm, which gives you the Euclidean length. In a nutshell, the closer you are to this perfect team the better you are. In addition, the power rankings are computed on a per game basis. In other words, I take the minimum number of games played by any team in the NBA and then compute the rankings for that number of games for every team. I have used these rankings for a few years now and they are pretty solid. For the first 13 games of the season, the results are: 1. Miami Heat 2. Los Angeles Lakers 3. San Antonio Spurs 4. New Orleans Hornets 5. Boston Celtics 6. Orlando Magic 7. Portland Trailblazers 8. Dallas Mavericks 9. Chicago Bulls 10. Atlanta Hawks 11. Denver Nuggets 12. Utah Jazz 13. Oklahoma City 14. Milwaukee Bucks 15. Indiana Pacers 16. Charlotte Bobcats 17. New York Knicks 18. Phoenix Suns 19. Toronto Raptors 20. Houston Rockets 21. Philadelphia 76ers 22. New Jersey Nets 23. Memphis Grizzlies 24. Golden State Warriors 25. Detroit Pistons 26. Sacramento Kings 27. Cleveland Cavaliers 28. Washington Wizards 29. Minnesota Timberwolves 30. Los Angeles Clippers Looking at the rankings, we can see that the Houston Rockets are what they are… a borderline pretender/lottery bound team. Here is a plot that shows how the Rockets have fared in the power rankings as a function of games played. <table style="width:auto;"><tr><td><a href="http://picasaweb.google.com/lh/photo/LgQMfZKPAWMOwbXrtJ-Xwg?feat=embedwebsite"><img src="http://lh4.ggpht.com/_8u4A05PwTfw/TPAzy1DMrSI/AAAAAAAACOA/Gozuoi_CQik/s640/nba10_pr_rockets.jpg" height="480" width="640" /></a></td></tr><tr><td style="font-family:arial,sans-serif; font-size:11px; text-align:right">From <a href="http://picasaweb.google.com/jsmee2000/PowerRankings?feat=embedwebsite">Power Rankings</a></td></tr></table> The first three data points are shaded in gray because that statistical noise (or variance) in the data is so high that the measurements should be taken with a grain of salt. So far, it does not look good! In a nutshell, based on the power rankings the Rockets are a team that should be ranked 15-20 in the league. So clearly, the Rockets are underperforming as they are ranked 26th in league winning percentage. However, the Rockets are far from the hyped contenders that we, Clutchfans members, were thinking that the Rockets were in the preseason. Are the Rockets out of the playoff hunt? Based on five years’ worth of data that I have collected, no! They surely can make the playoffs as they are the 10th team in the rankings behind the Phoenix Suns and Oklahoma City Thunder. Nevertheless, if the Rockets are around 20 in the power rankings for the next 10 games they don’t stand a chance at a playoff bid. In summary, the Rockets are certainly not dead but are on life support. Updated table with strength of schedule can be found here
I don't understand the first chart. Does this mean that if you play Defense on 10, and offense on 10, you're a pretender?
Why are you ranking our performance though without 2 major pieces? Of course the 4-10 Luis Scola led Rockets arent contenders. Put this together once we are at full strength.
Thanks for the comment. I have been evaluating all NBA teams for five years. I evaluate teams as is and make no assumptions or adjustments for coaching changes, injuries, changes in personnel, etc. With that said, all I can say at the moment is where the Rockets are currently compared to the rest of the field. Of course, we shall see some major improvements when Yao comes in as our defense will solidify. In addition, perhaps, our offense may be more efficient with Brooks and Lowry in the line up. But... all of this is hearsay at this point. I am a man of science and I go by the facts.... and these are the facts.
Be a man of science and go by the facts. Is this squad as is, going to be competing in the playoffs? The answer is no. Either we get our guys back, make a dramatic improvement and compete or we remain injury plagued and dont make the playoffs. How this current squad compares to other teams seems to me, more or less irrelevant. edit* I appreciate your effort and it was well done, but I just think it would be way way more useful to do maybe ten games in after Brooks and Yao are back.
I really appreciate the comments and your desire for the Rockets success. However, the current squad has a major impact on our standings. Therefore, this analysis is pertinent and more critical than you are willing to give credit to. If Ming and Aaron were back today then I understand your judgment; however, given my previous statement I don't.
Great- you just used all those words and lots of hours of analysis to tell us what we already knew from watching our games- that we've performed as a below average team so far this season. However, I don't think that your power rankings has much predictive power of our future performance or anything like that, as there are so many unknown variables, so your conclusion is of limited value.
it's pretty obvious we're not that good. our defense is pathetic, and our top 2 players (Scola and Kmart) aren't as good as some of the other duos on top tier teams (Kobe Gasol, Westbrook Durant, Lebron Wade, etc.) we just don't have the talent level to compete at that level. Yao when healthy, is the only player who can hang with the top tier players, but he will never be that player again.
That is sort of correct! The Utah Jazz are borderline contender/pretender. That means that they are overperforming in the winning percentage rankings. When it comes down to it unless they tighten their defense or improve their offense they will not win it all. Look at all the close games that they won...some of those games could have tipped either way.
I like the analysis. It tells us that the team isn't working the way Morey had envisioned (not that the record doesn't already do that!). Basically, the team was meant to have such a good offense that it would make up for defensive mediocrity. The fact that OE is barely above the x-axis and the DE is pretty far to the right of the y-axis tells us that that's not working. Hopefully the Rox are turning things around though. Just a suggestion here: OE and DE are clearly important, but because they are measured on a per possession basis, it doesn't take into account the number of possessions. If you can also model the number of possessions and couple that with your OE and DE, you can do a simulation on the different teams out there that'll give you rating calibrated on both efficiency and volume.
That it itself may indicate a flaw in the system. I'm surprised that Portland is also ranked higher than Atlanta. The fact that Portland is a contender and Atlanta is a pretender (albeit not by much in either direction) shows that the system isn't perfect.
