With the lackluster play of our role-players I decided I would do some analysis to see what it might turn up. I've used different angles of statistical analysis in the process. Presented here is one angle using Hollinger's Player Efficiency Rating (PER) as a method for comparison. PER Background Simplified the PER formula essentially adds the good (made shots, steals, assists, rebounds, blocked shots, FT's) and substracts the bad (missed shots, turnovers, fouls). PER is also adjusted to a per-minute basis as well. The actual calculation is quite involved but fortunately we can pull PER scores from a couple of sources. PER scores for 05-06 season: www.knickerblogger.com PER scores 04-05 and older: www.basketball-reference.com Hollinger has set it up so that the league average every season PER score is 15.00. 2005-2006 PER Analysis: Western Conference I used the PER scores for the Western Conference mainly to save on time with the inclusion of Detroit and Miami for my own interest sake. I took the PER scores and since it's a per-minute figure and filtered out only players that have played 15+min per game for analysis. Here's what the PER scores for the Rockets roster for players playing more than 15+ min-per-game looks like: Using these stats I then graphed each player's PER score for a team by counting the number of players with PER scores ranging from 8 to 34. Keep in mind the NBA average PER score is 15.0. This is reflected by the RED line that I've placed on each graph which represents the NBA average PER score. Here are the PER Graphs:
One thing to note is that teams with 5 or more players with PER scores higher than 15.0 comprise the teams in the current 1st-8th seeds for the play-offs. Outside of that teams have 4 or less players with PER scores of 15.0 or higher. The exception is Seattle. What about the Rockets? Here's the Rockets PER chart: This was really surprising to me. It's easy to guess who the two players are with PER scores above 20 on the Rockets. What really is dramatic to me is that are role-players are all pretty much below league average in player efficiency even compared to the lousy Western Conference teams. Maybe this isn't totally surprising considering the way our role-players have played but I think this chart suggests that for whatever reason our role-players contribute on the whole below the NBA average. Drawing Theories & Conclusions Naturally I'm pretty interested in why there is such a disparity vs. the rest of the league. INJURIES: My first thought was well maybe it is because of all the injuries this year. So to see if this was plausible I checked the Rockets PER for last season. Here's the 04-05 Rockets PER Chart: The two other players above 15 are Mike James and Deke. So on inspection save for James and Mutombo the pattern looks pretty much the same as it is for 2006 with or without injury. What this seems to indicate to me is that we consistently get less player efficiency out of our role-players relative to the NBA average with or without injuries. On a tangent, this seems to challenge the notion that we can keep the current roster and with a healthy TMac & Yao we would have no problems. Instead it points out that we might have an inherent weakeness in getting production out of our role players. LOWER THAN AVERAGE SKILLS: Could it be that our role-players have just had a lower than average PER rating? The following is a table of PER scores for members of the roster with any history. I also added on the bottom a few players we have moved as well: This was an interesting table. For some players they seemed to have a history of being above average before joining the Rockets, while others have a consistent history of being rated below average in PER. Overall everyone either has stayed pretty consistently rated in PER where they have trended or have taken a dip after joining the Rockets. The notable exception is Mike James to this trend. I infer from this that the lower than average PER scores for the role-players as a result of Rocket choices: (1)because of the offensive/defensive system that we run & (2)choices we have made in the players we have brought on - that we value possibly other traits than what leads to higher PER ratings. Assuming PER scores are accurate indicators of contribution then this could be a pretty condemning of the Rockets personnel & coaching staff. To be fair JVG & CD may have gambled here a bit with the past off-season moves and it may be evident in the PER scores here. The Alston for MJ trade seemed to have been to bring in more of a pure PG while hoping to make up the loss of production through Derek Anderson & Swift. If Anderson & Swift were able to contribute as per their historical levels then things would be different - they both were above 15. Instead we got or are getting less from them than they've historically put up. Rafer on the otherhand is trending what he has put up in the past save for last season. On closer inspection the only thing that has changed from this season to last is just the number of shots he put up. His FG% this season is consistent to what it has been in the past. To me this strengthens the argument that it was mainly for his PG skills that we've brought him in. Closing Thoughts: There's a lot that can be interpreted here. I'm really interested in ways others can add to the discussion especially in the form of analysis since I certainly don't feel like I've covered all angles. I've come away from looking at this thinking that we really need to make some serious off-season moves this summer because our PER ratings don't look encouraging whatsoever leaving the roster they way it is. I'm in the middle of doing some assessments in terms of shooting % and efficiency as it relates to comparable inside-out offenses in the league currently. I'll post that when I get a chance.
"The proof is in the pudding", as they say. Great job Tango! Man are you good at this stuff or what! Your thorough analysis only proves what myself and almost every other fan on this board has realized by now: save for a healthy Yao/T-Mac, our roster absolutely SUCKS! No questions, no exceptions, no ifs or buts. If anything, that tells you just how good T-Mac and Yao are together that -- when healthy -- they camouflage our weaknesses and make it seem as if everything is alright. Our two stars have carrying too much of the load and the other guys are literally dragging them down. If no changes are made, then even when they're healthy they're going to play way too many minutes and will end up running out of gas towards the end of the season and the playoffs, and eventually it's going to show. This team is THE definition of a two-man squad, it's not just us fans that realize that, it's also pretty damn smart people like Doug Collins who keeps mentioning how he hasn't a seen a team "in the last 20 years" more dependent on one or two players to win like the current Rockets are. Still, I will give management and coaches another chance to right the ship next season. In today's NBA, depth and talent are VITAL to your success as a team. Collecting trash (undersized trash, actually) that no other NBA team wanted and a bunch of NBDL talent won't EVER get us anywhere, I don't care how good Yao and McGrady get. It's time to get these guys some help...
Interesting data, thank you. Hope Rox GM and scouting staff can study these in future selection of role players for the team.
do you think to an extent when you run a system like the rockets compared to a team like the suns that its harder to trade the rockets players with ugly stats then to trade the suns players with inflated stats. for instance the joe johnson trade i think phoenix got 3 1st round picks for the guy. thats crazy p.s. great analysis
This also means we need new coaching strategies or a new coach. So what available coaches right now have a history of raising PERs?
Cliff Notes version - Using PER as a comparison the Rockets role-players contribute below the NBA average. Injuries do not appear to be the reason for the below average contribution. JVG's system, bad picks by CD & JVG in combination appears to be the main reason for our below role-players lack of contribution.
Seems like a reasonable assertion. Someone could probably just check Joe Johnson's stats this season to substantiate if this is plausible.
I'm not sure I'm quite there yet regarding replacing coaches and coaching strategies. The PER comparison is definitely a red flag for me though regarding JVG's system. Whether by design or by chance we are a two-trick pony team. My guess is this is more by chance/bad decisions than by the system. It looked like to me CD & JVG were trying to improve in our PF and SG positions. The Anderson signing and the aftermath has really bombed out for us. The Swift signing hasn't born the fruit that we were hoping. I can't tell if it was just poor scouting on our part or if it was that JVG's system is just too hard to adapt to. Both of those have left us pretty big gaping holes with the absence of Mike James. Rafer I don't think was ever the answer to shore up the loss of Jame's production. I'm willing to give them the benefit of the doubt in trying to shore up the SG & PF positions. JVG & CD are put on notice though. If we bomb out a 2nd time then they both need to be on the hotseat and will have no excuse whatsoever. I would dare say that their necks are on the line - and I would guess that JVG is the one that has the most to lose.
I would take David Wesley over any other role player in the NBA any day Are we that bored to have pointless threads about how crappy our team is minus Yao and TBack?!? I dont want my statement to underwrite your work though Tango, much appreciated!
The question is WHY do we have such crappy role players on the team? Were the players themselves crappy? Or does our system make them look crappy?
Clearly this is a system issue. IMO, PER stats fail to adjust for the system. Defense, for example, is underrated. Yes, it factors in stats like steals, blocks and rebounds, but it doesn't incorporate overall team defense, rotations, limting oppossing fast break points, etc. (at least I don't think it does?) That said, as many have pointed out in here, obviously we need better role players, and certainly the system as a whole may not be perfect.
This is clearly a PLAYER issue. Dont get me wrong, I dont thin the JVG is good for us in ther long run, but is ithis fault we cannot make a shot?!? Th last ten games, there have been PLENTY of open shots that we should have knocked down. I think more of the blame has to fall upont he shooters. It doesnt matter who is out, who is injured, you have to knock down shots. We cannot do that! I thought that being able to make a jump shot was a prerequisite for an NBA player? Correct me if I am wrong, but this is more than bad luck. This is a team that cannot shoot, simple as that. No excuses, no other reasons. We cannot make an open jumper. I don't know what else there is to it. The object of the game is to put the ball in the bucket, and we cannot do that. Actually it should be more of than , because for crying out loud, YOU ARE IN THE NBA!!!!!!!!!!!
Quantitative support for my season-long qualitative analysis; our role players are below par, and even with a healthy TMac and Ming are incapable of helping this team achieve contender status. If this team is not significantly different next season, we are all in for another season of disappointment. The supporting cast is less talented than last year, and the group last year was not talented enough to get past the first round. When you can't shoot, you really can't win. A couple of consistent shooters would have a huge impact here, huge. With a healthy Ming and TMac, a Jason Terry type of shooter would make an enormous difference.
I would tend to agree that it's more of a player issue than a system issue. JayZ750 is right -the PER doesn't account for defensive contribution very well. That being said take a look at the following... The top 6 best defensive teams in the NBA in terms of limiting opposition points are: (1) Spurs - 88.2 pts (2) Grizzlies - 89 pts (3) Pistons - 90.4 pts (4) Pacers - 91.3 pts (5) ROCKETS - 91.7 pts (6) Mavericks - 92.7 pts Spurs, Grizzlies, Rockets, & Maverick PER graphs are above so you can compare. I don't have the Pacers graph but here's the Piston's graph: Their PER charts all look better than the Rockets. So the notion that defensive oriented teams with no offensive firepower wins games I think is a sham. This would suggest that our role players just don't measure up. The kicker for me though is to go a step further to ask the question so how or why did the current role players end up on the Rockets? That leads us squarely back to JVG & CD - either having confidence in a flawed system or making bad judgement calls regarding the players we want on this team.
Great job, Tango. Some random observations: 1. PER mostly measures offence. Some defensive factors are taken into account (e.g. blocked shots) but not very many. Not surprising, of course, as defence is very difficult to quantify individually. I notice that most of our role players suffered a drop in PER after joining the team. Could the JVG effect be a stunting of a role player's offence but an improvement in the defence? I don't know. (Much as I dislike JVG, I want to give him the benefit of the doubt.) Maybe the historical plus/minus numbers would be illuminating. I would do the analysis but I have no idea where to find the data. 2. Yao's PER of 26 is right on schedule. According to http://www.basketball-reference.com, both TD and KG passed the 26-PER level for the first time at about Yao's age. (On the other hand, Shaq was already at 28.5 by age 21 -- and Hakeem didn't pass 26.0 until the age of 30.)
