DFDepressed FanDepressed Fan



, all the time

Sixers 2010-11 PERformance to Date

Statman checks in with a look at the Sixers production viewed through the PER or Production lens. There are some eye-openers in here, check it out, download his spreadsheet if you want a closer look, and check out 82games.com if you'd like to do some digging of your own. On a personal note, I'd like to thank Statman, Rich, GoSixers and TK76 for their contributions to the site this season. You guys are all doing a great job. It's nice to have a couple more voices and viewpoints around here.

82games.com is one of my favorite sites to visit, because they show individual performance -- for both offense and defense -- in an easy-to-digest format.  Their primary individual statistic is called "Production" and is a variation of John Hollinger's Player Efficiency Rating (PER).  82games doesn't explicitly say how they calculate Production, but one would assume that the calculation is similar to PER.  For those not familiar with PER, it attempts to capture all aspects of player production based on boxscore statistics:  points, rebounds, assists, steals, blocks, fouls, and shooting efficiency.  PER takes into account several team and league factors to ascertain the relative value of an assist, rebound, and possession.  After the raw PER is obtained, it is normalized to 15, so that an "average" PER is 15.  Typically, 20 is an excellent PER and 10 is a poor PER.  See here for a detailed explanation of how PER is calculated.

As I mentioned, what's interesting about 82games.com is that they report PER values for both offense and defense (meaning opposing players). [Note that for the remainder of this post, I'm referring to their Production stat as PER, since the latter is shorter to type.] What they don't do at 82games.com is gather all the starters throughout the league at different positions into a single chart, so that one can get a sense of how a player is doing in relation to his counterparts around the league. I've been interested in such a chart for a while, so I decided to compile it myself. Read on for the results.

All the results are from the most recent update of 82games.com, which was last updated on 12/27 (so for the Sixers, it doesn't include the Warriors and Suns games but presumably does include the Nuggets game).  Regarding defensive PER, the usual disclaimers apply:  it doesn't account for switches and fastbreak baskets, and it doesn't account for "difficulty" of wing matchups (i.e., which player takes the best scorers on the other team).  Presumably, though, all players at a given position (e.g., starting PG) would be equally affected by the first two factors.

The first table shows the PER (own and opponent) values for the 8 Sixers who have played at least 40% of the team's minutes this year.  These values can be found at 82games.com at the Sixers' team page.

AI SF 17.2 10.5 6.7
EB PF 20.9 16.0 4.9
SH C 13.0 19.8 -6.8
JH PG 15.3 19.8 -4.5
JM SG 12.2 11.0 1.2
TY BN 20.2 12.5 7.7
LW BN 16.8 13.3 3.5
ET BN 8.4 15.8 -7.4

Remembering that 15.0 is the "average" PER value, we see that Brand, Thad, Iguodala, Lou, and Jrue have above-average PER values and that Iguodala, Meeks, Thad, and Lou have above-average opponent PERs.  For Thad and Lou, it's helpful for their opponent PERs that they are usually guarding reserves.  Generally speaking, the PER calculation is such that Centers and PGs have the highest values (because of more rebounds for Centers and more assists for PGs), PFs are next (again because of more rebounds), and SGs and SFs are lowest (having neither high rebounds or assists).  So there is a positional bias as well, and the PER difference (own to opponent) is perhaps most illuminating.

Now in the next table, I've listed the rankings by position for each of the starters among all the starters in the league (30 in all at each position).  For the most part, it was obvious who should be listed as "starter" for each team (somebody who's played over 50% of the team's minutes at a given position and is the usual starter at that position), but there were some cases where I filled in the guy with the most minutes (e.g., Ilguaskas of MIA) or the guy who is the nominal starter (e.g., Bogans in CHI).  For bench players, I ranked all other players around the league, independent of position, who had played at least 40% of their team's minutes thus far.  There were 45 such players.
AI SF/30 11 4 6
EB PF/30 11 13 11
SH SF/30 28 24 27
JH PG/30 26 28 28
JM SG/30 23 2 12
TY BN/45 1 9 1
LW BN/45 4 13 7
ET BN/45 44 28 42

Observations from  this table:
  • Iguodala, not surprisingly, has an excellent defensive rating (only Kirilenko, Gay, and Mbah a Moute are ahead among starting SFs), and his PER differential only trails the elite of the league (Durant, Carmelo, Gay, Lebron, Pierce).
  • Brand has been above average, as expected.
  • Hawes has been well below average, as expected.
  • Meeks has been above average because of his defense, but I expect that to change with the last two games (where his opposition has combined for 40 points and 14 assists and his PER in the last game was -5.9).
  • Most surprisingly, Jrue has played poorly according to these metrics.  He's been excellent in the last three games, so expect the offensive numbers to rise (currently, his numbers are pedestrian: 44/35/79 and a 2-to-1 assist-to-turnover ratio).  But his defense has been below-average all year, and I've noticed this myself in most of the games I've tracked.  Iguodala's defense also suffered in his first year as a go-to option on offense (06-07 to 07-08), so maybe there's something like that going on with Jrue.
  • We all know Thad has played well, but according to PER he is the #1 bench player in the league, both offensively and in terms of differential (not too shabby defensively either).  Lou is also in the top 10 among bench players in PER differential.  Turner is toward the bottom (only Rasual Butler ranks worse in own PER), but expect that to change with more games like the Suns game (36.6 PER).
I've forwarded the Excel spreadsheet to Brian with all the raw data, so you can play with the sorting yourself.  If you sort the positions by PER differential, you'll find that even though PER has some "holes" in it such that it doesn't capture performance completely, it does a remarkable job in identifying the best players in the league.

Download the spreadsheet here.

by Statman on Dec 31 2010
Tags: Advanced Stats | Basketball | Sixers | Statman |