in-the-game.org

European Basketball Analysis

Players | Teams | Game List in the Stats Database

Adjusted plus/minus for Euroleague 09/10

with one comment

Thanks to some outstanding work by Eli on his blog Count The Basket, I got an idea of how to calculate adjusted plus/minus. From that point it was only a question of changing the Euroleague play by playa data to the right format. It might be interesting to note by the way that Eli was signed by an NBA franchise twelve days after he had published his guide to calculating adjusted +/-.

I’m sure you’re accustomed now to basic plus/minus, as you find it in boxscores. It is common to divide a player’s pure +/- by minutes or possessions and set it into context with offcourt data, which makes for the oncourt/offcourt +/- I published here (I mistakingly called it “adjusted” some time ago, but this isn’t the adjusted +/- everyone is talking about). I’m a big of a fan of +/-. I don’t think it is the end of all arguments, it doesn’t tell which player “is better” than the other, but it offers a whole new perspective, since it provides an individual stat solely based on team production, in contrary to the biases of the boxscore-based statistics you are all familiar with.

The problem with basic +/- is, that, say, a +10 in 10 minutes versus the five weakest players in the competition while playing alongside four good teammates weighs as much as a +10 versus five elite players while playing with rather average teammates. Adjusted +/- adjusts to both the performance of teammates and opponents during a time a player was on the floor. For player 1′s adj. +/- rating, the adj. +/- rating of all players that he either played alongside or matched up against, factors in, and those are depending on the players they themselves played with or against. This can be archieved by running a regression.

The results of such a regression for 09/10 and earlier NBA can be found here. I find the results surprisingly predictable. The fact that Ilardi’s defensive adj. +/- marks Ron Artest, Shane Battier, Manu Ginobili and Bruce Bowen as the four best perimeter defenders in a six year span from 2003 to 2009 sounds very interesting to me, since those are the four guys I’d have named too, and since you’ll hardly find a stat that provides these results.

When presenting Euroleague stats, sample size is always a problem. Last Euroleague season had 186 games. The 09/10 NBA season had over 1400 games. In the Euroleague, eight teams don’t get past the ten game regular season, another eight are eliminated after sixteen games (regular season + Top16), four more teams will be eliminated after the quarter finals (total of 19 to 21 games per team), the four Final Four teams will finish the season with 21 to 23 games. Experts consider a sample size of a 82-game NBA season as too small for adjusted +/-. You get an idea what that means for one Euroleague season. I would have to gather data from five Euroleague seasons to get significant results. And there you face the problem of player improvement and decline.

I thought about not presenting the stats because of the small sample size (a very high standard error), but in the end I just put it out there and let you decide what you do with it. Many of the results match my perceptions of a player, some don’t. Coincidence?

NAME CLUB ADJ +/- OFF ADJ +/- DEF ADJ +/- TOT STD.ERR
VUJCIC, NIKOLA OLY 16,93 11,06 27,99 13,81
DIXON, BOBBY ASV 16,88 9,59 26,47 20,80
SATO, ROMAIN SIE 16,76 14,62 31,38 11,58
CURTI, ALDO ORL 16,60 9,49 26,09 21,16
MAVROEIDIS, DIMITRIOS MAR 16,39 3,07 19,45 24,16
MICKEAL, PETE BAR 16,38 -2,20 14,18 16,09
JIMENEZ, CARLOS UNI 15,67 4,80 20,47 9,91
HUERTAS, MARCELINHO CAJ 15,46 6,27 21,73 13,36
DEAN, TAQUAN UNI/CAJ 14,83 16,16 31,00 10,74
ANDRIC, LUKSA CIB 14,81 -3,15 11,66 10,90

Sorted by offensive adj. +/-. Surely the presence of the more defensive minded Aldo Curti and Carlos Jimenez is a little surprising.

NAME CLUB ADJ +/- OFF ADJ +/- DEF ADJ +/- TOT STD.ERR
KLEIZA, LINAS OLY 7,19 20,35 27,54 12,00
DIAMANTIDIS, DIMITRIS PAO 6,78 20,20 26,98 9,95
MORRIS, TERENCE BAR 6,62 18,58 25,20 15,82
MORDENTE, MARCO MIL 7,26 16,45 23,71 14,49
DEAN, TAQUAN UNI/CAJ 14,83 16,16 31,00 10,74
VAZQUEZ, FRAN BAR 2,83 16,07 18,90 18,46
LORBEK, ERAZEM BAR 3,19 14,90 18,09 16,50
SATO, ROMAIN SIE 16,76 14,62 31,38 11,58
GARBAJOSA, JORGE MAD 2,85 14,25 17,10 10,93
PEKER, KAYA EFE 6,27 12,89 19,16 11,28

Sorted by defensive adj. +/-. Always thought Linas Kleiza was doing a decent job on defense, especially considering his size disadvantage. Solid, just not spectacular. Dimitris Diamantidis up there, Fran Vazquez and Romain Sato are considered elite defenders too, Terence Morris, Jorge Garbajosa (though not getting any younger) and Kaya Peker known as good defenders.

NAME CLUB ADJ +/- OFF ADJ +/- DEF ADJ +/- TOT STD.ERR
SATO, ROMAIN SIE 16,76 14,62 31,38 11,58
DEAN, TAQUAN UNI/CAJ 14,83 16,16 31,00 10,74
VUJCIC, NIKOLA OLY 16,93 11,06 27,99 13,81
KLEIZA, LINAS OLY 7,19 20,35 27,54 12,00
DIAMANTIDIS, DIMITRIS PAO 6,78 20,20 26,98 9,95
DIXON, BOBBY ASV 16,88 9,59 26,47 20,80
CURTI, ALDO ORL 16,6 9,49 26,09 21,16
MORRIS, TERENCE BAR 6,62 18,58 25,20 15,82
PELEKANOS, MICHALIS MAR 13,10 11,29 24,39 15,01
JAVTOKAS, ROBERTAS KHI 13,28 10,74 24,03 24,46

Sato leading the ranks, Vujcic, Kleiza, Diamantidis, Morris in there as well.

Written by sJacas

September 30th, 2010 at 3:36 pm

  • Emili Paz

    There is no need to purchase tickets for double or triple the face value prices. I have found this great site Ticket2Final, which allows you to purchase options and then decide later on whether you want to purchase a ticket or not, depending if your team makes it to the final four. It is therefore a great way to save money since you are paying only for the actual price!