You watch the games. You can see what’s going on out there. You may not know (or care) about what your favorite player’s True Shooting Percentage (TS%), Player Efficiency Rating (PER) or any of those other newfangled numbers say, you just know that when he is on the floor, the team plays better than when he isn’t.
I think you may just like the plus/minus (+/-) statistic.
Basic (Single Game) +/-
The basic mechanics of +/- are even easy to understand. Here’s a one-game example. “Your guy” starts the game and plays until the midway point of the 2nd quarter. When he leaves, your team is up by 5 points. He’s at +5 (told ya it was easy). He comes back in to start the 2nd half with the score tied and he plays the full 3rd quarter. Your team regained the lead and is up by 3 points. Your guy is now at +8. He re-enters the game with 5 minutes left and your team up by 1. He plays the rest of the game and your team ends up winning by 2 points. You grab a beer and a cigar and toast your team’s victory and your guy’s +9 performance.
If you don’t feel like keeping track of +/- in a game, you can just go to nba.com. A couple years ago they began including the stat in their box scores.
The obvious problem with the +/- stat is that a player doesn’t necessarily have to do anything good on the court to end up on the positive side or bad to end up on the negative side. The +/- stat is only a reflection of what happened when you were out there, and when you compare it to the final score, what happened when you weren’t out there. In a single game, there are often players who record what look to be significant +/- numbers who were nothing more than innocent bystanders. However, when you begin to look at +/- over many games, like say a full season, the results become more credible.
Basic Unadjusted (On-court/Off-court) +/-
To get at +/- over longer periods, the stats guys keep track of how many points a team scored and allowed per 100 possessions when each player was on the court and how many that team scored and allowed per 100 possessions when the player was on the bench. There’s going to be a little math coming up here, but hang in there…it’s only going to be addition and subtraction.
Example: the Celtics’ Paul Pierce 2010-11 Season thru 3/9/11
When Pierce has been on the court this season, the Celts have scored 112.24 points per 100 possessions. When he’s been on the bench, the team has only scored 96.37 per 100. This gives him an offensive net +/- rating of +15.88. On defense, with Pierce on the court Boston has given up 99.90 points per 100 versus 102.99 when he’s on the bench for a net defensive +/- rating of -3.09 (on defense, negative is good). To get Pierce’s overall net +/-, you subtract the net defensive rating from the offensive rating (15.88 less -3.09 equals +18.97).
If Pierce’s +18.97 sounds very good, it’s because it is. He currently leads the league in overall +/-. Rounding out the current top 5 are the Mavs’ Dirk Nowitzki (+17.74), the Suns’ Steve Nash (+17.58), the Trailblazers’ LaMarcus Aldridge (+17.43) and the Celtics’ Kevin Garnett (+15.33). One of the tests of any overall-type performance statistic is to look at the league leaders and ask “Are these guys great players?” As I see it, the overall +/- stat passes the test…all of the top-5 are great players having great seasons. Other notable +/- leaders include the Hornets’ Chris Paul (+12.42), the Heat’s Lebron James (+9.69) and the Magic’s Dwight Howard (+6.23).
About now you may be saying, “Hey! you missed some totally stud players! Where’s Kobe? Wade? Amare? What about Kevin Durant, Derrick Rose and Carmelo Anthony?” Well, up ‘til now I had only been talking about the top +/- guys. None of these players make the top-50.
For Kobe, Wade, Amare and Rose, all have modest, but overall positive +/- ratings for this season and it’s clear from the +/- numbers that all are outstanding offensive players. Their teams score between 7.29 (Wade) to 10.30 (Rose) more points per 100 possessions when they’re on the court than when they’re not. The problem is that their teams hold opponents to fewer points per 100 possessions when these players are on the bench than when they’re in the game, dragging down their overall ratings. Does this mean that they’re bad defensive players? That’d be a tough sell for Bryant or Wade since both have made multiple NBA All-Defensive teams. Amare has never been known as a plus defender and the fact that the Knicks give up a whopping 110.42 points per 100 possessions (yes, that’s a lot) when he’s in the game is further evidence of his relative defensive shortcomings.
Rose is the most interesting case of this group. His Bulls currently lead the NBA in most defensive categories. Their second unit is particularly effective and features the NBA’s #1 defensive +/- performer in guard Ronnie Brewer (-8.50…and remember, for defense, negative numbers are good). Their second unit’s center, Rookie Omer Asik has an ungodly defensive +/- rating of -10.73, but doesn’t quite have enough minutes played to qualify for the league lead. When Rose is on the court, opponents have scored 101.95 points per 100 possessions. This is not only respectable, it’s the 23rd lowest on-court defensive point total in the NBA. However, when Rose is on the bench, Bulls’ opponents score an absolutely microscopic 92.35 points per 100, giving Rose a positive (bad) net defensive rating of 9.60. As a result, when you subtract the 9.60 defensive rating from his 10.30 offensive rating, he ends up with a very pedestrian +0.70 overall +/-.
Since Anthony was just recently traded to the Knicks, most of his +/- data relates to his time with the Denver Nuggets. Among superstars, Anthony’s +/- numbers this season are unusual in that on both offense and defense, his presence or absence hasn’t seemed to make much difference. His offensive net +/- is 1.76 and his defensive net is -0.25 for an overall +/- of +2.01.
The Oklahoma City Thunder’s Kevin Durant is the most surprising +/- case I’ve come across, though when you break it down, it’s similar to what we saw with Rose. Durant, only 22 but in his 4th NBA season, is an undisputed uber-star. He was named All-NBA first team last season and led the league in points scored. This season, he again leads the NBA in scoring. While fans and the media just love Durant, the +/- stat decidedly does not. His net offensive +/- is a very good +6.61. On defense, it’s another story. When Durant is in the game, the Thunder’s opponents score 109.37 points per 100 possessions…not quite Amare territory, but close. When Durant’s on the bench, opponents only score 101.69 per 100 making his net defensive +/- +7.68. This gives him an overall +/- rating of -1.07.
In fairness to Durant, he is consistently among the league leaders in minutes played, so the off-court numbers lose some of their meaning (there just aren’t that many of them). His +/- rating notwithstanding, the Thunder would have no interest in life without Kevin Durant.
Durant and Rose are good examples of one way the +/- stat can lead you astray. The Celtics provide another. Earlier we saw that the Celtics’ Paul Pierce is the current NBA unadjusted +/- leader. His teammate Kevin Garnett is #5. Their point guard Rajon Rondo is #19 and shooting guard Ray Allen just missed making the top-50 by 0.02. These guys are an ensemble and they’re an ensemble that’s consistently better than the competition they play. Their second unit isn’t nearly as strong. They’re bound to have positive +/- ratings and it’s difficult to isolate the individual’s relative contribution to the group’s achievement. For better or worse, the stats wizards have taken a shot at addressing this issue.
Adjusted +/-
I’m not going to spend a lot of time on this, but I suppose you ought to know this beast is out there. So how do the statisticians level the +/- playing field to attempt to better isolate individual excellence? Here’s their explanation of the adjusted +/- stat:
“Adjusted +/- ratings indicate how many additional points are contributed to a team’s scoring margin by a given player in comparison to the league-average player over the span of a typical game (100 offensive and defensive possessions). These ratings are considered “adjusted” since they start with the simple +/- rating and apply a regression model as outlined by Dan Rosenbaum to adjust for the impact of all other players on the court.”
Don’t worry if that description made you go “Huh?”…it’s a perfectly natural reaction. Statisticians can’t be expected to be both brilliant and intelligible…it’s just unreasonable. Essentially, what they do is statistically make a given player’s teammates and opponents more like average players and then see how that player comes out.
The current NBA top-5 for adjusted +/- this season? They’re Aldridge (16.84), Nash (14.01), Paul (13.17), Nowitzki (13.12) and Howard (12.98). As you’ll recall, Aldridge, Nash and Paul were also in the top-5 for unadjusted _/- (Paul was 6th). Like its unadjusted parent, the adjusted +/- stat also passes the top-5 “eye test.” Garnett dropped to #7 and Pierce to #9, showing the leveling effect (Rondo is not in the adjusted top-50). Lebron is at #13, Rose #15, Durant #45 and Anthony #50 (4.32). Wade and Amare didn’t make the top-50, but both have a positive adjusted +/-.
The Lakers Kobe Bryant is the Durant of adjusted +/-, coming in with a -6.99. This isn’t merely bad, it’s 27th worst in the league! It’s at this point that the statisticians would quickly explain to you that in order for the results of their +/- adjustment to be accurate, very large sample sizes are needed. Obviously, in Kobe’s case, ¾ of a NBA season is way too small.
In fact, the statisticians feel that a minimum of 2 full seasons of data is needed for adjusted +/- to be valid. This is a problem for me and a key reason why I’m only giving it a brief mention. In sports, each season is unique. Players change teams and new coaches bring in new systems. Last season’s darlings can be this season’s dogs (and vice versa). Sports is a “what have you done for me lately” environment. Sports simply doesn’t have the kind of patience necessary to appreciate the technical elegance of a statistic that requires two years to properly develop its bouquet. With apologies to the propeller-heads, somebody had to say this.
As I’ve said before, there’s no perfect basketball stat. This said, I’m a fan of the plus/minus stat. Used as a supplement to the other individual statistics, it can help confirm excellence. It can also help identify those players who may not put up gaudy individual numbers, but do the little things that help a team succeed…and as long as you stick to the unadjusted version, it probably won’t make your head hurt.
No comments:
Post a Comment