Continuing the quest to bridge the gap, another edition in the Hardwood Paroxysm series of Understanding Advanced Stats
One of the hardest things to quantify in basketball is the value of individual defense. The eyeball test can at times be as reliable as statistics here, but it can also fool you all the same. Remember that although it helps, “intensity” doesn’t automatically equal good defense (looking at you Mr. Kobe Nine-Time All-NBA Defensive First Team Bryant).
The standard blocks or steals per-game numbers are also very misleading — a good help defender can net dozens of rejections while being a poor man-D and/or rotation-D defender, and no one in their right mind would accuse steals leaderboard regular Monta Ellis of being a good defender. Even Chris Paul, who regularly leads the NBA in steals, is something more a gambler than an actual lockdown defender as many assume when browsing numbers.
Having a preconceived notion of a player’s abilities due to the standard accepted old school stuffed box score of stats, or a prominent analyst’s opinion, can then skew the viewer creating a bias in the mind’s eye. Because individual defense is so difficult to measure in the NBA we often make assumptions and excuses for particular plays or players. Reputations often rule masses, single spectacular plays net new contracts (has anyone made more money and gotten more burn off of a single clothesline than Raja Bell?).
Unlike on offense, there is no one place to go to get a fairly definitive defensive feel of a player — it really takes trips to a few different sites before we have enough information assimilated to even begin getting an accurate feel for a player on the D end that we can then take with us to the game.
Let’s start by taking a look at an All-NBA Defensive First Team selection from last season, one Kobe Bean Bryant, at BasketballReference. As you may recall from previous posts here at HP, lower is better for DRtg (defensive rating).
Kobe actually looks pretty on par by this rating, right at his career average. Let’s dig deeper, next stop 82Games, and opponent PER (player efficiency rating).
While PER isn’t universally accepted as an accurate offensive measure, it lends itself as a very useful tool as a piece of the defensive stats puzzle — there isn’t, after all, “opponent win shares,” or “opponent wins produced” — and while imperfect, it does have a place in furthering our quest. Keep in mind that we are always wary of small sample sizes.
Bearing in mind that 82Games says “all stats reflect assigned responsibility to a player” when positional assignments are charted, Kobe looks pretty good here as well (15.0 is considered the “average NBA player” PER). As with DRtg, opponent PER can’t be trusted as a stand alone D-stat — there’s just too many variables, such as maybe his counterpart simply stinks at shooting the ball, a big portion of the PER formula, or that a team’s D-scheme may include a couple of towering seven-foot beasts that tend to force opposing guards to want to take a lot more lower-percentage long twos than they normally would, etc. Just because Kobe is assigned an offender on paper doesn’t strictly mean he defensed that player on a particular play.
For comparison, since we brought him up, Chris Paul’s opponent PER
My eyeballs have been telling me a different story concerning Kobe for some time, so since we’re seeking truth and not confirmatory bias, and knowing now that D-stats are difficult to quantify, we will continue on to another metric to see if it again holds up.
Over at mySynergySports we find dedicated cameras recording every play by every player in every game, then categorized by type. The result is a comprehensive points-per-possession rating, otherwise known as PPP, with accompanying video.
Note: While quite comprehensive, the calculation of PPP, as with any stat, does have it’s flaws as well, one of which is the inclusion of turnovers as a negative mark. This has the unfortunate effect of having a tendency to rate high usage guards, such as Kobe and point guards, lower offensively since they handle the ball much more, therefore turning it over more often, generally speaking. But that has no effect on defensive stats other than that maybe a particular player may have a penchant for floating around in passing lanes or anticipating passes, like say, Kobe or Chris Paul. While this is technically defense, it’s not exactly what one peruses when pursuing defensive measures of a player.
Giving up 0.89 PPP overall isn’t awful. It’s not great either, especially when one considers that of the approximately 450 players in the NBA at any given time only about 300 or less are getting significant playing time (Kobe ranked 216 overall on defense, as you can see).
Kobe is good at isolation D and quite often guarded the opposing point guard, as we can glean from his second-largest sample size of 24.1% on the P&R (pick-and-roll) Ball Handler (duh, Derek Fisher. If I was Phil I’d have put Kobe on ‘em too). The best way to expose Kobe on defense is to come off a screen (he doesn’t like chasing guys around. At all.), or make your spot-up shot, where he allowed 1.01 PPP, which was by far the most used method to take advantage of the aged wonder at 40% of the time Kobe defended opposition.
You see, Kobe will play off you, dare you to shoot, because he just loves getting in those kinds of contests where he excels. They get him goin’. He thrives on ‘em. What he doesn’t thrive on is playing defense. Sure, he’ll gnash his teeth, Â jut his jaw, talk smack, and stare you down something serious, but when it’s time he mostly just likes to occupy a spot to one side of the free throw line and take jabs at the ball as it passes in the paint. If he has to, he’ll offer a token couple of steps at a close-out with one hand up at the jump shooter. He’s not wasting precious offensive energy playing defense these days.
This is typical of a Bryant defensive setup. He’ll mostly just stand right about there waiting for the ball to come back to him on offense.
Kobe’s man here is CJ Miles, whom Bryant is playing way off of. Bryant has little or no respect for opposing shooters (notice he has his back to his man, granted Kobe knows his scouting reports and all, but this is also typical of Bryant on defense regardless), opting to occupy a space he doesn’t have to move much more than a couple of steps at.
Miles sees Kobe’s manner of defense on him and will astutely take advantage of it a few possessions later. No, not with an ill-advised three. Watch.
Again, same basic setup.
Paul Millsap has the ball, and Kobe’s full attention, inexplicably, on the low block. Pau Gasol has had a little trouble with Millsap this night, but Ramon Sessions is right there ready to cheat down on a double if necessary. Yet Kobe floats over anyway, and as he does he opens up a huge lane that Miles recognizes.
As Miles makes his move Kobe continues to gravitate toward the ball, y’know, cause it’s Kobe and all. He can’t help it. Notice the passing lane Millsap now has thanks to Kobe as Miles cuts to the paint.
You could drive a Kia through that passing lane and a Mack truck through that cutting lane. Millsap takes full advantage of his underrated court vision after pulling three of the floor’s five defenders to him and dishes off to the cutting Miles for an easy layup attempt.
No, Bryant doesn’t always make it this easy on opposing players to get clean looks, but he rarely challenges with more than a token effort. If you can shoot even a little, Bryant will let you try.
Watching every defensive spot-up play defended by Bryant in last season’s playoffs, Kobe’s man made 23, missed 26, and fouled three times. Assigning one point per foul, Bryant gave up 49% shooting to his man in the playoffs last year.
Perspective? Only three qualified shooting guards in the NBA last season made at least that from the floor, Dwyane Wade, Arron Afflalo, and Ray Allen, the latter two extremely good mid-range-to-long shooters by perimeter player standards (here’s your cue to go check em out at HoopData). After that the drop-off is pretty steep to DeMar DeRozan and Kobe himself, at .467 and .451 respectively. The median qualified shooting guard last season shot 44% from the floor, so Kobe gave up at least 5% more than he should have in the playoffs with his “defense.” To guys like Trevor Ariza, Marco Belinelli, Peja Stojakovic, and Jason Kidd, who together averaged .405 from the floor in the regular season.
And yes, he gave ‘em plenty of room to do it in. Ring up another victory for advanced stats.
_____
• Chris Paul’s 2010-11 Synergy defensive numbers
• Kobe’s current Synergy defensive numbers for the 2011-12 season. Interesting to note here that while spot-up and iso numbers remain about the same we see a pretty big drop-off in his ability to cover the ball handler this year. And while his overall rank is better this year than last, that’s as likely attributed to a league-wide fall-off in FG% due to the shortened season as anything else
• Handy sortable “simple ratings” from 82Games already sorted for you by opponent PER (remember, click the column heading)
• Dozens of sortable team and player stats by position from HoopStats based on efficiency differential (which they call Diff. Eff. Same thing), yet another metric you can use to fill in a piece to the D-puzzle
• A little something I’d like to see revisited and updated from Rohan Cruyff, defensive pace factor























 