Why rumors of Kevin Love's poor defense are exaggerated

Tim Kawakami of the San Jose Mercury News did an excellent job of laying out the big picture about the Golden State Warriors' position on trading for Kevin Love, which is worth a read just as a single statement of the Warriors' reasoning in this process.

Bottom line, in his words: "Lee is getting severely under-valued and Love is getting a little over-valued and in their minds, the true delta is less than Klay T's inherent value." In other words, the Warriors aren't going to offer up Klay Thompson in a package for Love even as the competition for the latter's services is only getting tougher, which just increases Minnesota's leverage. And I think Ronaldinho did a solid job describing why the Warriors' position is one that reasonable people could disagree with but not moronic - and I, for one, do disagree, for whatever that's worth.

But piggybacking off of a point made in both the Kawakami and Ronaldinho articles - as well as Sam Amick's article for USA Today last week - one of the concerns fueling the Warriors' reluctance about trading for Love is apparently his defensive deficiencies. Ronaldinho did about as well as anyone laying out the specific concerns, but - for the most part - people have generally sort of accepted that Love is a defensive liability without much evidence supporting it beyond, "Duh."

Love's defense in context

However, ESPN's Kevin Pelton wrote up a pretty detailed analysis of why trading for Love is a no-brainer generally, highlighting both the fact that he's an elite offensive player (not just elite forward) and that he's actually not the terrible defender he's made out to be.

Pelton re-framed Love's defense ability pretty much describing why the numbers don't actually support this notion that he is some massive defensive liability. It's an Insider article, but the relevant words went as follows:

It's possible to highlight stats that showcase Love's defensive shortcomings, particularly as a rim protector. New SportVU player-tracking data on NBA.com showed that last season opponents shot 57.4 percent when Love was within five feet of attempts near the rim, the league's fourth-worst rate among players who defended at least five rim attempts per game.

However, opponent shooting percentage tells only half of the story. Love's reluctance to contest shots also kept him out of foul trouble and opponents off the free throw line. His 1.8 fouls per 36 minutes were the fewest of any regular big man last season (no one else was below 2.0 per 36), and not coincidentally, the Timberwolves allowed the league's lowest rate of free throws per field goal attempt. The trade-off between not fouling and surrendering layups didn't always work out for Minnesota and former coach Rick Adelman encouraged his team to foul more frequently, but looking at opponent shooting percentages without context is unfair to Love.

Pelton ultimately concludes that Love was "an above average defensive player" last season - again, not simply because he's an outstanding rebounder and that matters to creating possessions, but because he's not nearly as bad otherwise as he's made out to be.

Coincidentally, a league source that I have been going back and forth with throughout this entire Love situation relayed similar thoughts just the day before and I'll summarize that conversation, supplemented with the analysis of others from throughout the last year or so (as well as some additional numbers):

• Love is actually a very good one-on-one post defender. The Synergy numbers bear that out, especially in comparison to David Lee: Love gives up .72 points per possession, Lee gives up .88. If you don't immediately grasp the significance between that difference, consider that Love ranked 43rd in the league whereas Lee ranked 156th. As an interesting side note, Bogut was just ahead of Lee at 148th (.87 PPP).
  

• As Zach Lowe has described at length in the past, the problems for Love and the Wolves occur primarily because Corey Brewer and Kevin Martin are "serial gamblers". From Lowe's article in March:
  
  

  Corey Brewer and Kevin Martin, two key wing free agents, gamble their way into crazy fouls all over the floor...These guys are serial gamblers, and a lot of their crunch-time fouls happen before Minnesota’s opponent is in the bonus. But those fouls also put opponents in the bonus.
  
  • Quick opposing point guards can puncture Minnesota’s scheme. The Wolves play a conservative pick-and-roll defense in which Love and Pekovic hang near the paint to corral ball handlers instead of chasing them far from the hoop. It works in the aggregate; neither big is a major plus defender, but they both understand the scheme and approach it with solid footwork.

• And really, "serial gambler" is a generous way to describe what Martin is defensively; long before any Kevin Love trade talks heated up, Kawakami described Martin as "...a horrendous, horrendous defensive player, who doesn't fight through picks and often zones out off the ball, leading to more than a few wide open shots when his man cuts to the basket or across the floor."

• That porous perimeter defense helps to explain why opponents shot so well at the rim - yes, the Wolves give up a high percentage, but Love and Pekovic are also a) put in that situation more often and b) whether they weren't fouling by design or as a tactic to stay on the floor, they were one of the 20 least foul prone teams in NBA history. Put in simplest terms, playing the Wolves was like getting free runs at the basket if you had capable perimeter players. And that didn't favor either Love or Pekovic, who neither operate well in space nor protecting the rim.
  

• So as a quirky by-product of that system alluded to by Pelton and borne out by the numbers, even the Wolves' opponent field goal percentage overall is a little overstated: they managed to limit the number of points surrendered by allowing the fewest opponent free throw attempts in the league last season despite playing at the fourth-highest pace in the league.

How does this matter for the Warriors presumed interest in Love?

None of this is to negate the defensive critiques of Love that Lowe also summarized more harshly in June. But it is to say that a more nuanced look at the context of Love's numbers do in fact support Pelton's summative claim that he was an above average defender or, perhaps better put, not as bad as the surface level numbers make him look.

All of that helps to explain Love's positive D-RAPM as some less of a rebound-inflated illusion and far more as a real phenomenon that at most reflects the low quality of the defenders surrounding him.

Cobbling all those very complementary insights together, you have two takeaways in imagining Love's defensive fit with the Warriors:

• I'm going to go out on (a very sturdy) limb and say that the Warriors' combination of Andre Iguodala, Draymond Green, and Shaun Livingston are all better perimeter defenders than the majority of the Wolves perimeter defenders, but specifically and especially Brewer and Martin.
  
• A healthy Andrew Bogut is a better frontcourt partner than Pekovic defensively and Festus Ezeli could very well be one as well once fully recovered from injury.
  
• The combination of better perimeter and interior defenders should leave Love less exposed in positions of weakness.

So now imagine Love surrounded by better perimeter defenders and a high IQ, veteran post defender like Bogut. Immediately, the Warriors' interior defense improves due to Love's ability to defend post players one-on-one and secure defensive rebounds (which is a valuable contribution to a team's effort to keep an opponent's points per possession low).

While the theory that Bogut would cover for Lee's deficiencies didn't quite pan out, Love is actually a much better complement as a better post-up defender than either of the Warriors' starting posts last year. With Draymond Green also in the mix as a versatile defender, the improvement in the Warriors' interior defense would actually benefit the Warriors' defensive unit as a whole. That matters even more given concerns about the health of Bogut, Ezeli, and Jermaine O'Neal (should he even return) in the post.

How much would Thompson's loss hurt the defense?

There is certainly a trade-off there that's not to be dismissed: as Evanz suggested last week, the gains in the post (on both ends) would be offset somewhat by the loss of Thompson on the perimeter (on both ends). And there's certainly value to the notion of just keeping the team together and growing together independent of how you evaluate that trade-off.

But SI's Rob Mahoney made a good point about Thompson in his article last night exploring the dynamics of a hypothetical Warriors' bid for Love that would have to include Thompson: looking at Thompson as a good defensive complement to Stephen Curry actually, "...lends itself to argumentative warping, namely to the end of claiming that Thompson is a better situational defender than the evidence can support."

Mahoney focused on Thompson's struggles against the elite guards in the league, but a close look at the situational statistics also suggests that Thompson is not only not the elite defender he has been framed as but also not that much better than Shaun Livingston defensively. The one major advantage Thompson has over Livingston is defending spot-ups, where he yields .88 PPP to the latter's .98 PPP. That's significant and also possibly the result of Thompson defending more isolations while Livingston defended more spot-ups - Livingston was especially poor on that and put in that situation 30% of the time.

But if we're comparing Livingston's fit next to Curry, the drop-off from Thompson is considerably less severe: in pick and rolls, Livingston (.82) was about even with Thompson (.81) and in isolations neither Thompson (151st) nor Livingston (192nd) ranked in the top 150 in the league. In other words, yes, Thompson would be a loss defensively but neither so big that it can't be overcome nor bad enough to offset the considerable gains found by adding Love.

Context matters

To Pelton's point, context matters when looking at statistics - basketball, despite the increasingly availability of quantitative data to help the average person analyze the game, is still not a math problem that can be solved simply. We can quantify certain events in isolation and even a player's contribution to the outcome of those events to some extent, but really getting a sense of how all those discrete events come together requires digging a bit beneath the surface. That's at least doubly true when talking about defense, which has a ton to do with not only individual ability but also a team's philosophy and a unit's commitment to it - even in a NBA world dominated by individual play, defense is still not a matter of "me vs. you" but "us vs. them". It's not as simple as adding up single number metrics to determine defensive fortitude, but a matter of fit.

In this case, when you look at the whole picture, the benefits of Love bringing better spacing and having a complementary scoring option for Curry offensively in addition to the actually significant defensive upgrade in the post could very well make the Warriors a better team. The cost, of course, is Thompson's perimeter defense but the Warriors are much better equipped there than the conventional wisdom suggests and possibly more in need of help on the interior than some would suggest.

It can be really difficult to project how those things will come together, especially when so many other aspects of the situation are already uncertain (e.g. Can Steve Kerr coach? How good is this team at mostly full healthy?), but - unless you strike gold in the lottery - you have to take risks in order to improve to championship level. And the defensive risk of adding Love at the expense of Thompson is a worthy one once you get past assumptions and accepted defensive reputations.