This week’s mailbag features your questions on shooting differentials from last season, Toronto’s achievement level, and more.
“Are NBA referees protecting shooters more than drivers?”
— LeBron J., Cleveland
“We’re at a point now where we protect the shooter more than the driver,” James said. “There’s no reason I should be going to the line four times in a game when I drive 100 times to the paint and I’m getting hit and slapped and grabbed and whatever and whatnot. We protect the shooter. That’s what it’s turned into. ‘Chicks dig the long ball,’ and that’s what it’s about.”
Let’s take a look at LeBron’s contention using Second Spectrum data. I calculated the rate of shooting fouls to all attempts to shoot — including shooting fouls, which aren’t technically shot attempts — both inside the restricted area (where most shots of drives come) and beyond the 3-point line (excluding heaves). Naturally, shooting fouls are far more common in the restricted area, but I really wanted to compare how the two rates have changed over the five years for which we have complete tracking data:
As you might suspect, the rate at which players are fouled on attempts from 3-point range increased dramatically from 2015-16 to 2016-17 — 27.5 percent higher last season, a change my former colleague Tom Haberstroh wrote about during the NBA Finals. It has gone back down a little this season, presumably because the NBA made one of its points of education the difference in definition of when a player is in the act of shooting on a jumper as opposed to a shot off a drive.
At the same time, after declining slightly in recent seasons, the foul rate on attempts in the restricted area in 2017-18 is actually the highest it has been during the camera-tracking era. The kicker is that James’ own rate of fouls on attempts in the restricted area (19.5 percent) is the second highest in this span. So while it’s certainly possible that James is also getting hit much more often on shots around the rim, the evidence suggests his complaint was more frustration than legitimate problem.
I’ve watched ever cavs game this year and it looks like teams are hitting ridiculous shots at a ridiculous rate (heavily contested shots) any stats to back this up #peltonmailbag
— zach pruhs (@devil7923) February 23, 2018
Speaking of Cleveland, I’m not sure I would say that opponents are displaying ridiculous shot-making against the Cavaliers. Second Spectrum tracks shot difficulty using the quantified shot quality (qSQ) metric, and Cleveland’s defense ranks dead last in the league there at 52.7 percent — the effective field goal percentage (eFG) we’d expect opponents to make given the location of those shots, their type and nearby defenders.
Now, opponents are in fact shooting a better eFG against the Cavaliers (53.9 percent), so there does appear to be some good shot-making against them. But 10 teams have a larger discrepancy between their qSQ and eFG, a metric Second Spectrum calls qSI (quantified shooter impact).
Looking specifically at contested shots using Second Spectrum data on NBA Advanced Stats doesn’t reveal much more in the way of ridiculous shooting against heavily contested shots. On attempts for which the closest defender is between two to four feet (called “tight” contests), opponents are making 27.0 percent of their 3-pointers against Cleveland, the league’s sixth-lowest mark. The Cavaliers’ bigger problem is that they’re giving up a league-high 17.3 3-point attempts per game with no defender closer than six feet (called “wide open”).
“Why was RPM’s projection for the Toronto Raptors so wrong? Did RPM just love P.J. Tucker, Patrick Patterson and Cory Joseph so much? (Because I do too.) Is it because it didn’t predict the growth of the bench?”
The projection by ESPN’s real plus-minus (RPM) that Toronto would win an average of between 44 and 45 wins this season, the sixth-best projection in the Eastern Conference, has been a source of much discussion recently in Raptors Twitter. As is typically the case when a projection is so far off, I think there are several factors at play.
You’ve identified a couple of big ones relating to the bench. While Joseph and Tucker didn’t rate appreciably better by RPM than their replacements, Patterson’s loss looked like a huge hit to the Toronto bench. In 2016-17, the Raptors were about a break-even team with Patterson on the bench (plus-0.9 points per 100 possessions, according to NBA Advanced Stats) — and that was actually an improvement on getting outscored with Patterson sitting during 2015-16.
Patterson and Kyle Lowry got most of the credit in RPM for the success of Toronto’s highly effective bench lineups the past couple of seasons, so the Raptors’ revamped second unit projected as average at best. Instead, with Pascal Siakam developing into a capable replacement for Patterson, and Fred VanVleet emerging as the new plus-minus standout, Toronto’s second unit has been better than ever despite that Lowry has played fewer minutes with this group.
Also interesting is the improvement of the Raptors’ starting five. Their most common four-man lineup last season (Lowry, DeMarre Carroll, DeMar DeRozan and Jonas Valanciunas) outscored opponents by just 1.7 points per 100 possessions, according to NBA Advanced Stats. This season, with rookie OG Anunoby in place of Carroll, and Ibaka solidifying the power forward spot, Toronto’s starting five has a robust plus-12.1 net rating.
For all the talk of the Raptors’ offensive changes, that improvement has come largely at the defensive end. Toronto’s starting five has a 100.7 defensive rating, better than the Boston Celtics‘ league-leading mark this season (101.1). Anunoby, projected to play limited minutes before the season, has been a big factor in that defensive success.
The other question here is why the Raptors have continually outperformed their RPM projections. As I noted over the summer, Toronto has beaten them by an average of six wins over the past four seasons, which will only go up this year. I suspect part of what’s happening is that the multiyear version of RPM used to build projections incorporates playoff performance, and naturally, the Raptors haven’t been as effective there. We’ll see whether their overperformance can carry over to the postseason this season.
Wassup Can you direct me to somewhere that I can find who has seen the biggest jumps in 3PT% (NBA)?#PeltonMailBag
— KidKash202 (@KidKash202) March 3, 2018
Here they are, minimum 100 attempts both seasons.
And, for comparison’s sake, the players with the biggest declines this season:
To state the obvious, these lists are heavily governed by regression to the mean. Every player on the second list shot better in 2016-17 than every player on the first list, in most cases, dramatically so. As a result, odds are both lists overstate the magnitude of any actual change in players’ shooting ability.
That said, it’s encouraging to see second-year lottery picks Brandon Ingram and Dragan Bender making strides as 3-point shooters after disappointing in that regard as rookies. Improved 3-point shooting is also a big part of Kelly Oubre Jr.’s development, and Trey Lyles has gotten back to the kind of 3-point shooting we saw from him as a rookie.
On the other side, the Wizards expected they’d get capable 3-point shooting out of Meeks, a career 37.3 percent shooter entering this season. Instead, he has made 3s at a career-low rate, a tough break for a team whose depth is limited by luxury-tax concerns.