The Truth Behind The Heat’s Makes And Misses

Shortly after the Miami Heat thumped his Charlotte Hornets 115-103 in Game 2 of the teams’ first-round series Wednesday, coach Steve Clifford was asked what adjustments he had planned for Game 3, with his team trailing by two games. He answered politely but was clearly annoyed by the suggestion that it was up to him to make big changes. Reporters — in general, but in this case also in particular — have no idea what they are talking about, he suggested.

“Jeff Van Gundy always used to say, ‘Writers love to say, they made an adjustment,’” Clifford said. “Usually the adjustment is, some guy that went 1 for 8 went 6 for 8.” He added, “Sometimes the other team just makes shots, OK? So that’s really what’s going on.” You can watch the video below.

At that point in the series, the Heat offense was unimaginably efficient. Game 1 was the most efficient game in the franchise’s history (1.42 points per possession), and through two games the Heat’s effective field goal percentage (eFG%) was 63.4, which was the best ever for any team at that stage of a series.

Clearly, this was unsustainable: Miami was steamrolled in the second half of Game 3 and lost Game 4 Monday night 89-85. But arguments like Clifford’s that NBA games swing on simple makes and misses more often than on hifalutin in-game “adjustments” have often provided cover for crusty old cranks to wave off spreadsheet-wielding poindexters. With the NBA’s newfangled SportVU player-tracking technology, we can finally start to assess arguments like Clifford’s by separating the lucky makes from the good ones, and the good misses from the bad, with a new stat researchers have developed called quantified shot quality (qSQ). That helps us figure out whether a hot-shooting team like the Heat was getting better shots or merely got better for a couple of games at hitting tough shots. In the long run, it also separates the art of getting open shots from that of making the no-no-no-no-YES variety.

Quantified shot quality uses SportVU data to estimate how difficult any given shot is based on factors such as where the shot was taken; the position, speed and angle of approach of a defender; and the average effective field goal percentage for similar shots. In other words, we can tell you if a team is getting wide-open looks like the Hawks’ (Atlanta had a league-best 53 percent qSQ during the regular season, meaning we’d expect average shooters to have an eFG% of 53 on the Hawks’ shots) or if they’re shooting with a rusty bucket over their heads like the Knicks (a 47.8 percent qSQ that, naturally, just barely edged out the Lakers for the NBA’s worst).

We can then apply that to every shot by a team’s offense or against its defense and see not only how good or bad the shots were in a vacuum, but also how well teams did compared with expectation. Over a full season, teams generally stay within 2 percentage points or so of expectation — the Warriors and Spurs were the big and unsurprising outliers, and they came in only 4.1 and 3.4 points above expectation, respectively.

This brings us to Miami’s hot start and the billowing crater the Heat left behind after a disastrous Game 3 and an underwhelming Game 4. On this point, the story the stats tell sounds an awful lot like what Clifford had to say to the press.

Coming into Game 3, the Heat had an unreal 62.7 percent eFG% in two games, but that came on shots averaging a 48.5 percent qSQ (this is bad: not much better than the Knicks, who are very bad) for a difference of 14.2 percentage points. In Game 3, Miami’s raw eFG% fell to 38.6 percent, which is awful and about what you’d expect if you’d watched a second half in which the team scored just 36 points. It was less obvious that the shots themselves didn’t seem to be much different in quality than those of the first game: Miami had a 45.3 qSQ for the game. This time, however, they shot 6.7 percentage points worse than expected, and the Hornets ate them alive. In Game 4, things finally evened out: Miami’s eFG% and its qSQ were an identical 46.7 percent.

Aggregated X and Y coordinates don’t tell you everything that happened on the court. SportVU cameras can’t tell if a smaller defender is trapped on a switch and getting pummeled in the post, or if he’s in the neighborhood but clueless. (Going the other way, I’m reasonably certain the difficulty of a twisting fadeaway jumper isn’t fully captured, either.) Here are a few tough shots Dwyane Wade got to go in Game 1, a couple of which probably look a lot tougher to the cartography software than they did to Jeremy Lin in the film room:

Still, over the course of a season, or a series, teams tend to find their level. During the regular season, the Miami offense was in the middle of the pack, at 50.1 in qSQ, and the Hornets defense was the best in the league at 49 percent. (Charlotte allowed opponents to shoot 0.8 percentage points better than expectation; Miami shot 0.9 percentage points better than expected.) That gets us most of the way to explaining the Heat’s bipolar performance, but “the Hornets have a great defense, and the Heat had to come down eventually” isn’t much of a takeaway from the fanciest stats in the business.

Since Feb. 28, when Joe Johnson played his first game with the Heat, the Miami offense has sprung to life. It’s been shooting 3.7 percentage points better than expectation (still a lousy 50.1 percent qSQ), which puts it right there with the Warriors in both shot-making and overall efficiency. The Heat offense looks different with Johnson than it did with the now-injured Chris Bosh. Bosh is a wonderful offensive player, but he’s at his best as a slow-it-down big man, while Johnson playing as a small-ball power forward opens up more up-tempo plays and space on the floor, especially around the rim.

With the Nets, Johnson was known for taking really difficult shots (he averaged a 45.6 qSQ, which is just impossibly bad for anyone who isn’t Bargs or Kobe) but shot 1.9 percentage points better than expected. On the Heat, he improved to 46.5 qSQ (still butt) but has been 12.2 percentage points better than expected. Small sample size and all, but Johnson averaged 32 minutes and 10.5 field goal attempts per game in his 24 regular-season games with Miami.

So that leaves us not too far from where we started: The Heat offense is good — and lately, very good, even on tough shots — while the Hornets defense has been great all year long. Miami is taking bad shots; Charlotte is forcing them. A lot of things can dislodge result from expectation, like exceptional shot-makers (Steph Curry, Kevin Durant) on offense or players who alter shots with gigantic wingspans and exceptional quickness on defense (Kristaps Porzingis and Anthony Davis both give up relatively “good” shots but hold opponents well below expectation). But in this series, it really seems to come down to whether the Heat make their shots or miss them.

Check out FiveThirtyEight’s 2016 NBA Playoff Predictions.