The Utah Jazz strode confidently into Denver on the last day of January. They boasted an 11-game win streak and the best record in the NBA; they’d recently beaten the Bucks, the Warriors, and the very Nuggets they were set to play that day.
Little did they know they were walking into a flood. The Nuggets deluged the scorching Jazz with triple after triple, taking a 25-point lead into halftime thanks to 15-for-17 shooting on 3-pointers. Their first-half shot chart is hilarious.
Even with some second-half misses, Denver finished the game 18-for-28 from distance—the best single-game performance for any team this season. Utah simply couldn’t compete with that kind of marksmanship, and needed to start a new winning streak in its next game. The Nuggets’ outlier afternoon is the only reason the Jazz haven’t now won 17 games in a row.
It’s not just the Jazz, or one game. The NBA is taking the “make-or-miss league” maxim to an extreme, injecting increased randomness into an already scattered season. More than ever before, 3-point defense is determining who wins or loses games—except teams don’t really know how to defend 3s at all.
Consider the Bucks, Utah’s opponent Friday in a matchup of the NBA’s top two teams by point differential. Milwaukee is 0-8 this season when its opponents shoot 44 percent or better on 3s. But when opponents don’t hit that mark, the Bucks are 16-1.
In other words, you don’t need to watch Giannis Antetokounmpo dunk, or Khris Middleton splash midrange jumpers, or Brook Lopez wall off the paint. As long as you know how the Bucks’ opponent shot from 3, you don’t need to know anything else about the game to intuit the outcome.
Let’s examine two buckets of games: outlier 3-point performances on the high end, in which teams make at least 50 percent of their 3s, and outliers on the low end, in which teams make just 20 percent of their 3s or fewer.
Teams on the high end are 48-5 this season (a 91 percent win rate), and one of those losses came in a Clippers-Bulls game in which both teams exceeded 50 percent. Teams on the low end, meanwhile, are just 1-21 this season (5 percent win rate). Both of these figures would easily set records.
The same story appears with point differential, where high-end teams have won their games by an average of 15 points. Low-end teams, conversely, are minus-18 on average. A higher percentage of 3-pointers taken will naturally result in 3-point accuracy having a larger impact on performance, but teams also haven’t increased their 3-point attempt rate so much this season to account for such a steep shift. (There are fewer low-end teams this season because the league is more accurate overall—more on that data point in a moment—but the same pattern appears when considering teams at 25 percent or below, instead of 20.)
Crucially, defenses can’t do much to force opponents into the red lines on those graphs and avoid the blue ones. There isn’t any strong evidence that teams can consistently coax poor shooting nights from distance. One way to examine this possibility is to split the season in two and compare how a team performs in its first 41 games versus its next 41 games; a team with this ability should be able to maintain it across both halves. Yet looking at the period from 2012-13 through 2018-19 (to avoid the lockout on one end and the pandemic-shortened season on the other), the correlation between first-half 3-point opponent accuracy and second-half 3-point opponent accuracy was only 0.12, on a scale in which 0 means no relationship and 1 means a perfect relationship. That’s a negligible result.
(For comparison, the correlation between first- and second-half 3-point accuracy for a team’s own shooters was a considerably stronger 0.42. Teams have a lot more control over their own offensive performance, versus their opponents’.)
As an illustrative example, the team that defended 3-pointers best in the first half of a season in that span was the 2014-15 Trail Blazers, whose opponents shot just 29.7 percent from 3 in the first 41 games. But in the second half of the season, that percentage shot up to 37.4 percent, well above average.
Conversely, the 2016-17 Mavericks had the worst first-half mark, allowing opponents to shoot a blistering 40.3 percent on 3s. In the second half of that season, Dallas was perfectly average at 3-point defense (35.8 percent).
So the league is shooting more 3s than ever, and 3-point performance is more meaningful than ever, and teams can’t actually do much to influence their opponents’ accuracy. Those factors all combine to produce wild scoreboard swings from game to game.
Indeed, a bunch of 2020-21 teams rank among the most inconsistent squads in NBA history (as measured by the standard deviation of their point differentials in the first 25 games). The Warriors are the fourth-most inconsistent team ever through 25 games, and the Clippers, Rockets, Mavericks, Bucks, and more aren’t far behind.
Most Inconsistent Teams, 2020-21
|Team||Percentile (in NBA History)|
|Team||Percentile (in NBA History)|
Absences related to COVID-19 certainly contribute to this atmosphere of uncertainty. As writer Owen Phillips noted this week, a strong predictor of over- or underperformance so far is simply how many starting lineups a team has needed to use.
But almost any surprising single-game result this season has made much more sense when examining each team’s 3-point output. The Knicks blew out the Bucks? Well, the Knicks shot 59 percent on 3s while the Bucks shot 18 percent. The Cavaliers beat the Big Three Nets? They made 50 percent of their 3s. The Clippers lost in record fashion to the Mavericks? They made just 12 percent of their 3s (4-for-33, including 1-for-19 in the record-setting first half).
This trend is especially important when considering the league’s boosted accuracy from distance. NBA teams are shooting a combined 36.9 percent on 3-pointers, which would be the best mark ever if it lasted for the full season. Accuracy on wide-open 3-pointers in particular—which comprise about half of all attempts—has improved, per NBA Advanced Stats, from the 38 percent range in recent seasons to 39.6 percent so far this season.
There are too many new variables in this strange season—a lack of fan distractions, more space to maneuver on the sidelines without courtside tickets, no going out the night before games—to isolate one specific cause. But it’s possible that defenses designed to combat the NBA’s 3-point direction have themselves contributed to this improvement. “More and more teams have gone to drop schemes,” The Athletic’s Seth Partnow noted recently, referring to the defensive style in which the big man sags on pick-and-rolls to protect the rim, “meaning players get many more reps against that scheme. … Familiarity is breeding if not contempt, at least comfort.”
This kind of dynamic isn’t unique to the NBA. In MLB, the number of 100 mile-per-hour fastballs increased by about 50 percent from the first half of the 2010s to the second half of the decade, and batter performance improved against those once-nigh-unhittable pitches as they gained more practice. In the NFL, the Dolphins’ wildcat offense took the league by storm—but only briefly, because as other teams started copying their success, defenses grew familiar with the special formation and developed schemes to combat it.
It’s possible, then, that notions of how to play defense in the modern NBA have flipped. Last season, the Bucks and Raptors allowed two of the highest 3-point attempt rates in the league, but they were the two most efficient defenses because they controlled the paint and weren’t hurt by those extra long-range tries. This season, however, as more teams have followed the Bucks’ lead, the teams that have gone the other way to allow fewer 3-pointers have generally been the best defenses overall.
This Season’s Best Defenses Prevent 3-Point Attempts
|Team||Opponent 3PAr||Rank||Defensive Rating||Rank|
|Team||Opponent 3PAr||Rank||Defensive Rating||Rank|
As I noted last month, the correlation between 3-pointers allowed and defensive rating was negative last season, so that allowing more 3s meant a better defense across the league. This season, that correlation is positive, so allowing more 3s means worse defensive performance.
On a game-to-game level, 3-point extremes make for odd scorelines and offer an important reminder to not overanalyze a single game. The Bucks, for instance, lost to the Lakers at home last month—a theoretically worrying result for Milwaukee in a potential Finals preview. But the Lakers shot 19-for-37 (51 percent) on 3s in that game, with Kentavious Caldwell-Pope canning seven of 10 tries. Presumably the Lakers wouldn’t hit more than half of their 3s throughout a playoff series—and if they did, Milwaukee probably couldn’t do much to stop that barrage, anyway.
But the uptick in leaguewide 3-point accuracy goes beyond impacting single games; it could change broader strategy. The Bucks’ bet under coach Mike Budenholzer is that it’s worth sacrificing a few extra 3-point attempts to strengthen paint defense, but the margins to make that bet pay off are mighty thin. An additional point or two of 3-point accuracy could flip that calculation. If this level of accuracy is the new normal, teams might have to adjust how they defend the perimeter.
Or at least they have to hope their opponents miss all the open looks. As loath as we might be to admit it, NBA success is dictated by random shooting luck.
Thanks to David Corby of Basketball-Reference for research assistance. Stats through Wednesday’s games.