The Boston Celtics Broke 3-Point Defense

How the Celtics have defied the odds for two decades

There is one hard and fast rule in basketball analysis: opponent 3-point defense isn’t a repeatable skill. Last season, the Oklahoma City Thunder led the NBA at 34.2% opponent 3-point shooting. This year, they’re 23rd at 36.6%. And in both seasons, they led the NBA in defensive rating: two seasons, same great defense, wildly different opponent 3-point shooting. However, the problem with rules is that they can always be broken, and the Celtics have been breaking this one for two decades.

Starting in 2005-06, the Celtics embarked on an unprecedented and near-uninterrupted run of opponent 3-point shooting excellence. Outside of 2020-21, the Celtics have held their opponents below the league average 3-point efficiency in every single season. Despite conventional wisdom, the chances of this being happenstance are incredibly remote. You simply don’t get that lucky for that long. The Celtics know something that we don’t, and I think I’ve uncovered a few answers.

The Beginning

While the Celtics’ 3-point defense run technically started in 2005-06, it spiritually began in 2007-08. Following two blockbuster trades for Ray Allen and Kevin Garnett, plus the hiring of Tom Thibodeau as an assistant, the modern era of Celtics dominance was upon us. That team had the best defense in the league, won 66 games, and took home a title for their efforts. What drove their success was one of the best defenses of the 21st century, and it was bolstered by the league’s best opponent 3-point shooting.

From that moment on, over the course of 19 seasons, the Celtics have held their opponents to a 3-point percentage at least one percent below the league average in 17 seasons. This consistently elite 3-point defense has led to only two seasons with a worse-than-league-average defensive rating, nine Eastern Conference Finals berths, four Finals trips, and two championships.

It’s no surprise that as the league has taken more and more threes, the one team that can consistently beat the mean has been unwaveringly excellent. However, the fact that the Celtics have done this over the lifecycle of four head coaches and two to three distinct player cores makes this one of the great organizational triumphs in NBA history. Most extended runs coincide with a singular force, be it a player or coach, but the Celtics’ run over the past two decades has been, in no small part, powered by their 3-point defense.

Constant Evolution

One weird trick might work for sharks, but for most species, their only ticket to continued relevance (aka existence) is to adapt. Likewise, the Celtics’ quest to constantly break opponent 3-point shooting has been about consistent defensive evolution.

Defensively, the goal is to prevent points, not limit opponent 3-point shooting, even if it helps. Due to the size of the court and the skill of offensive players, you can’t effectively guard every board of hardwood. Early in the Celtics’ run, they were more than happy to concede shots from corners, but that all began to change in 2016-17, the first season the league eclipsed a 30% 3-point attempt rate.

Corner 3-point attempt rate is the rate of 3-pointers taken from the corners, which means that as the raw volume of 3-pointers increases, so does the raw number of corner threes, even if you experience a decline in corner 3-point attempt rate. Looking at corner threes as part of the total field goal diet, you can see why the Celtics decided that preventing corner threes was more important than it had been in the past.

Looking at this dynamic in chart form provides us with an interesting glimpse into coaching and strategic thought. Under Doc Rivers, from 2007-08 to 2012-13, the Celtics allowed 5.87% of field goal attempts from the corners, compared to the league average of 6.02% for a Corner 3PAr+ of 97.51. Under Brad Stevens, 2013-14 to 2020-21, their 3PAr+ was 94.44, and from Ime Udoka/Joe Mazzulla, 2021-22 to 2024-25, they posted a 3PAr+ of 82.87.

*3PAr+ is the Celtics’ 3PAr based on all field goals, divided by the league average, and then multiplied by 100. 100 is exactly league average, with each point above or below 100 representing one percent from the league average. 95 would be five percent below league average.

The Celtics, as the volume of 3-pointers increased league-wide, and thus corner threes, have emphasized defending the corners more and more. The reason for this is quite simple: corner threes are far more efficient than above-the-break threes. However, none of this explains this.

Celtics Above the Break and Corner Three Defense

Since 2007-08, the Celtics have only allowed opponents to shoot above the league average on above-the-break threes once, and on corner threes three times. So while the Celtics have adjusted their schemes and thinking to limit corner threes, that can’t be the only thing going on.

Contesting the Uncontested

Predicting opponent 3-point shooting is nearly impossible. It’s not because defenders have no sway over the shot; it’s because defenders are rarely there to defend them. This season, teams have attempted 83,856 threes, and of those, 43,723 have been classified as wide-open (nearest defender six-plus feet away), and another 32,040 have been open (defender four to six feet away). Doing the arithmetic, 75,763, or 90.35%, of all 3-pointers are either open or wide-open.

This fundamental fact is why analysts usually roll their eyes when fans claim their team has dramatically improved on defense, only to see their opponent’s 3-point shooting reside around 30%. When only 10% of threes are actually contested, it’s hard to say you’re the reason they’re missing. Unless, you’re the Celtics.

Since 2014-15, the first year NBA.com has data, the Celtics have been the best team at “defending” opponent wide-open threes, and it’s not particularly close. On average, they’ve held their opponents to 3.2% worse shooting on wide-open threes than the league average, which is the farthest from the mean of any team over that period.

Team Wide-Open 3-Point Defense Relative to League Average

Now, they’ve had a few seasons where they allow better than league-average shooting on wide-open threes (2025-26, 2020-21, and 2024-25), but by and large, they’ve kept their opponents well under league average, and this suggests two things.

First, the organization does an excellent job of communicating the who, what, where, and when of 3-point defense. Many wide-open threes are a defensive choice. You’ll see close to a dozen a game, where the defense decides that a certain player can just have the shot. Knowing who and where can have that handful of uncontested threes is a combination of analytics, scouting, and coaching. The second is that the Celtics’ defense has been excellent over this period of time, and elite defense is the ticket to preventing threes from falling.

Defense Matters

Remember when I said that opponent 3-point shooting wasn’t a repeatable skill? Well, that’s not entirely true. You see, on a season-to-season basis, opponent 3-point shooting doesn’t tell you all that much, but over a much larger sample, say 19 seasons, you start to see some correlations.

Using opponent 2-point defense as a proxy for team defensive ability, I found that between 2007-08 and 2024-25, the single-season linear correlation between opponent 2P% and 3P% was only 0.317. That’s not nothing, but that’s also far from a smoking gun. And this relationship is why people say 3-point defense isn’t really a real thing.

However, when you expand the sample to take teams’ opponent 2P% and 3P% over that entire sample, suddenly the linear correlation explodes to 0.661. So, it wouldn’t shock you to learn that the Celtics, who own the lowest opponent 3-point efficiency in the league over this span at 34.2%, also own the lowest opponent 2-point efficiency at 49%.

This all makes intuitive sense. If you’re better at defending shots near the basket, why would that suddenly stop at the 3-point line? Defenses do have to make trade-offs, which is why the correlation isn’t much stronger, but in general, if you can defend within the arc, you can probably defend beyond it. And this is where things get interesting.

The Celtics Have a Type

In my quest to figure out how the Celtics have consistently built elite defenses with a rotating cast of players and coaches, a thought dawned upon me. The Celtics must be targeting excellent defensive players, and that should show up in their measurement data.

The NBA provides combine data beginning in 2001, which gives us a wealth of data to analyze. With close to 1,800 players in the sample, 1,737 of which had some sort of measurements taken, we can pinpoint the positional averages for height, wing-span, weight, hand size, and, to a lesser degree, agility, strength, and leaping ability. While not every player who is invited to the combine measures in every area, when your sample is this large, a few absences aren’t going to fundamentally fudge the figures.

With all of these figures at my fingertips, I sorted players by nine position classifications: Center, Center/Power Forward, Power Forward, Power Forward/Small Forward, Small Forward, Small Forward/Shooting Guard, Shooting Guard, Shooting Guard/Point Guard, and Point Guard. From there, I found each positional average for lane agility time, shuttle run, three quarter sprint, standing vertical, max vertical, max bench press, body fat percentage, hand length and width, height with and without shoes, standing reach, weight, and winspan. Now that I had my positional average template, I was ready to sort through some Celtics.

Starting with the Celtics’ 2007-08 team, I took every individual player season of over 800 minutes through the 2024-25 season and ended up with 36 players who both measured at the combine and played significant minutes for the Celtics in a season, and compared their figures to their positional combine average. Unfortunately, Kevin Garnett, Ray Allen, Paul Pierce, and Jayson Tatum were not in the sample.

The results were somewhat shocking. Contrary to conventional wisdom, the Celtics had not been prioritizing for length. On average, their players were 0.33 inches shorter than the positional average, with a -0.29 inch average wingspan, and -0.79 standing reach. Obviously, the Celtics weren’t completely ignoring height, but it clearly wasn’t a massive priority.

However, where the Celtics did pop was in weight, strength, and athleticism. Positionally, Celtics’ players were 6.28 pounds heavier on average, despite having 0.67% less body fat. On top of that, they performed 1.36 more bench press reps on average, had a 1.16-inch higher standing vertical, and were 0.15 seconds quicker in the lane agility drill. Instead of prioritizing long and lean players, the Celtics had spent the past two decades targeting strong athletes.

When you think about defense, height and length seem paramount, but the ability to hold up physically and move your feet might be just as important. Draymond Green has been wrecking teams as an undersized center for years due to his strength and lateral agility. And as the league has gotten more and more spaced out, the ability to defend in isolation and space has only become more important. Nothing is going to render the defensive advantages length provides moot, but overwhelming length might be far less important for non-rim protectors than we suspected.

The Celtics have been beating the opponent 3-point shooting odds for close to 20 years, not because they set out to do so, but because they committed to an organizational philosophy. They targeted the right type of players, surrounded them with excellent coaches, and were quick to adapt to the changing tactical landscape. I’m sure this run will eventually come to an end, but I doubt it’ll be anytime soon. The Celtics beating the odds was an accident of excellence.

For any inquiries about work, discussion, and the like, you can email me at nevin.l.brown@gmail.com.