Mazzulla-ball goes deeper than you realize

You’ve seen this movie so many times over the last couple of seasons, but it never ceases to satisfy.

The opposing offense, Cleveland this time, runs a play hoping to free up a roller or shooter, but nothing materializes as Boston defends it perfectly. The Celtics communicate on switches and stick with the screeners, and Sam Hauser holds up against Darius Garland’s probing dribbles. The ball ends in Donovan Mitchell’s hands, and with time running out, he is forced to sidestep past an improvised Jarett Allen pick and launch a no-hope grenade:

Clank. It’s obviously an excellent defensive possession, both individually and collectively. But zoom out, and it also hides a key part of what makes the Celtics’ defense so good: they force opponents into the worst kinds of basketball plays.

Synergy Sports categorizes how teams use possessions into different buckets: cuts, hand-offs, isolations, off-ball screens, pick-and-roll ballhandler and roller actions, post-ups, spot-ups, and transition plays. They then track how often (and how efficiently) teams end a possession using those actions.

For example, if a playtype is labeled P&R Handler, that means the sequence ended with an action from the ballhandler in the pick-and-roll, like a turnover or shot attempt. It’s not comprehensive — especially today, many of these actions are combined, so the data undercounts things like pick-and-roll that initiate so many possessions — but it’s still a solid directional view of how teams approach offense.

Here’s the breakdown of leaguewide offensive actions, the associated points per possession (PPP), and the percentage of total possessions used in this manner, as of 12/6/24:

Note that the total time doesn’t add up to 100%; I’ve excluded offensive rebounds and Synergy’s catch-all miscellaneous category, since I’m mostly interested in first-shot attempts, and the miscellaneous category defies easy description. Each is ~5% of possessions.

You can see above that plays ending in cuts are by far the most efficient shot in the NBA, which makes sense. Plays ending in a cut almost always lead to open layups or dunks, because players don’t pass to a cutter unless they’re open.

What might be more surprising to some is what’s at the bottom of the list. Would you have guessed that P&R ballhandler shots are the least efficient offensive action?

This isn’t a one-year blip, either; it holds for past seasons. Pick-and-rolls aren’t really designed to get the ballhandler a shot. It’s great if they can get an open layup or three-pointer, but more typically, the defense will crowd the action, opening up a shot for someone else. Many pick-and-roll ballhandlers take shots after the designed offense has failed or late in the shot clock as an improvised attempt to create some space, like in the Mitchell clip above. Relatedly, they tend to create self-generated jump shots, which are generally low-efficiency shots.

Isolations rate poorly for similar reasons. Movement is the key to creating good offense, but isolations often involve minimal involvement or action by the players without the ball. That’s why, particularly in the regular season, they are often a weapon of last resort. Can’t generate an open shot for someone through scripted offense? Give it to your best guy and hope he can make magic happen. Sometimes (often memorably), it works; more often, it doesn’t, killing possessions with quiet rim-outs.

(Playoff basketball leads to some different outcomes, of course; teams are heavily focused on optimizing game plans for specific opponents, meaning playtype frequency and efficacy can ride a roller coaster between series depending on who is playing and how they perform in a scant handful of games.)

Think about how many teams try to attack Sam Hauser in isolation and regret it afterward! Boston is meticulously coached and overflowing with intelligent, talented defenders. We know this on an implicit level, but seeing it explicitly laid out is fun. If you’re designing the ideal regular-season defense, you’d want to minimize cuts and maximize how many possessions end with an isolation or shot from a pick-and-roll ballhandler.

Guess what? Unsurprisingly, that’s exactly what Joe Mazzulla’s teams have done:

As you can see above, the Celtics have allowed very few cuts in the last three years while forcing teams into the worst offensive actions more than almost anyone in the league.

This is by design, a product of both the Celtics’ defensive principles and their overflowing talent. Nobody is more analytically inclined than Mazzulla; he knows exactly what he’s doing, and his vision is made manifest by the roster. Having players stay home on their marks without overhelping eliminates a lot of the looks opposing offenses want to get, flattening the offense until there’s nothing left but for the ballhandler to toss up a Glory Be.

Of course, there are exceptions. For example, Boston is actually one of the league leaders in offensive isolations — they’ve run the second-most in the league this season, behind only James Harden and the Clippers. But an isolation for Boston looks a whole lot different than most teams, thanks to the plethora of shooting up and down the roster. All that spacing, and the open lanes to the rim it provides, means that a Boston isolation is quite efficient — 1.08 points per possession leads the league by a mile (the Dallas Mavericks, thanks to Luka Doncic and Kyrie Irving, are the only other team who crack 1.00 PPP).

But overall, it’s good to direct teams to the least efficient playtypes. Nobody has done this more consistently than the Joe Mazzulla Boston Celtics.

[If you’re interested in this subject matter, I did a deeper dive on the offensive frequency and points per possession for every NBA team to prove that all NBA teams don’t play the same way. It even has a couple of interactive charts, so check that out if you’re so inclined!]