If you’ve spent enough time around Warriors fandom, you’ve most likely come across this statement:
They need to run more pick-and-rolls for Steph Curry.
By all accounts, it isn’t a statement without merit. Curry is one of the elite pick-and-roll operators in the league today, both as a scorer and as a playmaker. What makes him such an effective ball handler in such situations is how defenses are forced to pick their poison.
Going under the screen as the on-ball defender is a non-starter, for reasons that don’t need further elaboration. Going over the screen is, by necessity, the default option — but save for the upper echelon of elite screen navigators, most defenders would be hard-pressed to navigate a rock-solid screen from the likes of Kevon Looney and Draymond Green.
Whoever is the screener’s defender must then decide the best course of action in order to help his compromised partner. Do you drop back and hope that your teammate recovers in time to erase Curry’s airspace? Do you meet Curry higher up than usual — as far up as the level of the screen — to discourage a potential shot and provide enough time for proper on-ball recovery?
Or do you commit to a more aggressive form of pick-and-roll coverage — e.g., a hard hedge or an outright blitz — to get the ball out of Curry’s hands and force a lesser offensive teammate to make the decisions and shoulder the scoring burden?
Such is the conundrum that defenses face when the Warriors choose the simplest — and arguably the deadliest — option.
The question then becomes this: If such an option puts the fear of the basketball gods in even the best of defensive squads, then why have the Warriors — Steve Kerr, in particular — been reluctant to spam continuous high ball screens for Curry?
It has been well documented that Kerr eschews turning his team into another run-of-the-mill pick-and-roll-based offense, in favor of heavy ball movement that maximizes touches for everyone. More touches equals a feeling of empowerment and involvement — which translates to vigor and energy on the defensive end of the floor.
Such a philosophy has elicited its fair share of praise, but it also has invited room for criticism — both constructive and overly harsh.
Despite his pedigree and accolades, Kerr is not above reproach. He has previously found it difficult to find a healthy balance between running the egalitarian offense (fueled by Curry’s otherworldly talents) and opting to go the simple route by giving the ball to his best player and letting him cook.
But Kerr has gotten much better in that department. He knows when to rein in his desire for the collective, in favor of letting an all-time great be great.
Despite the high efficiency — save for his injury-riddled 2016 postseason stint, he has at least been within the top 26% in terms of pick-and-roll efficiency — Curry has never profiled as a high-volume pick-and-roll operator throughout the Kerr era, at least when compared to other elite high-usage primary ball handlers who monopolize their team’s offensive possessions.
When comparing Curry to himself, his pick-and-rolls per 75 possessions — including passes out of the pick-and-roll to teammates who shoot, get fouled, or turn the ball over — have mostly stayed consistent throughout the Kerr era. Save for two Kerr seasons (one where the Warriors missed the playoffs), Curry’s pick-and-roll averages have either stayed largely the same from the regular season to the playoffs, or have slightly increased.
Still, they pale in comparison to the gargantuan numbers that the likes of Luka Dončić, Trae Young, and James Harden have put up, especially during the regular season.
A cocktail of factors resulted in significant playoff dips for all three of the players listed above: the generally slowed pace of postseason play, the predictability of their offenses (i.e., elite defenses expecting a heavy dose of pick-and-roll and being prepared to limit it), and a dip in usage (in Harden’s case, who had to share touches with another superstar in Joel Embiid).
The closest analogue in terms of skill set to Curry — at least in terms of shooting range and potency — is Damian Lillard. But Lillard is much more of an on-ball operator than Curry; in Lillard’s last fully healthy season (2020-2021), he ran 24.4 pick-and-rolls per 75 possessions (1.083 PPP, 86th percentile), and around the same amount (1.157 PPP, 84th percentile) in the Portland Trail Blazers’ first-round series against the Denver Nuggets.
But Lillard and the Trail Blazers flamed out in six games, despite his otherworldly pick-and-roll numbers, gaudy stat line (34-4-10), and sublime shooting splits (46/45/94, 66 TS%).
That puts a point in the argument that Kerr is justified for being measured and calculated in whipping out spread ball screen alignments for Curry, and opting for an offense that doesn’t rely on a predictable pattern, but rather, one that changes speeds to throw off opponents on a per-possession basis. Defenses who prepare for a boatload of motion are caught unprepared when Kerr dials up ball screen after ball screen for his superstar; on the other hand, defenses who have been conditioned to expect a continuous stream of Curry pick-and-rolls are caught on the backfoot when the Warriors return to their randomized concoction of cuts, off-ball screens, and chaotic movement.
A specific trend can be discerned — the Warriors tend to increase the amount of ball screens for Curry for two reasons: 1) when opponents find a way to stagnate the motion offense, and 2) when there is a lack of an equal or near-equal self-creator alongside Curry.
Which is why Curry’s pick-and-roll frequency largely remained constant during the Kevin Durant era. With someone being able to share the shot-creation responsibilities — sometimes outright hoarding them — Curry saw fewer ball screens his way.
(Take note of the efficiency of Curry-led pick-and-rolls during the 2017 playoffs – 1.193 PPP, a mark that placed such actions within the 97th percentile. That speaks to the ridiculous offensive firepower the Warriors had on the floor, especially when the dreaded Curry-Durant pick-and-roll mercilessly shredded the Cleveland Cavaliers.)
Take away an equally threatening teammate on offense (Durant’s departure) and couple with it other factors (Klay Thompson’s post-injury form, Draymond Green’s decline as a scorer), and you have the perfect conditions for an uptick in Curry ball-screen possessions during high-leverage situations.
Curry increased his diet of pick-and-rolls from 14.0 per 75 possessions during the 2021-22 regular season (3rd-highest during the Kerr era) to 16.6 per 75 possessions during the playoffs (2nd-highest during the Kerr era). The Boston Celtics, in particular, saw the majority of that increase.
The Celtics were built to stifle the typical Warriors offensive attack. Their combination of scheme versatility, personnel, and switchability was the perfect blend to counter the intricate motion.
Length, athleticism, and collective IQ limited playmaking opportunities for Green, whose value on offense was reduced by the Celtics’ blend of total ignorance (due to Green’s near-nonexistent shooting and general reluctance to score), and jumping at him in order to pressure the pass.
Kerr responded in kind by bypassing the middle man, giving the ball to Curry, and letting him make the decisions on the floor. Even with the Celtics choosing to play varying levels of drop coverage against Curry pick-and-rolls — theoretically minimizing help from other parts of the floor — the prospect of Curry getting loose triggers the instinct to help anyway.
Once the Celtics blink and commit a third defender, opportunities open up elsewhere on the floor:
Rarely does Curry get the chance to showcase this particular ability, but like all elite pick-and-roll operators in history, he is skilled in navigating the middle ground between his defender and that of the screener’s defender. He probes patiently, looking for openings and waiting for the slightest hint of commitment from the roll-man defender — all while keeping his own man in check and close behind on his hip.
Even when the rest of the Celtics defense chose to stay home on their assignments — a product of their decision to drop — Curry found pockets of success by finding Green on the roll.
On some level, the Celtics’ reluctance to commit another man to Curry around screens had logical merit. They did not want to stretch their defensive machinery too thin by having to defend a backline disadvantage, with multiple shooters on the floor and capable short-roll playmakers who can slice and dice their way to easy buckets.
It was a calculated risk built on the amount of trust and equity placed upon their on-ball defenders — particularly Marcus Smart and Derrick White — and in their ability to navigate their way around screens in a steadfast manner. But despite their screen navigation chops, Curry made use of the little airspace he had and, as he is wont to do, overcame extremely good defense with all-time-great offense.
Eliciting favorable matchups and playing the mismatch-hunting game typically runs anathema to their philosophy, but when half-court possessions slow to a crawl and an offensive singularity like Curry is on the floor, high-leverage situations call for the Warriors to seek out possible chinks in the armor and pound at them repeatedly.
The Celtics don’t have bona fide weak spots within their defense, but targeting their main offensive lifelines and making them expend energy on defense will increase the chances of their fuel depleting just enough for them to falter on the other end of the floor.
Such is what happened when Curry received ball screens from teammates defending Jayson Tatum and Jaylen Brown, who predictably switched onto Curry. While both of them are highly switchable and decently capable, a slight positional mishap — leading ever so slightly with their front feet, for example — is enough for Curry to mercilessly pounce.
Besides the Jays, Al Horford — a highly competent switch-big — was sought out on ball screens for Curry:
Grant Williams was another notable mismatch target for Curry and the Warriors. Using a classic and common set out of “HORNS” (a formation with a ball handler at the top of the arc, two players on the elbows, and two players — preferably floor spacers — on the corners), they coax a switch onto Williams:
“HORNS Twist” is when the two players stationed at the elbows set successive ball screens for the ball handler — the first screener dives immediately after setting his screen, while the other screener subsequently comes over to set his, all while corner spacers keep their defenders nailed to their spots to discourage them from helping on the main action.
In the case above, the endpoint is a Williams switch, and a Curry three over Williams.
HORNS Twist is arguably the one HORNS-based action that most blatantly involves the pick-and-roll — and therefore, something that shouldn’t be in Kerr’s playbook, given the notion that he doesn’t like to run pick-and-roll half-court sets.
Truth be told, this has always been in the playbook during the Kerr era:
The HORNS Twist possession above happened during the Warriors’ November 4, 2015 matchup against the Los Angeles Clippers — nearly seven years ago.
Whenever the Celtics — in lieu of dropping or switching — bit the bullet and decided to send two bodies to Curry around ball screens, the worst-case scenario happened almost every time:
What has yet to be mentioned is the amount of pick-and-roll adjacent play types the Warriors ran — most of them part and parcel of their natural motion flow — for Curry. Even while belonging in the bottom tier of teams in terms of pick-and-roll frequency last season (13.8 per game, 2nd-least), they ranked higher in the number of handoffs (7 per game, 7th-most) and off-ball screens (10.6 per game, most in the league) they ran.
For example, take these two possessions:
Technically, these empty-corner possessions aren’t tallied as pick-and-roll plays — they’re considered “handoffs.” But when we zoom out from the definition of what a pick-and-roll is and use an umbrella term such as “screen-and-roll,” the delineation between handoffs like the ones above and the classic ball screen becomes less defined and more blurred.
Or how about this possession from Game 5 of the NBA Finals:
This is a pet Warriors half-court motion set: a “Ram” screen or screen-the-screener action (i.e., when Gary Payton II screens for Thompson, who then sets a ball screen for Curry), coupled with an “exit” screen for Andrew Wiggins. This then flows into a classic low-post split, with Wiggins setting the off-ball screen for Curry.
What happens next is the real meat of the possession. Horford and Tatum double Curry around the screen, setting Wiggins loose to dive to the rim with the ball for a layup. The question then becomes this: How different is any of this from a typical pick-and-roll where Curry draws two defenders around the ball screen?
Or even consider this possession from Game 6:
What will you call what happened above? A handoff? A ball screen? A combination of a handoff and a ball screen? Either way, what transpired looks highly similar to if Curry was handling the ball from the start, called for a ball screen, and pulled up for a three with the opposing big dropping deep in the paint.
When you combine Curry’s pick-and-roll possessions with the numerous handoffs and off-ball screening actions he’s involved in — and cast them under the aforementioned umbrella term of “screen-and-rolls” — you’ll find that he virtually matches both the pick-and-roll volume and efficiency of Dončić, Young, and Harden during the regular season and the playoffs.
Unlike the other three superstars, Curry goes about his business differently. Pick-and-rolls — in the form of early offense drag screens, double drag screens, etc. — aren’t the be-all-end-all of the offense that he captains. But if needed, they can end possessions with much needed buckets, and ultimately end a playoff run with a championship.
When adding the pick-and-roll adjacent actions — the handoffs, the off-ball down screens, multiple staggered screens, exit screens, etc. — that generate the same kind of looks and scoring opportunities that a typical ball screen would, Curry’s relationship with the pick-and-roll has been more complicated than one would think.
It’s not that Curry doesn’t run enough pick-and-roll — it’s that he runs it in varied and creative ways that throw opponents around for a loop. For a player that has a unique skill set that no other player in the history of the NBA can replicate, it’s only appropriate that Curry — and the team that is his extension — zigs when other players and teams zag.