Developing an understanding of what made the Knicks tick on offense this season
Much hullabaloo has been made about the state of the Knicks’ offense this year. We’ve all heard or asked variations of the following questions: Is it too isolation-heavy? Is there a system? Where is the ball and player movement? Is David Fizdale — pissed off that EA doesn’t make NBA Street anymore — just trying to get Trier to bust out a Level 2 Gamebreaker on some poor defender? And so on.
In part one of this two-part article, I’ll identify and explain several concepts that the Knicks rely on in their offense, and try to figure out whether those leading questions are leading anywhere reasonable.
Let’s start by taking a look at a few stats that will help give us a general overview of the offense.
(Some of these stats are a few weeks dated because I’ve been working on this piece for a while. Fortunately little of significance has happened since I collected these.)
A good place to start is with team offensive rating (ORTG). This is a calculation of the number of points an offense scores per 100 possessions, and it provides a nice bite-sized overview of the health of the offense. The Knicks have an ORTG of 104.3, which is good for last in the NBA. For comparison, the Warriors lead the league with a robust ORTG of 114.7, and the median NBA team ORTG is 109.5 (as of March 7). To the surprise of no one, the Knicks’ offense is not healthy.
Let’s dig a bit deeper. First, let’s chat about ball and player movement. It’s commonly thought that lots of both of these are essential for a healthy offense. That’s straightforwardly false. In fact, over the last few seasons, offensive ball and player movement have been negatively correlated with offensive rating, and many of the league’s best offenses have very little of either — consider this year’s iso-heavy iterations of the Rockets and Thunder, for example. That said, there are excellent NBA offenses that thrive off of ball and player movement (e.g. the 76ers and the Warriors). The bottom line is that there is no single recipe for excellent offense.
So where do the Knicks fall with respect to these stats? Right in the middle of the pack. As of March 7, they were 18th in passes made per game and tied for 17th in miles run on the offensive side of the ball per game.
In a bit, I’ll use film to explain where most of this ball and player movement comes from within the offense. But first, there’s a bit more to be said about their passing statistics. Quite frankly, they’re atrocious. Despite moving around a fair bit and passing the ball quite a bit, they produce assists like a team that neither moves nor passes. In fact, the Knicks average fewer assists per game (20.1) than any other team (the Rockets average the second-fewest).
Even worse, they have the statistical profile of a team that doesn’t capitalize on the advantages they create. So, for example, while they average a decent 44.9 drives per game (good for 13th in the NBA), they almost never pass out of those drives (they have the second-worst pass percentage on drives). When they do, it’s usually with little success (they sport the NBA’s worst assist percentage on drives). But the problems don’t stop there. Even when they make good passes off drives, the team is terrible at swinging the ball to the open man. This is reflected by the fact that they have the third-fewest secondary assists per game (2.3). Per NBA.com, “a player is awarded a secondary assist if they passed the ball to a player who recorded an assist within one second and without dribbling”.
The Knicks also lack any high-level individual creators. In general, NBA players are not efficient isolation scorers. Even the very best isolation scorers often average fewer points per possession (PPP) in isolation than their teams do as a whole. This is true of James Harden, for example, the most efficient volume isolation scorer in the NBA. On isolation attempts that result in him shooting he averages 1.11 PPP. If every one of the the Rockets’ shots came off a Harden iso attempt, the Rockets would have a league average offense. But when the Rockets work together to create advantages that result in better shots, they significantly outperform Harden’s solo mark. Specifically, they average a whopping 1.14 PPP. And they do this even though many of those shots come from shooters who are much worse than Harden.
Of course, because of Harden’s skill, the Rockets able to create advantages out of isolation sets. Harden is able to collapse a defense and demand second and third defenders help off their men, which allows the Rockets shooters to come open without much in the way of ball or player movement.
The Knicks’ best isolation scorers/shot creators pale in comparison to Harden, so they do not have this luxury. Here are the Knicks’ top five iso scorers: Trier (1 PPP), Vonleh (.88 PPP), Mudiay (.81 PPP), DSJ (.81 PPP), and Ntilikina (.76 PPP). Quite clearly, the Knicks do better (1.04 PPP) when they work together to create advantages that result in better shots.
Not surprisingly (given their terrible passing and lack of a high-level shot creator) the Knicks do not create very many open shots. Only 47.7 percent of the Knicks’ shots are listed by NBA.com as either wide open (no defender within six feet of the shooter) or open (closest defender within 4-6 feet). If you don’t pay attention to these sorts of numbers, you might think that this seems… good? Nearly half of the Knicks shots are open or wide open?! But, lo and behold, the Knicks are bottom feeders. Only four teams create fewer wide or wide open shots than the Knicks. (If you’re curious, the Boston Celtics lead the NBA — about 57 percent of their shots are open or wide open. Yowzers).
Alright, so the Knicks have a terrible ORTG, and that’s probably heavily influenced by the fact that they are a terrible passing team which doesn’t have a high level initiator or shot creator and as a result they don’t generate enough open shots. So far… not surprising.
In this section I want to do a couple of things. First, I want to chat a bit about the concepts that structure the Knicks offense. While the Knicks — like every team — run some set plays, there are a number of offensive sets the Knicks go back to time and time again. I’ll offer a brief overview about how a few of these work, and what the Knicks are trying to achieve with them.
Let me start by noting that I don’t have access to fancy Synergy data. Below I’ll say some things like: “the Knicks often start their possessions with this action…” Since I don’t have access to the fancy data, these claims will be based off my observations. As a result, these will be ball park claims, and should be understood as such.
Pin-down screens are a staple of many NBA offenses, and the Knicks follow suit here. In a pin-down screen, the screener faces the baseline and attempts to “pin” a teammate’s defender to that spot, while their teammate sprints away from the baseline. In this (somewhat blurry) image you can see Mitchell Robinson setting a pin-down screen for Kevin Knox.
Pin-downs set on the baseline like the one pictured above are sometimes called “wide pin-downs”, while pin-downs set in the paint are called “middle pin-downs”.
Ideally, wide pin-downs result in dribble penetration, an open three or a backdoor basket, depending on how the defense plays the screen. If the defense doesn’t switch the screen, then the defender guarding Knox (in the example above) will either trail him around the screen, or go under the screen. If he does the former, Knox should curl around the screen like this:
With the defender on his hip, Knox is free to attack the rim. Now, there are a bunch of problems with this particular clip — Vonleh has completely screwed up the spacing, Knox doesn’t keep his head up, etc. — but that’s the basic idea.
Before moving on, let me make a note of why I paused the previous clip partway through, and why I pause some of the later ones. Every time down the court, 10 different players move around simultaneously. That’s an impossible amount of activity to keep track of in real time. As a result, I often pause things at key moments. For example, in the above video I paused it right when Knox comes around the screen and the defense has committed to trailing. Moving forward, I’ll also use it to point out when an open player has been missed, and sometimes just to give you an opportunity to survey the court, check defensive alignments, etc. Alright, back to the action!
Here’s another example of a wide pin-down screen where the defender trails. This time, Robinson’s man is forced to stay in front of Knox which leaves Robinson open for an alley-oop… Unfortunately, Knox fails to recognize this and puts up a difficult layup.
You’ll also notice that, in the above clip, the defense collapsed into the paint in order to stop the penetration. This creates opportunities for the ball handler to kick the ball to an open teammate on the perimeter.
While trailing usually forces the offensive player to curl into the lane, if you run hard enough around the screen you can sometimes generate enough space for a three. Dotson is probably the only Knick with this ability (it’s possible that Jenkins can also do this, but I haven’t seen enough film of him to make a confident judgment).
When the defender instead goes under the screen, the offensive player typically should pop out for a three like this:
Like some of the previous examples, the latter play isn’t perfect. Ideally, Knox would have taken two more steps back towards the baseline which would have made it harder for his defender to contest. But you get the basic idea.
When the defense is worried about the shooter, they may sometimes hedge the screen or briefly commit two players to the shooter. Or the defenders might just get their wires crossed and accidentally both follow the shooter. When this happens, the screener often comes open on a back door cut.
I’ve been waiting all year for an example of this to make a clip of, but to no avail. Defenses often pay extra attention to Dotson on pin-downs, and the back door has been open countless times. But none of the Knicks’ passers have (to my knowledge) successfully made that pass. But one recent fateful night, the Rockets played defense against the Knicks like they were up by 40 (because they were…) and Luke Kornet found himself unguarded and diving towards the hoop after setting a wide pin-down. Here is the proof:
The Knicks loooove to start possessions with wide pin-downs (especially after an opponent’s make), and they often create good looks this way. But too often they fail to execute them properly. We’ve already seen the team make several mistakes. But perhaps the most common and most infuriating is their failure to set good screens. Here are a few examples to brighten your day:
A related action that the Knicks like to use to start possessions (especially baseline-out and side-out possessions) are middle pin-downs. Here, the ball handler runs to the wing, leaving the top of the key open. A pin-down screen is then set in the paint, and an offensive player runs off it to the top of the key. At that point they can either shoot, attack the close out, or transition into a high pick and roll or designated handoff action. Here are two examples where it fails to create a clean look, and the offense flows into a PnR:
The Knicks set stagger screens all over the court. They set them on the ball, they set them off the ball, they use them in combinations, they use them to disguise other actions, etc. In some instances, the stagger screen works a whole lot like a wide pin-down screen. However, since there are two screeners, it is significantly more difficult for the defender to trail the shooter around the screen. This allows players without Dotson’s quickness/deception/balance to put up threes even when the defender trails like this:
Still, this defense typically results in the player curling toward the hoop like Dotson does here:
Like the plays mentioned previously, stagger screens can flow naturally into DHO/PnR or other on ball sets. For example, in after-time-out play, the Knicks string together two stagger screens followed by a DHO.
Here’s a few things to note about that play. First, the defense has to do an enormous amount of communicating. In the span of about seven seconds, the defense has to deal with five screens. This often leads to mistakes, and it does so in this case. As you can see, two players follow Dot on the second stagger, leaving Knox uncovered and diving to the rim for a split second. Unfortunately, Mudiay doesn’t see him in time (a frequent theme), but these are the sorts of openings standard stagger screens can create.
There are a large number of things that screeners can do once they set their picks. Often, one man will dive and one will pop, like here:
But they can also mix things up. For example, in this next play Kornet and Robinson set a stagger screen (Robinson is late, and fails to set the 1st screen). Immediately after, Robinson steps up and screens Kornet’s man. This creates a 2 on 1 opportunity for Kornet and Robinson which should lead to an easy bucket. But, unfortunately, Kornet bobbles the pass.
In this next variation, Robinson pretends he’s going to dive to the rim, drawing the attention of his defender, Joel Embiid. After showing dive, Robinson turns and forms a wall on the sideline with Luke Kornet for Dotson to slide through before they close it off (this is sometimes called an “elevator screen” because it looks like elevator doors closing). This makes it nearly impossible for JJ Redick to go under the screen, and — because of how wide Kornet is — makes it very challenging for him to go over the screen quickly enough to contest Dotson’s 3-point attempt. Kornet’s man is also sealed to the baseline and has no chance to fight through and contest. That leaves Embiid as the only one with a realistic chance to contest, but as you’ll recall, Robinson’s dive action attracted his attention and he never makes a move to contest.
As you’re probably noticing, lots of the Knicks actions involve closely-related concepts. This next one follows suit. Horns sets involve two screeners who effectively sandwich the on-ball defender. Here’s an example where the Knicks show a stagger set, but switch to a horns set just before initiating the action.
For the ball handler, horns sets work a lot like a PnR. As you can see, DSJ comes around the screen facing drop coverage just like he might when running PnR. Where this set differs is the number of options that the screeners have. Rather than having a roll man only (or a pop man only, in PnP), you can have both on one play.
Alternatively, as with the stagger set, the Knicks sometimes screen one of the screeners.
The Knicks use this trio of on-ball actions relatively frequently. Not only do they use them to start plays, but they frequently stack them with the previous sets to preserve or increase advantages they’ve created, or use them as bailouts at the end of the shot clock.
Since these actions are discussed a bit more commonly, and this is getting a bit long, I won’t discuss these in depth here (not that there isn’t a lot to be said about them). but here are a few things worth noting:
(1) Because of the Knicks poor on-ball passers, almost all of the Knicks PnR attempts end with the PnR ball handler shooting.
(2) When the PnR results in a pass, it’s almost always a lob at the rim. None of the Knicks guards can reliably make pocket passes, so they were mostly unable to utilize the popular and effective short roll. And, none of the Knicks guards can reliably read and react to where the help defense comes from, so we rarely saw guards find shooters out of the PnR.
(3) Since the arrival of Deandre Jordan, the PnR sets have been more sophisticated. He uses tactics that our younger guys haven’t yet mastered and that Kanter didn’t reliably implement. For example, he’s adept at switching the direction of the screen, slipping the screen when he recognizes the ball-handler has beat his man, etc.
(4) PnR/DHO are by far the Knicks most used actions. They use them to start plays, but also as bailouts when up against the shot-clock.
Drive and kick, and putting it all together
None of the actions I’ve discussed so far can be expected to produce open shots every time down the court. But, they can — when run correctly — be expected to reliably generate advantages for the offense. Of course, advantages can be maintained, increased, or lost altogether. The keys to a successful offense are (1) creating advantages, (2) increasing advantages and (3) turning those advantages into good shots (preferably at the rim or the 3-point line).
One of the central ways NBA teams attempt to maintain and increase advantages is via an action called “drive and kick.” As you’ll note, many of the concepts I’ve outlined here result in an offensive player attacking the basket with the ball with a 2-on-1 or 3-on-2 advantage. Wide pin-downs and staggers result in this when the defender follows the offensive player around the screen, and PnR/PnP/DHO and Horns sets are designed to do this as well. Some teams don’t require special actions to create these advantage situations. For example, high-level isolation players like James Harden or Giannis Antetokounmpo are capable of creating them on their own. But the Knicks don’t have any players who can reliably do this, so they rely on a variety of the concepts I’ve outlined.
Regardless, when these advantages are created, defenses — intent on preventing easy shots at the rim — will often bring help in order to cut off the penetration. Fortunately for the offense, this often leaves a player uncovered on the perimeter. It is the responsibility of the driving player to find the open man and kick the ball to them. Hence the name “drive and kick.”
Once the offensive player kicks the ball to the open man, their responsibility is to immediately sprint to the 3-point line. This is because the open man will be faced with the following options upon catching the ball: (1) shoot the ball if he’s open enough, (2) if the defense brings perimeter help, swing the ball or (3) if the defense closes out from the paint, attack the rim. If (3) is the optimal option and the initial ball handler is still standing in the lane, the lane will be clogged and the penetration will be unlikely to be successful. However, if the initial ball handler has cleared out and positioned himself on the 3-point line, the offense can re-run the drive and kick action in order to maintain or increase their advantage until they end up with a good shot.
The Knicks are absolutely terrible at this, and — in my view — this is where many of their offensive woes stem from. While their offense isn’t particularly complicated, the Knicks create plenty of advantages when they get into their sets. However, too often they fail to maintain or increase these advantages. As a result, they end up running isolation sets or desperation PnRs against the shot-clock, which is a sure-fire recipe for poor efficiency.
That sums up the first chapter of this two-parter. The next chapter will be film focused. In it I’ll (1) provide examples of the offense working how it’s supposed to (2) attempt to explain when and why things go wrong when they do. I hope you’ll check it out!
A few acknowledgments:
I’ve learned a lot from reading and listening to Ben Taylor’s work, especially stuff related to offensive advantages (he likes to call them “power plays”). I’ve also picked stuff up from Mo Dakhil, Caitlin Cooper, and Ian Levy. And if you haven’t read Zach Diluzio’s really excellent piece on the Knicks offense from earlier in the season, you should!