This is more of a ramble (or though process) but I wanted to explore my thinking on spacing as it relates to the Memphis Grizzlies, and the NBA, and spacing in offence.
I’m using two examples from the most recent San Antonio Spurs game and will at some point in the season revisit this. Important to note that I am not criticizing coaching or the player’s execution of the offence, I am simply putting together thoughts on some of what we see and why I do/don’t like it.
First examples is a free flowing offence, from Memphis, without a set play. The initial drive, by Ja Morant, starts in transition offence before flowing to Jaren Jackson Jr.
I wanted to post this play on twitter but I couldn’t come up with a tweet that I felt captured what I wanted to say. From here I am mostly just going to break down the play and my thoughts on each point.
- First, Ja Morant transition offence — stopped by Jeremy Sochan and Victor Wembanyama. It’s fine, I am not sure you’re going to create a + scoring opportunity against Wembanyama and two other defenders in transition.
- Now we get JJJ trailing, receiving the ball, and driving to the rim. Here is my first problem. JJJ is driving right back into Morant’s initial defender(Sochan) and his own defender (Tre Jones). He technically has a mismatch on his defender but he’s driving into two players + nail help + Wemby as the rim protector.
>This, to me at least, is not optimal spacing. Back-to-back (b2b) drives into the same space is, again to me, negative spacing and especially if Morant is not going to immediately fill another spot above the break and initiate the 5-Out spacing.
- On the initial Morant drive, Bismack Biyombo filled the dunker spot in the half court. This is spacing I like. It is more of a 4-Out 1-In spacing but occupying this spot doesn’t hurt the offence. It is positive spacing because Wemby needs to make a decision (Biyombo or the drive — because a dump pass it always an option) on who to guard. If Biyombo was spaced out to the corner it creates a new spacing issue because Wemby wouldn’t have to make a decision (whether to guard the drive or Biyombo) when he could guard both the rim and the pass.
- Ultimately Wemby guards the drive from JJJ, into occupied space, and Biyombo is there to receive the pass and reset the offence.
- This changes the shape of the offence but maintains the spacing the team is looking for to execute offence.
a) Morant is now fully above the break, b) JJJ, after his drive, fills the weakside corner, and c) Marcus Smart, who had originally filled the corner, cuts the baseline to the strong corner (effectively this replaces Biyombo spacing). This is the shape of a 5-out offence the Grizzlies regularly run their offence through (78% of all halfcourt offence by my tracking).
- Now comes a P&R. Does a P&R* provide spacing? A P&R is the most frequent possession type for the Memphis Grizzlies. I am not saying it does not provide spacing but it is bringing two defenders together — should note spacing is less about the offence and more about how the defence is defending you — to defend two players on offence. Scheme counters and player dependant, a P&R is exploitable but, to me, it does not provide spacing. Of course the P&R is not going anywhere but was it really required in this scenario? Bane is not going anywhere against Wemby so he passes out to Morant who has now filled Bane’s previous spot above the break.
(*Interesting that at the same time I am writing this, Louis Zatzman wrote a piece for SBNation on The NBA is shifting away from pick-and-rolls, but offensive efficiency is higher than ever** — I do recommend reading this)
(**Second note — I really like the way Matt Hrdlicka writes his Hrd notes in his Patreon — and I may steal that idea at points while I write, including now)
But here’s the thing, there are now two defenders in the paint after Bane’s drive and Biyombo’s roll to the rim. What does Ja do? He drives directly into space these two defenders, and now his own defender, are occupying.
Now Bane has relocated to fill the third spot above the break joining Smart and JJJ.
The three defenders did not rotate and Bane has a wide open 3. Is this good spacing? Or is it poor defensive rotations? I’m not certain it’s clear (despite what loud people will tell you) either way but it’s fun to think about and is a good illustration of how I think about spacing and why it is important.
This play is a little different due to two reasons a) this is against a set defence and, b) it is a set After Timeout (ATO) call. While the other clip was a free-flowing offence with principles the coaching staff implements, this is a drawn up play for the Grizzlies to execute (though there are also principles to it and that’s where I am getting to).
The initial formation of this play looks like a 3-Out 2-In where Santi Aldama sets the initial step-up (“Knicks”) screen for Morant. Again this is a screen (see above comments) where two defenders are coming together to guard two players on offence — affecting spacing principles.
Aldama ghosts the screen (i.e. doesn’t make contact) and receives a pass from Morant where we see the same 4-Out 1-In formation as the play above.
> Biyombo fills Aldama’s original spot (~dunker cut)
> Aldama initiates a dibble handoff (DHO*) with Luke Kennard and both players switch spots maintaining space
(*Is a DHO good spacing — again it is bringing two defenders together)
- Biyombo sets a second step-up screen (“Knicks” from the ATO), while Vince Williams Jr. drifts the baseline from one corner to the other, and Morant receives a pass from Kennard (“Reverse” from the ATO).
This is where the play breaks down and Morant gets the ball with 9 seconds left**
(**will write on this below***)
With 9 seconds left comes the second part or the secondary action — a Veer Screen. The Veer Screen starts with a P&R into an off-ball screen (are either of these good spacing?) for Kennard. The Spurs are in a heavy drop (the veer screen is a “drop beater” after all) and Kennard gets a high quality 3PA*.
(*regardless of the shot falling I am more interested in the process/execution of the play than the actual result because over time it will even out)
Throughout the execution, when the play starts to break down, the Grizzlies maintain spacing through 5-Out offence.
***
Here is the part I mentioned earlier. Two additional things I am tracking this season are a) shot clock when the offence is initiated (not including time to take the ball up court) and, b) how long a play takes to execute. This is specific to the second example with the Kennard 3PA when set plays are being executed — I am not as interested in how long free-flowing offence takes to execute as these plays are too varied.
There is no real shape to this graph* at the moment — I have posted it at various points of the season on twitter and here is the update as of January 03, 2024, after the San Antonio Spurs game. At the beginning of the season sets were clearly not working but the secondary sets and “tough” shot making when the offence broke down was scoring at a high rate. After 25 games the line was starting to resemble a line trending upwards based on play length and now (above graph) we are at a point where set offence is improving and “tough” shot making is coming back to normal.
(*these numbers are all hand tracked again — take it with a grain of salt)
But how long does a set take to execute, what is a tough shot, and does this matter?
On average a play, this season for Memphis, takes between 6–8 seconds to complete. I have highlighted the wins in yellow and losses in blue*. Based on wins and losses there doesn’t appear to be any difference in play seconds but I am bringing this up for another reason. Though using the two graphs we can see what offence between 6–8 seconds, for Memphis, is significantly worse than offence between 9–12 seconds (tough shot making is still better than the Grizzlies set plays, on average). It is not an exact one-for-one (remember this is the average play length) but we can see the most frequent (in the first graph, larger bubble = more possession) plays take between 4–7 seconds (i.e. when Grizzlies run most of their sets) and again is significantly worse than the 9–12 second timeframe.
(*a better idea for the future would probably be to colour code based on offensive rating of HC sets but that’s for another day)
In the Morant/Kennard play above the set started at 16 seconds (remember it takes time to get the ball past half court) when Aldama screened for Morant and the play broke down at 9 seconds — this is a 7 second lapse (15 total seconds) while the Grizzlies executed the play during which the Grizzlies did not get a shot off. This is my long way of saying is set offence the optimal offence in the NBA?
I’m not sure. The Grizzlies have, by my tracking, scored 0.9533 Points per Possession (PPP) on halfcourt sets (1.0184 PPP on ATO) and 0.9042 PPP on freelance/free-flowing plays (none of this includes transition offence). All three of those numbers are bad bad (!!) but the Grizzlies are clearly scoring better on sets. The other part is free-flow offence executed in the NBA (or at least with how the Grizzlies execute these possessions) often does not contain the same principles as set plays would. Is there a lost opportunity (points) within these free-flowing plays?
All of this ties together — average play seconds, spacing, and screens (P&R) on offence. This is not meant to be conclusive, again, this is my thought process and me talking out loud. I just hope it’s coherent.