To take a break from the heavy mathematics of recent articles, today we take a look at the early offense set called Pistol. Like many offensive wrinkles such as elevator doors, floppy, horns (etc.), a key goal of the offense is to force the opposing defense to be out of position. More importantly, not only to get a defense out of position, but also maximize the probability of scoring when a defense is out of position. The best offensive executions typically either result in an uncontested lay-up or an uncontested three point attempt from the team’s best shooter.
In the Pistol action, the offense attempts to catch the defense before it sets in hopes to find optimal mismatches or blown coverage by a lack of defensive rotation. It’s called an early offense as it usually begins on the heels of a transition or is the initiation of an offensive set. One of the key components is that the Pistol breaks up other revolutionary defensive rotations such as BLUE, as the offense is set up to go sideline already. The difference here is that if Pistol is run correctly and efficiently, there is rarely a two-nine defender who can deter the layup.
Pistol, also known as the “21 series” to many teams, starts in a triangular format. Typically, utilizing the point guard, the shooting guard, and a big; hence the “21” in the name. The triangle usually runs its base parallel to the sideline as almost to be set up as a give-and-go far away from the basket.
Typically, the Pistol will be initialized through one of two options: the pass and the dribble. These two different initializations will dictate the how the primary and secondary actions will break out. Ideally, the court is spread out such that there are no offensive players near the key.
It should be noted that the Pistol need not start in spread-option. This is a common set-up, but not always the case. By using this spread option, however, the defenders that are coming down into position must make initial decisions to guard space or their man; both leading to decidedly difficult situations.
Option A: Dump and Chase
The pass option mimics a “dump-and-chase” outlet; similar to a give and go. In this situation, the point guard passes ahead towards the shooting guard. Usually the point guard has drawn a defender well outside the three point line. By dumping the pass, the point guard is free to accelerate by chasing the ball. This forces the point guard defender to have to rotate and accelerate; thereby putting him at a disadvantage for staying between his man and the basket.
Situationally, this places more emphasis on two defenders: the shooting guard and the two-nine; if the latter even exists.
As we see the Point Guard chase the pass, the defender saddles near the player, forcing them towards the sideline. In this case, the offense must make their primary decision with the basketball.
Option A1: Give and Go
The standard option is to perform a give-and-go. The Point Guard can rub his defender off the Shooting Guard using a handoff, forcing the Shooting Guard’s defender to either show or drop. A show will uncover the shooting guard for a roll to the basket, while a drop to give the Point Guard’s defender a chance to fight through the screen with minimal contact will allow the point guard breathing room.
The primary action actually catches many teams off-guard. Let’s take a look at the execution from a Portland Trail Blazers game against the Houston Rockets on November 3rd, 2013.
Here, Damian Lillard (POR – 0) tosses the pass ahead to Mo Williams (POR – 25) with LaMarcus Aldridge (POR – 12) playing the big position. As the Pistol initializes, notice how Wes Matthews (POR – 2) fades to the opposite corner, drawing James Harden (HOU – 13) away from the action, and Dorell Wright (POR – 1) staying at the opposite wing, drawing Francisco Garcia (HOU – 32), above the key.
This ensnares Dwight Howard (HOU – 12) on Aldridge, leaving Patrick Beverly (HOU – 2) and Chandler Parsons (HOU – 25) to guard the give-and-go. Notice Beverly stays well beyond the three point line; giving up the dump-ahead pass to Williams. Lillard then accelerates hard, leaving Beverly in rotation and forcing Parson to show. With a full head of steam, Lillard blows by Parsons for a nearly uncontested layup as Harden merely watches from the weak-side.
Option A1S: Secondary Action off the Give-and-Go
If the defense reads the give-and-go well, then a path to the basket does not appear and the offense must resort to a secondary action. This action is usually either a screen or a pass. Regardless, the secondary action must operate through the use of an off-ball player’s action.
The most common is the Screen-the-Screener action. We will see this pop a few more times later in this breakdown. The action is simple: after the give-and-go handoff screen, the Big comes to set the screen for the Shooting Guard. This action creates a flare to the top of the key for a potential open three; or opens up a slip by the post if the defense reacts poorly towards the shooter.
As example, consider Portland’s December 6, 2013 game against the Utah Jazz. In this play, Mo Williams (POR -25) initializes the offense through CJ McCollum (POR – 3) with Brooke Lopez (POR – 42).
In this situation, Diante Garrett (UTA – 8) plays well off Williams; but is hedging towards the center of the court. After Williams dumps to McCollum, Garrett and Alec Burks (UTA – 10) switch the play, and Enes Kanter (UTA – 0) drops into the paint to help deter the drive. Note that Kanter can get away with this as Lopez is not a well-known three point shooter… even moreso in 2013.
This leads us to the secondary action: Lopez comes to set a screen on Garrett. However, as Garrett already hedged to the center of the court; this forces McCollum out of the play. Lopez then moves into another secondary action and screens Williams. At this point, we have already reached 8 seconds into the shot clock and the Pistol is nearly a broken play. Despite this, the screen forces Burks to go under, leaving Williams for a lightly contested 21-foot jump shot; which is effectively a win for Utah on the play except for one fact: Williams knocked it down to give Portland the lead.
Option A2: Keep
The second primary option on the Pistol is for the Shooting Guard to keep the ball. In this case, the Point Guard will typically run away from the Shooting Guard to establish a secondary action from the Big. This is effectively a screen-the-screener action, but the screener kept the ball for a fake handoff as opposed to a real handoff.
To illustrate, let’s consider a quick Lakers action against the Detroit Pistons. Here, D’Angelo Russell passes ahead to Jordan Clarkson. In this case, Clarkson keeps the ball and turn towards the center of the court, where he is met with a secondary action screen from Brandon Bass for the mid-range two.
Typically, keep action from a pass-ahead requires a secondary action as the Shooting Guard usually has their back to the basket out on the wing. Therefore, this action is the rarest of the four major primary actions for a Pistol series.
Option B: Dribble Handoff
Instead of a dump-and-chase option, the Point Guard may instead dribble towards the Shooting Guard and set-up a dribble handoff, or DHO. This type of initialization of the Pistol forces the on-ball defender to converge to the Shooting Guard; whereas in the pass option, the on-ball defender can change coverage away from the ball (as it is passed).
Under the DHO initialization, we head into the two primary actions: handoff and step-up. These are slight variations on the dump-and-chase primary options above, only because the ball is in the Point Guard’s hands as opposed to the Shooting Guards hands. Due to the dribble action, a screen is much more likely to occur (step-up) than in a passing initialization.
Option B1: DHO
The first primary option is to perform a dribble handoff. In this action, the Point Guard will typically set a screen to clear out a path for the Shooting Guard to drive. As the handoff action usually pushes the Shooting Guard towards the center of the court, defenses can naturally react and contain this motion.
As an example, let’s consider an example from the Houston Rockets versus Sacramento Kings game from November 25, 2016. In this example, the Rockets won the tip and went directly into a Pistol action.
The reason for such an action is to force the smaller guards on Sacramento, Ty Lawson and Darren Collison, to interact with Matt Barnes against the stronger duo of James Harden and Patrick Beverley, with Ryan Anderson stalking the perimeter. With the spreading of Clint Capela and Trevor Ariza, Sacramento’s big men of Rudy Gay and DeMarcus Cousins are pulled away from the defensive action. In the end, this Pistol action should create a 3-on-2 game of Beverley, Capela, and Ariza against Gay and Cousins.
Option B1S: Wide Pin
As we see at the end of the clip, Gay and Cousins leave Ariza on the perimeter and pack the lane. Barnes shows to push Beverley along the perimeter. Without a full head of steam, Beverley is unable to penetrate the key and Lawson is able to recover. This leads into a secondary action, the wide pin screen.
A wide pin screen is a pin-action screen that sets the screen away from the center of the court with the purpose of opening up a lane for the screened offensive player towards the center of the court.
In the Rockets-Kings example above, we see at the end of the clip that Anderson sets a wide-pin on Harden’s man (Collison). As the ball has been taken back to the top of the key, the Pistol series has run its course and the offense resets by placing the ball back into Harden’s hands.
And we see that the possession is rather lengthy, taking a total of 21 seconds to convert for a score; hardly an instant offense.
Option B2: The Step-Up
The step-up screen is one of the most common screens in the Pistol series. This action turns a DHO into a step-up screen by serving as a pick-and-roll/pop/other action that potentially frees either the Point Guard or the Shooting Guard by baiting the defense into a double team. The primary actions on a step-up is a Drive to the rim, or a pullup.
To illustrate consider this example from a Mavericks – Rockets game from December 4th, 2015. With the Mavericks railing 79-80 after a Jason Terry (HOU) layup, the Mavericks go into Pistol with Raymond Felton at the point, Chandler Parsons at the wing, and Dwight Powell at the top of the key.
Parsons hits the step-up screen on Ty Lawson, effectively pinning him on the play. With Corey Brewer hugging Parsons on the screen and Josh Smith not getting back fully; tightly guarding a non-threatening shooter in Powell, Felton drives unabated to the rim to reclaim the lead.
Option B2S: Secondary Screen
If a defense manages to disrupt the ball handler’s path, a secondary action is commonly put into place. This action is either a screen-the-screener action as before, or a misdirection screen.
For a misdirection, from the same Rockets-Mavericks game as above, we are hit with a step-up screen from Wes Matthews (DAL) that brings Deron Williams (DAL) down the sideline. Marcus Thornton (HOU) disrupts the drive and allows Patrick Beverley to recover on the play.
The secondary action is then a slip-screen from Dirk Nowitzki (DAL). The beauty of the action doesn’t come from Nowitzki’s slip screen. It actually comes from Matthews’ zipper cut. Matthews performs a zipper cut to the top of the key, rubbing Clint Capela at the elbow. This is done in tandem with Nowitzki’s feigned screen, discombobulating Capela into double-teaming Williams at the perimeter.
This action frees Nowitzki up for a wide open fifteen footer, which unfortunately hits the back of the iron and misses. Nowitzki is clearly disappointed with the miss, as such a sweet secondary action found an open shot for an elite shooter.
For a screen-the-screener action, we can refer to a Suns-Lakers game from October 29, 2010. In this play, Jason Richardson (PHX) hits Steve Nash’s (PHX) defender, Derek Fisher, with a step-up screen. As he does so, Kobe Bryant (LAL) shows to close Nash’s driving lane. With a secondary screen by Robin Lopez (PHX) on Fisher; who is closest to Richardson at this moment, Richardson pops and gets an open look for three.
Option C: Designed Plays
Up to this point, we broke down the simple actions of the Pistol series. However, this is just a brief introduction as there are literally dozens of options to run off this series. Some designed plays such as Backdoor Options or Shuffle Cuts exist. The primary take-away from this example is to develop an intuition on how to read when a Pistol formation is being established and how it interacts with a defense.
Into the Data…
Traditionally, play diagnosis is a job that performed using video analysts and video tools such as Synergy, Eagle, or AVGEN. However, with the use of templating or sketching, we are able to start performing play diagnosis through the tracking data. The challenges then translate to repeatability, exchangability, and variability.
Do we see enough instances of the play to determine it is in fact the play. This is one of the biggest challenges in play diagnosis. Having a full scope of the play helps us diagnose whether a play is a Pistol or if it some other sort of set. For starters, we see a triangle pattern in three locations with the ball at the “point guard” spot. Are we in Pistol?
To define Pistol, we require that action between the Shooting Guard and Point Guard; after all it is called a 21 series. Therefore, any old triangle pattern does not constitute a Pistol. Similarly, this is an early offense action. Sure, the same action could be run late in the shot clock; but that’s not the identity of Pistol; as one of the primary functions is to disrupt the defensive setting.
Furthermore, to diagnose the wrinkles of Pistol, we need to see the action multiple times. Not necessarily the same players, but at least the same roles. Which leads us to…
If we attempt to template a play, we cannot simply designate player motion. Instead, we must look into role motion. That is, if I have James Harden and Chris Paul on the court, at any single moment they may switch roles on the court. Simply tracking Harden and Paul will cause increases in variability and a reduction in repeatability.
Therefore, we impose a role definition, much like we did throughout this description: Point Guard, Shooting Guard, and Big. If we choose not to select roles, we are on the hook for requesting of upwards to 120 times the number of necessary plays of a single wrinkle of Pistol. Good luck with obtaining those.
The other challenge is to identify variability in the plays. Using tracking data, it is very rare to see the same pass twice. If say twenty iterations of Pistol on the right side of the court involve dump and chase during a game (this is ridiculously high by the way), then just measuring the location of the pass leads to considerable amount of variability.
Take for instance a series of ten Pistol plays ran during a recent season by the Portland Trail Blazers. Here, we isolate ten random events of right-side only Pistol. We even curated it to be Lillard to McCollum. Even with this fine-tuning, we see considerable variability of where the pass was initiated and the pass was received.
Therefore a method for capturing variability is required before we can start clustering different plays.
Given these requirements, the goals are straightforward at this point. We template plays to start to identify an encyclopaedia of offense for each team. We can start to measure quantities associated with offensive players such as burst/acceleration, reaction time to screen, and court vision. Similarly, we can start to diagnose tendencies of defenders such as reaction time or coverage range.This helps us measure the overall athleticism of a player and their associated IQ relative to specific situations.
But first, we have to understand the situations they are in. And Pistol, being a large part of offenses, is just a start.