Of the two major sports that we will discuss, the National Basketball Association of today would undoubtedly look the most foreign to a time traveler from 15 years ago. Teams are now adopting “positionless basketball,” seven foot tall giants are now shooting threes (!??), and the game is being played at a more rapid pace than in recent memory. The games are fast-paced, exciting, and full of scoring, mainly from three pointers and dunks. In this breakdown we will expand our view beyond the game tape and into the land of numbers and spreadsheets to analyze the evolution of shot preference, and how this evolution is changing offensive spacing.
The Rise of the Three-Pointer
NBA players are shooting a TON of threes. In fact, the three-point rate (the percentage of field goal attempts that are three-pointers) has increased every season for the past ten seasons. Before the 2019-20 season was suspended, teams attempted over 38% of their shots from beyond the arc, up 16% from ten years prior. Even before the 2009-10 season the three-point rate was up 5% from the 2001-02 NBA season. So essentially, the three-point rate is increasing at an increasing rate. In fact, the 2017-18 Houston Rockets were the first team in NBA history to actually shoot more threes than twos en route to a 65-17 record. The following season, Rockets guard James Harden shot an unprecedented 13.2 threes per game, totaling 1028 on the season. For reference, in 1979-80, the Rockets as a team only attempted 379 threes, which was actually the third highest rate in the league that season.
Reggie Miller, widely considered one of the greatest shooters in NBA history attempted 6.6 threes in his most trigger happy season back in 1996-97. This rate would place Reggie at 29th in the 2019-20 season, next to guys like Terry Rozier, Taurean Prince, and Lonzo Ball. These guys aren't considered deadly weapons from deep. So where are all these shots coming from? If teams are shooting an astronomical amount of threes, which shot type is decreasing?
The Simultaneous Death of the Midrange
Many consider the midrange shot to be an art, craftily mastered by some of the greatest to ever pick up a basketball. Dominant players like Michael Jordan, Kobe Bryant, and Dirk Nowitzki centered their whole offensive repertoire around the midrange jump shot while leading their teams to championships. But especially in recent years, the midrange shot, characterized as a field goal attempt ranging from around 8 feet from the hoop to having your toe on the three point line, has all but evaporated from NBA offenses.
The chart above depicts the three point attempt rate compared to the long twos rate (16+ feet) in the last 20 seasons.
Some of the most iconic shots in basketball history are from the midrange (think Jordan sealing his sixth championship against the ‘98 Jazz). Teams are now avoiding the midrange shot at an alarming rate. Shots from 16 feet to just before the three point line now constitute only 7.6% of an offense’s shot attempts, a rate that has sharply decreased for 12 consecutive seasons. In 2000-01, teams shot 6% more of these shots than threes. That number is now 30% in the opposite direction. Contrasting with the threes, these midrange shots are actually decreasing at an increasing rate. In today’s NBA, the San Antonio Spurs “lead” the NBA, shooting 14.7% of their attempts from the midrange (“midrange” here is defined as 16 feet to just before the three point line). Nine years ago when teams averaged 21.4% of their shots from midrange, this rate would place the Spurs dead last in attempts.
So what’s going on?
The simplest way to explain this evolution in shot distribution is to look at a very basic formula. 3>2. All things being equal, it is more beneficial to shoot threes than twos simply because each made shot earns your team 50% more points. But all things are not equal; common intuition is that the farther you are from the basket, the more difficult the shot will be to make. But how much better should you be at shooting twos to pass up on a chance at that extra point?
Let’s play out a scenario where you are the point guard and you need to decide where to pass the ball. Player A (we’ll call him Jrue Holiday) is standing in the corner beyond the three point line, 22 feet from the basket. His 2019-20 three point percentage is 35.7% which is exactly what the league average is on the season on such shots. How convenient! Player B is standing just beyond the elbow, somewhere in the range of 16-23 feet from the basket and successfully makes these shots at the league average 40.5% (we’ll call him Lou Williams). It’s important to note here that the NBA three point line is 23.75 feet from the basket except for the corners where it is 22 feet. Theoretically, Holiday could be closer to the basket and score more points on a made shot than Williams. Both players are calling for the ball and are ready to catch and shoot. Who do you pass to? An informed way to answer this question is to choose the option that maximizes the expected points for the possession.
For the Mathematically Inclined...
Let’s do some math. We multiply Holiday’s three point percentage of 0.357 to the value of the points earned if he makes the shot. This is an expected value equation where we multiply a probability with a benefit. (0.357*3) yields 1.071 which is our expected points if we pass Holiday the ball. Basically if we simulate this scenario 100 times and pass to Holiday for three each time, we would expect to score 107 points. For comparison, the 2018-19 NBA champion Toronto Raptors scored 113 points per 100 possessions. If we passed Williams the ball 100 times, we would expect to score (0.405*2*100) only 81 points. In order for Williams to be at least as efficient as Holiday, he must shoot above 54% on his long midrange shot. Keep in mind that Williams represents an average NBA shooter in this scenario. A given player must shoot 14 percentage points above average for the attempt to yield a better return than an average three pointer. So Williams’s benefit of standing closer to the basket is met with the risk of needing to shoot a whole lot better to achieve the same level of efficiency.
Since reliable shot tracking data became available during the 1997-98 season, the best statistical shooter on long twos (16+ feet) is John Stockton at only 48.2%. With the influx of threes and death of the midrange, it’s safe to say that NBA teams are noticing this equilibrium imbalance and are adjusting their strategy accordingly.
(Image above courtesy of "Sprawlball" by Kirk Goldsberry)
Wide Open Spaces
So we have established that teams should opt for a three pointer over a midrange jumper when they have the choice and we have reason to believe based on trends alone that teams are following these instructions. Now let’s look at some game tape to see how teams have evolved their offenses to engineer more opportunities for open threes.
Here we are shown a clip from a matchup between the Phoenix Suns and the Toronto Raptors from February 27, 2000. Phoenix center Luc Longley catches the ball in the post with five seconds remaining on the shot clock and looks to pass. Raptors wing Vince Carter is forced to guard two defenders and smartly positions himself under the hoop to defend any layups or dunks from cutters. Longley has the option to pass to Clifford Robinson beyond the arc on the right wing, or Tom Gugliotta who is standing about 20 feet from the basket in the corner. Phoenix forward Tom Gugliotta spaces the floor enough to have his shot unaffected by Carter’s defensive recovery, but stops about 2 feet short of the right corner three point line. Instead of swinging the ball to Robinson who is shooting 37% from deep on the season, Gugliotta opts for the long two and misses. Note: I want to stress that even if Gugliotta makes the shot, it was still a bad decision. He should have positioned himself beyond the three point line for the opportunity at an extra point.
This next clip is from Game 7 of the 2004 Western Conference Semifinals between the Minnesota Timberwolves and the Sacramento Kings. Minnesota wing Latrell Sprewell catches the ball beyond the arc and immediately starts a pick and roll with center Ervin Johnson. This action probably would have led to an easy score, however, forwards Kevin Garnett and Wally Szczerbiak are both positioned terribly. They are standing in an inefficient zone on the court; (a shot from that distance is a win for the defense) if they are in fact passed the ball they are shooting a low value shot. Not only that, they also bring their defenders with them, meaning that the lane is now clogged for Sprewell and Johnson’s pick and roll. If either player breaks free of the defense, Sczerbiak and/or Garnett’s defenders are in a position to quickly help and stop the penetration. Most teams and players now realize this and space the floor accordingly.
Moving on with the play, all is not lost because Garnett sets a really nice flare screen to set up guard Sam Cassell with a wide open three. This is a high value shot; Cassell shot 39.8% from deep on the season, a career high. So what does Cassell do?
He takes one dribble inside the arc, towards the defense, into a long contested two that he misses. They never learn. We’ll come back to this play later.
The last play we’ll analyze is from February 6, 2020 in a matchup between the Houston Rockets and the Los Angeles Lakers. Houston prides itself on their analytically sound approach to the game and their general manager Daryl Morey has been a key figure in pushing the sport to its limits.
Notice right off the bat how much more spaced the floor is (Robert Covington is so far out of the play that he is literally off the screen!). Guard Russell Westbrook is able to back down his defender pretty deep without any help defense converging on him. Once Lakers forward Kyle Kuzma finally double teams the ball, Westbrook is able to kick out the ball to Covington who swings the ball to Eric Gordon who finds P.J. Tucker for an open corner three. Covington and Gordon both pass up good shots for a great shot. Houston’s spacing makes it such that their defenders are forced to cover more ground, driving lanes are open, and they are in positions themselves to take a high value shot.
The Crystal ORB
Let’s revisit the Timberwolves possession from 2004. The sequence ends with Szczerbiak securing the offensive rebound and putting the ball in the hoop for two. This is representative of the trade offs that teams now have to make to space the floor. If players now are spaced in a way that creates more three point opportunities, they are also spaced in a way that creates less offensive rebounds because they are further from the basket and thus out of position to secure an “extra possession.” This is less important now that the shot clock resets to 14 seconds instead of the full 24 after an “o-board” but it is worth investigating how much teams are sacrificing with their spacing. The graph below depicts the only real quantifiable way to show that offenses are more spread out without using player tracking data.
Now, offensive rebounds are valuable for obvious reasons. They prevent the opposition from gaining possession which leads to less points scored against you, and allow you an extra chance to score points. But is the NBA’s new love of spacing to create more threes actually decreasing the offensive rebounds to a point where our offenses are becoming less efficient? Is spacing doing more harm than good? Let’s examine.
The chart above depicts the average Offensive Rating (Points per 100 possessions) in the last 21 seasons. As offensive rebounding rates are declining, offensive efficiency rates are still climbing, indicating to us that the teams should be willing to trade off offensive rebounds for more three point attempts.
This change in philosophy around spacing has led to the development of the “stretch four,” a power forward who foregoes posting up to spotting up for three. Players like Kevin Love who are able to adapt their skill set to the modern NBA are able to prolong their careers and add value. Asking a seven footer to work on his offense beyond five feet from the rim would’ve seemed absurd 15 years ago while today, it’s seemingly necessary for a big man to have a jump shot “in his bag” in order for him to make any noticeable impact on the league.
Conclusion/Predictions
NBA franchises are finally coming around to the realization that threes are worth more than two as is illustrated by my research on changing play styles. Teams should prioritize attacking the basket which creates high percentage layups and increases the opportunities for free throws, and shooting threes in their offenses. Long twos and packing the paint are both inefficient and thus we are seeing those styles being weeded out by offensive systems. The three point rate continues to climb and I predict we will see even more threes attempted before reaching an equilibrium. The Golden State Warriors dynasty put to rest the fallacy that jump shooting can’t win championships. As jump shooting big men become more valuable, we are going to see more young players with a jump shot and as ugly as it may look, we’re going to see more centers shooting threes and the term “stretch five” will arise. I personally think the NBA of the current age is more fun to watch than ever but I worry that too many threes (which also means less shot attempts actually being successful) will damage the product, especially if the game becomes too homogeneous. If the NBA wants to combat this they could experiment with moving the three point line back but I don’t think that is necessary in the very near future. Teams have always been paying attention to what works and what doesn’t but the added power of having more data available is pushing the league into a rapid evolution.
Comments