The NBA will probably never shoot enough three-pointers.
Even though those shots are worth an extra point, they still don't get launched with enough frequency. Yes, they're harder to make than the looks from closer to the basket, but the boost in value is more than enough to make it worth the additional effort and need for increased skill.
When Grantland's Zach Lowe wrote about the new SportVU cameras back during the end of the 2012-13 season, he revealed that the triple wasn't being shot as much as analytical minds and computers would like:
The analytics team is unanimous, and rather emphatic, that every team should shoot more threes—including the Raptors and even the Rockets, who are on pace to break the NBA record for most three-point attempts in a season.
He goes on to write the following, referring to Dwane Casey, the head coach of the Toronto Raptors:
Casey is obviously right that DeRozan is a bad three-point shooter. But the analytics team argues that even sub–35 percent three-point shooters should jack more threes, and that coaches should probably spend more time turning below-average three-point shooters into something close to average ones.
Well, now the general public has access to SportVU data courtesy of NBA.com, and the results suggest that teams still aren't taking this philosophy to heart. There's a striking lack of willingness to fire away from downtown with reckless abandon.
One of the many stats that the camera tracks is simply the amount of touches that a player generates during the course of a game. It looks at where the touches occur (in the half-court set, in the frontcourt, on the elbows, etc.) and, of course, the results of the touches.
Take a gander at the qualified players who are recording the most points per touch. In this case, qualified means that the player is involved in the action (arbitrarily set at 20 touches per game or more) and has played in a significant number of games (at least five at this early stage in the season):
- Klay Thompson, 0.56
- Brook Lopez, 0.53
- Jodie Meeks, 0.50
- Nick Young, 0.45
- J.J. Redick, 0.44
- Eric Gordon, 0.43
- Kevin Martin, 0.43
- Kevin Durant, 0.43
- Gerald Green, 0.41
- Anthony Davis, 0.41
- C.J. Miles, 0.41
- Jamal Crawford, 0.41
Doesn't that feel a little bit strange? The list is comprised of jump-shooters and then two ultra-efficient scoring big men who may be playing at slightly unsustainable levels of efficiency on the offensive end of the court.
It's a trend that continues on and on.
Well, the idea for any elite team should be to strike a balance. If a player is scoring an inordinately high number of points per touch, then he should be fed the ball more. As long as he's over the NBA average (which is clearly the case for these league leaders in the category), it makes sense and is beneficial for him to become more involved.
There's always been a trade-off between volume and efficiency in the NBA. As players start to earn more playing time, it's tougher for them to maintain their per-minute numbers. As they get more involved on offense, it's tougher for them to maintain their efficiency.
The same principle applies here, and it's worth milking three-point shooters until they're only shooting slightly above the league average. If that involves asking Stephen Curry to shoot 38 three-pointers per game, so be it.
I honestly have no idea if that would be mathematically beneficial or not, although I suspect that he could do that while still shooting at least 28 percent from the field. And as you'll see in a bit, that would mean it's worth doing.
According to Hoopdata.com, the league average on jumpers from 16 feet out to the three-point arc was 38.4 percent. That means that the average shot was worth 0.768 points.
Since a three-pointer is worth an additional point, the percentage can be much lower.
For a triple to have that same value (0.768 points per shot), players would only have to make them 25.6 percent of the time on average.
Now, let's do the same exercise for the other areas of the court.
|Range||Percentage Made in 2012-13||Points Per Average Shot||Necessary 3P%|
|16 to 23||38.4||0.768||25.6|
|10 to 15||41.9||0.838||27.9|
|3 to 9||39.9||0.798||26.6|
Based on last year's numbers, it's awfully tough to justify shooting a mid-range jumper over a three-pointer. Which players in the NBA can't make at least 28 percent of their triples? I'd wager that—with some practice, of course—even most centers could top that mark.
Now as you can see, it's not worth passing up looks at the rim, but that's a different story entirely.
Last year, the NBA as a whole made 35.9 percent of its looks from three-point land. You do the math, but last time I checked, 35.9 was greater than 28.
Basketball-Reference reveals that 204 players are currently shooting better than 28 percent from beyond the three-point arc. Among the 206 players who have taken at least 10 triples this season, only 43 are checking in below the magical 28 percent, and it's a safe bet that some of them will start to rise up when there's a larger sample.
Steve Novak, for example.
If it doesn't already, that name is going to mean a lot more to you in a second.
Interestingly enough, Lowe's original article confirms this general principle.
He quotes Alex Rucker, the Toronto Raptors director of analytics, as saying the following:
When you ask coaches what's better between a 28 percent three-point shot and a 42 percent midrange shot, they'll say the 42 percent shot. And that's objectively false. It's wrong. If LeBron James just jacked a three on every single possession, that'd be an exceptionally good offense. That's a conversation we've had with our coaching staff, and let's just say they don't support that approach.
So why don't they support the approach?
Here we can turn to Casey, whom Lowe quoted as revealing, "You can shoot as many threes as you'd like, but if you don't make them, that philosophy goes out the window. There's always going to be disagreements. Analytics might give you a number, but you can't live by that number."
Unfortunately, Casey is correct.
A coach can't live strictly by the numbers because he isn't operating in a world that is free from the possibility of termination and ridicule from the media. In a perfect scenario, numbers would dictate decisions without fail, but no such scenario actually exists.
It's all about loss aversion, a psychological principle that attempts to explain human behaviors by claiming that we're more likely to avoid losses than pursue gains. Here's my favorite example:
Pretend that you and I are complete strangers—which we likely are in real life as well as this hypothetical. All of a sudden, I come up to you and promise you that I'll give you $100 tomorrow morning. I'm completely emphatic about this, and you believe me with all your heart.
But when the next morning comes around, I renege on my promise and don't give you anything. You're going to be upset, right?
Well, not as upset as you'd be in the following scenario.
This time, we're still complete strangers, but I walk up to you, grab your wallet and take $100. Now you're really mad.
Objectively, you lost $100 in each scenario. But in the former you avoided a gain; in the latter you actually experienced a loss. Even though they're mathematical equivalents, you're psychologically more upset by the loss.
And that's why loss aversion matters.
What happens if a coach tells his players to start shooting three-pointers at the expense of everything else, and it doesn't work? Not because the numbers don't back it up, but because there's an inherent flux when it comes to human activities. Some games, teams will make more than the average, and they'll fall below it other times.
But eventually, it all balances out. Just not fast enough in some cases. It works just like a coin.
If you flipped a coin 10,000 times, the percentage of heads and tails would each be right around 50 percent. But if you flipped it just three times, there's a much greater chance of it varying away from the expected mean.
The media backlash would be too much if a team went cold at the beginning of this strategy. Shot selection would inevitably be questioned, and the coach would be under too much pressure to change his methodology. Without letting the sample grow large enough, it's impossible to guarantee success.
Let's go back to the coin.
Say you're flipping a weighted quarter, one that is going to come up heads 70 percent of the time. You're banking your success on that coin landing on heads each time, but you could get into too big a hole after just five flips or so because there is still a 30 percent chance it doesn't work each time.
That's the scenario coaches find themselves in. Even if the odds will work in their favor eventually, their jobs are too dependent on the early results. And it works the same way for players.
Additionally, NBA players are taught to find the open shot, and once you learn something, it's tough to break that habit. Just look at how much trouble everyone is having remembering not to touch the ball after a shot is made this season. The delay of games are racking up, simply because that's an instinctual move.
So too is eschewing a contested three-pointer for an open mid-range attempt, even if the former is the better shot.
And for the most part, it is.
The next time you're watching an NBA game and see a player jack up a contested triple, don't shake your head. Instead, let a smile creep across your face even if the shot clangs off the rim. Even if he doesn't know it, that player is starting to play more efficient basketball.