Column: How has analytics changed shot selection?
“High risk, high reward” is a familiar phrase. It rules investment strategies and dictates policies. The basic principle of risk and reward states: if an action is risky, it needs to be highly rewarding to compensate for the risk taken. Conversely, if an action is safe, it is likely not rewarding, or that if something has a low reward attached to it, the risk must be minimal to justify taking a smaller reward.
This concept of risk and reward seamlessly translates into basketball. The rise of advanced analytics has caused a new way of thinking and looking at the game , leaving teams scouring data searching for ways to increase upside potential while minimizing potential risk.
Shot distribution is one of the most important factors of an effective offense. Although often overlooked, a team’s ability to create and take smart and efficient shots is paramount in offensive efficiency and can make massive differences in shooting percentages, total offense, points per possession and other metrics.
The 40/40/20 rule is often used to describe shot distribution and is excellent in balancing a risk-reward tradeoff. It states that 40% of a team’s shots should be from the 3-point line, another 40% at the rim and 20% from the mid-range. This means that a combined 80% of a team’s shots would be at the rim or from behind the 3-point line, or what are considered “efficient shots.” Based on NBA averages from Cleaning the Glass, abiding by the 40/40/20 rule would result in scoring an average 148.12 points per game — more than enough to win a majority of matchups. Analyzing the data, particularly points per shot attempt, provides clear insight into the efficiency of certain shots and the make-up of the 40/40/20 rule.
The value of the three pointer is obvious and becoming clearer every day. It has a value 50% higher than its two-point counterpart. It may not seem like much, but an entire point is enough to decide the outcome of a game. In a league where teams average 115 possessions per game, the difference between a two-point value and a three-point value can be consequential.
Shooting a three does come with a lower shooting percentage, but as the risk/reward analysis demonstrates, teams can afford to shoot a much lower percentage, because the three is so much more valuable than the two. The NBA league-wide average is 37% from three which leads to an average of 1.12 points per shot attempt.
Shots at the rim are extremely effective shots. The league boasts a 64% average shooting percentage at the rim. Shots at the rim also put the shooter in position for an offensive rebound and they also draw fouls at a high rate. One in five shots at the rim draws a shooting foul, resulting in either two free shots or an and-one opportunity for a three-point play. 64% shooting with a 20% foul rate equates to an average of 1.12 points per shot.
Mid-range jumpers are anything from about four feet away from the rim all the way to the arc. Knocked down at only 42% they often leave the shooter in poor position to get the rebound, but usually too deep in the floor to quickly get back to defend. Despite the tendency to take closely contested shots, the rate of drawing a foul is negligible. It is possible for shooters to create space between themselves and their defenders to take an open look, often by taking a fadeaway or a floater, both of which can decrease accuracy and foul rate. A mid-range jumper — especially the deeper mid-range shots (taken from about 14 feet and beyond) — is quite risky and unrewarding. These shots combine the worst parts of a shot at the rim and combine them with the worst parts of shooting a three without adding the positive aspects of either shot. It is only worth as much as a layup but knocked down as often as a three. The low percentage combined with its low value defies the risk/reward analysis and makes it statistically the worst shot in basketball. 42% shooting on a shot worth two points results in 0.84 points per shot.
The 40/40 allocation of the two types of most efficient shots may not matter, however, because they are both efficient shots. Offensive efficiency is often measured by points per possession. In this metric the goal is to average at least one point per possession. Put simply, only shots at the rim and behind the arc can do that, because they average more than 1.0 points per shot while shots from the mid-range come in below that line. So, in order to stay above one point per possession, teams need to supplement their offense with threes and shots at the rim. So, the key to efficiency is not necessarily shooting more shots at the rim or more threes, rather, it is limiting shots from the mid-range. By keeping the rate of mid-range jumpers down, at or around 20%, offensive efficiency goes up, regardless of the types of shots that fill the other 80%. It’s no coincidence that according to Ben Falk of Cleaning the Glass, the top 5 teams in the league in Loc eFG% (location effective field goal percentage) are the bottom five in the league for mid-range shots taken. Additionally, the bottom three mid-range shooting frequency teams were all in the top 5 in eFG%.
In response to analytics, three-point frequency is skyrocketing, and one statistician thinks it could go higher. David Locke, one of the most articulate and outspoken proponents of basketball analytics and shot selection, also has apparent qualms with the mid-range. At length, he has used statistical data to disparage that shot. He has explored the idea of teams shooting well over 50% of their shots from beyond the arc. When asked where the three-point revolution is going, Locke replied, “I think we’re going to 60 to 65 percent of all shots are threes and about 30 percent are at the rim and about 5 percent are mid-range shots.”
Citing the inevitability of defensive adjustments, he immediately adjusted himself saying that maybe we never get that high. But he does maintain that due to improved analysis, teams will continue to increase the number of three-point attempts. This scenario laid out by David Locke would break down to 65/30/5, resulting in 75 threes, 35 rim shots, and 5 mid-range shots per game equating to an average of 147 points per game. This seems to disprove the 40/40/20 rule, and instead indicates that shooting as few mid-range shots as possible — and replacing them with more efficient shots — is the number one way to increase shot selection efficiency and, in turn, total offensive production.
It does seem that given all the variables and nuances of dealing with an actual basketball game, the 40/40/20 rule is still a good standard for shot selection. Shooting zero mid-range shots would be ideal, but opposing teams know this and will try to take away the three and the rim leaving an open mid-range opportunity. Of course, Jerry Sloan was once credited as saying “maybe there’s a reason you’re open, kid.” This idea is applicable, but an open look can certainly be dangerous, and a great shot is almost always better than a good shot.
@pshark14
—sports@usustatesman.com