Many people erroneously believe that NBA games are “decided” in crunch time. Hopefully, everyone is starting to realize that it’s not, and that a team’s clutch performance has about as much to do with it winning as any other five-minute chunk of the game. One common explanation for why clutch play is so important is that the “game becomes different” down the stretch. After revisiting Late-Game Bias during the Sloan MIT Conference, I think it’s important for people to understand the game dynamics in the clutch and throughout an NBA game.
First and foremost, the game is always subtly changing. The simplest explanation for this is that with human performance, it’s nearly impossible for the game not to change: we fatigue during athletic movement for a variety of reasons. Furthermore, the brain has been exposed to a new environment during the game; a play executed in the first quarter will look different to the defense in the fourth quarter. This means we can expect changes throughout the game – not just during crunch time – based on both strategy and on physical performance.
Look at what happens with shooting. In the opening minutes of the game, more shots are taken within 15 feet than at any other point in the game (55.2%). The number drops lower and lower as we progress through the game. Visually, it looks like this:
As we move through the game — left to right for each of the four stats presented above — teams:
- Take more outside shots
- Specifically, they take more 3-pointers
- Assist on a lower percentage of field goals
So, this should be crystal clear: The game does not just change in the clutch, it changes constantly.
As the game progresses, teams are either settling for more and more outside shots, or defenses are doing a better job of forcing them. The fact that assist % continues to decline strongly suggests that more isolation basketball is being played and in general, players are not taking more outside shots because they are suddenly getting more open outside looks.
By the time we reach the clutch, only 50.5% of shots are inside 15 feet (almost exactly the same share as at the start of the 4th quarter). Only teams start launching more 3-pointers while assist percentage plummets. In the clutch, a whopping 27% of shots are 3-pointers, and without digging deeper, it’s mathematically unlikely this is a result of teams heaving desperate 3′s in the final seconds of four and five-point games (For reference, Boston, Cleveland and Chicago had 9 shot combined shots last season. At that rate, 30 of the 2303 3-pointers in question would fit that bill, or about 1%.)
What does efficiency look like throughout the game on those different kinds of outside shots?
While teams are taking more and more outside shots throughout the game, they are converting the outside shots less efficiently. FG%, long 2-point% and 3-point% all slowly decline, with the exception of the start of the second half. Presumably, teams are fresher and playing their best lineups, and in the first five minutes of the 3rd quarter 3-point shooting (and as a result, eFG%) jumps back up to its opening game levels.
What’s Happening to Offenses at the End of Close Games?
Look at what happens in the last five minutes of games that aren’t close but aren’t blowouts (6-12 point games, included in dark red in both graphs above): FG% still drops, 3-point% still drops, fewer shots are assisted and teams launch 3-pointers (and as a result, outside shots) more frequently than they do at any previous section of the game. This is similar to what we see when the game is closer, so perhaps the game progressing has more to do with these tendencies then something unique about “clutch” situations or the closeness of the score.
We can telescope deeper into the game and examine the “Super Clutch,” defined as a 3-point game in the final two minutes. Below is a table comparing the changes in outside shooting during the clutch and super clutch compared to the rest of the game (Note: The changes in shooting percentages within 15 feet are similar.)
Get the rest of this article over at Back Picks.