Pages Navigation Menu

Creating Shots In The Clutch

Icon SMI Media

I’m always a big fan of new ways to graphically represent basketball data. Corey Schmidt and his latest piece at Basketball Prospectus got me fired up with a nifty picture of clutch performance. Corey looked at a few of the projected top teams in college basketball and plugged some of their key rotation players into his Clutch Gauge. The Clutch Gauge shows each player’s Points Per Weighted Shot (a close relative of TS%) in clutch situations, relative to the percentage of their team’s clutch shots taken by that player. The graph is divided into four quadrants, easily distinguishing players from high-usage to low-usage and high-efficiency to low-efficiency.

The issue of clutch performance is one that just won’t go away. The tangibility, quantifiability, and style of player performance in close games has been discussed and dissected ad nauseum. Just this week Magic Johnson poked fun at how LeBron James has performed in those crucial moments.

Inspired by Corey’s Clutch Gauge, I decided to see if I could roll his idea in with some of the shot-creation work I’ve been doing, and add a little to the discussion of what an efficient crunch time offense looks like. I’m calling my tweak to his technique the Clutch Creation Gauge. The changes are small – I shifted the Points Per Weighted Shot over to TS%. Then looking at only statistics for clutch situations, I compared the TS% to the percentage of a player’s made baskets which were assisted on. All the numbers come from the clutch stat database at Clutch situations are defined as the last 5 minutes of a game or overtime, with neither team ahead by more than 5 points.

The common belief that to win in the NBA you must have a player who can create his own shot at the end of the game is ingrained in the minds of many. My contribution is the addition of the qualifier “high percentage” to the word “shot.” Creating a shot is nothing more than throwing the ball at the basket. The point is to create shots that are likely to go in. This Clutch Creation Gauge is meant to help us identify some players who create their own shots in pressure situations, but also do it in an efficient manner. I choose 16 players to start with, that I think most NBA fans would identify as guys they’d like to take a game-winning shot.

The Clutch Creation Gauge ends up with the same four quadrant model as Corey’s work. TS% is plotted on the Y-axis, Ast’d% on the X-axis. The Y-axis mid-point of 0.543 represents the league average TS% this season. Looking at the four quadrants gives us this breakdown:

  • A – Low Ast’d%, High Efficiency
  • B – High Ast’d%, High Efficiency
  • C – Low Ast’d%, Low Efficiency
  • D – High Ast’d%, Low Efficiency
If you’re looking for a player to create a shot at the end of the game, and you’d also like him to make that shot, then the players in Quadrant A are your statistical best bet. I have to admit I was shocked to see Monta Ellis score so efficiently in these situations. It’s especially amazing since his TS% for the entire season was just 53.6%, 6.2 points lower than what he shot in clutch situations. I was not surprised to see Tyreke Evans near the bottom in terms of efficiency. Injuries slowed him down last year, but his shot selection still leaves a lot to be desired. The thing that I thought was most striking was that of the 16 players I selected, each having a reputation as a potent individual scorer, just 3 created more than half their clutch shots themselves, while still shooting above the league average. This pattern held true for the larger sample as well. 71 players attempted at least 18 shots per 48 minutes of clutch play. Only 14 of those 71 had the low Ast’d% and high TS% to fit into Quadrant A:
There were are certainly a few surprises there, and small sample sizes are definitely having an effect. One player who did stand out was Eric Gordon. His clutch numbers were so impressive that I had to add him to the previous graph.
Take a look:
Gordon’s clutch TS% of 72.5% was almost 20 points higher than his average for the full season. I have mental list of guys I’d like to take a final shot, and Gordon just jumped up quite a few spots. But again, when so few guys have this ability to create and create well, why maintain a focus on that one narrow avenue to success. Another 14 of those 71 players we mentioned earlier fit into Quadrant B. They shoot above the league average but more than 50% of their made baskets are assisted on. There are more options for successful scoring opportunities in the clutch than the 1-on-5 attack.
%d bloggers like this: