The Value of Clutch
The NBA is in a full-fledged Age of Enlightenment, an awakening that at times has manifested in a pitched battle between long-held truisms and a new way of looking at things. There is a gleeful desire by many to rip back the curtain and point out that things do not actually exist the way we have seen them, that precedence is not a substitute for reality. But that instinct can speed the process beyond prudence, causing more strife than is necessary. Although there are exceptions, I don’t think this is a disagreement between noble, learned scientists and dogmatic, fingers-in-the-ears, medieval clergy. There is a dichotomy to be sure, but the dividing line seems to be more begrudging cautiousness, than idealogical repulsion.
When it comes to clutch performance there seems to have been a judicious meeting of the minds. I sense a general acceptance that handing the ball to your best isolation scorer and clearing out may not be the most effective crunch-time offense. However, it’s also generally acknowledged that context is the ultimate authority in these situations, giving any fan an out when talking about their favorite team. However, if evaluative consensus is ever to be reached on clutch performance, it’s importance must first be established.
The first defense for any Miami Heat fan confronted on LeBron James‘ historic close-game struggles, is to point out their team’s ability to blow opponents out in the first three quarters. While there is truth to this, dominant, resounding wins are rarely gained in a game’s closing seconds, it should take nothing away from the importance of a team’s ability to execute in clutch situations. While every team would like to avoid close games, on the positive side of the equation, they are an inevitability. Even the best teams in the NBA find themselves in scenarios that run down to the wire. The Chicago Bulls have the league’s best Net Rating at +8.6 points per possessions. They have also played 25 games this season where at some point during the last five minutes, neither team was ahead by more than five points. Being good doesn’t protect you from close games. But the ultimate question is, what exactly do you gain by making the most of those tight contests?
Pythagorean Expectation, or Pythagorean Win Percentage, is a way of calculating how many games a team should have won given their point differential (how many points they score against how many they allow). Since it’s based on point differential, Pythagorean Win Percentage gives a more complete picture of a team’s abilities than just looking at their win totals. Every year however, the majority of NBA teams do not end up with a win total that matches their Pythagorean Expectation. Rarely will a team drastically under or over-perform their expectation, but just as rarely do they hit exactly. The reason for this margin of error is clutch performance.
The table below covers the 2010-2011 season. For each team it shows their Pythagorean Expected win total, their actual win total, and the difference. It also shows their season-long Net Rating (ORtg. – DRtg.), their clutch Net Rating, and the difference. I assume the standard definition is familiar but, for the sake of clarity, we’re defining clutch situations as the last five minutes of a game or overtime, with neither team ahead by more than five points.
[table id=48 /]
And here is the data from this season.
[table id=49 /]
What you’ll see in both seasons is a strong connection between teams that out perform their Pythagorean Win Expectation and those that either maintain or improve their Net Rating in clutch situations. For last season there was a correlation of 0.767 between a team’s win difference and the difference in their Net Rating from the entire season to the clutch. This season that correlation is coming in at 0.550. Those are fairly significant numbers and I feel confident that looking back to previous seasons would yield similar results. Pinpointing the value of positive clutch performance is actually quite simple then – it allows you to overstep the boundaries of your talent, it allows you to be more than you are.
Take for example, the 76ers this season. When I ran this data on Monday they had the 5th best Net Rating in the league at +6.0. They had the 15th best Win% in the league, in large part because their Net Rating in clutch situations is -13.7, a swing of 19.7 points per 100 possessions. The 76ers are under-performing their Pythagorean Expectation by 8 wins. Looking at just their season long numbers, they appear to be a top five team. Their inability to carry that top five performance through the full 48 minutes means they are on the edge of sliding out of the playoffs.
While there is room for context to dictate how we evaluate clutch performance, we have to acknowledge that it is a definable phenomenon with tangible value. A five-point lead in the 4th quarter is much more valuable than a five-point lead in the 1st quarter. This is a fact which can be agreed upon definitively and universally, and the chronological constraints of a 48 minute game make it so. If you trail by five entering the 2nd quarter you have 36 more minutes to make up the deficit. If you trail by five with just five minutes to go, your opportunities for taking the lead have shrunk dramatically. In his presentation at the MIT Sloan Sports Analytics conference on using cumulative win probabilities to rank college teams, Mark Bashuk found that the first 25 minutes of a college basketball game were worth 36% in determining the outcome. The last 15 minutes were worth 64%. While those numbers may be accentuated by the difference between the college and pro game, there certainly wouldn’t be a 50-50 split in the NBA either.
The ‘Moneyball’ movement is about finding value where other’s aren’t looking, and there is value to be had for both good teams and bad in becoming a 4th quarter juggernaut. Every team, regardless of how good they are, will eventually find themselves in a close game. Executing down the stretch, when the clock, officials, and opponents get tight is a chance to steal success, to win games you probably shouldn’t. It’s a chance to be more than you are.
Statistical Support for this story from NBA.com