What's Clutcher Than Being Clutch

Very High Leverage, Clutch Squared, things of that nature

The beauty of the Elam Ending is that even in a blowout we still get a memorable finish. That’s how things played out at the All-Star game on Sunday night when Team LeBron won by 20 points after Damian Lillard hit a 40 foot walk off shot to end the game. All-Star game stats don’t count for anything, but Lillard’s closing shot was one of many game-altering plays he’s made in clutch situations this year. Pound-for-pound, no one has been a better closer than Lillard this season

In the table below I’ve pulled together a few numbers that summarize Lillard’s late game heroics. I’ve also handpicked a few other clutch players (or guys that I’ll get yelled at if I don’t include) to serve as points of comparison.¹

The NBA defines clutch time simply as the last five minutes of a game when neither team leads by more than five points. In these situations, Lillard has an Effective Field Goal Percentage (eFG%) of 78 percent.

Usually this is where I tell you just how good an eFG% of 78 percent is among players who have taken at least some minimum number of field goals. But I hesitate to do that because no matter what threshold I set, it’ll either understate Lillard’s incredible shooting efficiency or exclude someone unfairly. So I’m going to try something different (if you hate it, sound off in the comments).

Instead of showing you everyone who has taken at least one attempt, the chart below shows the player with the highest clutch eFG% at each attempt number. In other words, the chart shows you who has the highest eFG% on one attempt, two attempts, three attempts, and so on and so forth. This makes the chart less crowded by removing some of the less relevant data points and it also allows you, the reader, to decide what field goal threshold you want to set to contextualize Lillard’s clutch shooting efficiency.

Based on the chart above we can see that no one has a higher eFG% than Lillard on more attempts. In fact, the first player to the left of Lillard on the chart to have a higher eFG% is Devonte' Graham, who has only taken 12 clutch attempts this season.² (That means we could technically get away with saying Lillard’s 78 percent eFG% in clutch situations is the highest among anyone who has taken at least 13 attempts. But that sounds dumb.)

While the NBA’s definition of clutch is fine in most cases, it leaves some meat on the bone. For example, buzzer beaters get lumped in with shots that occur when there’s still five minutes left in the game. Both of those types of shots have different levels of impact on a team’s chances of winning, but they’re considered equally clutch by the NBA’s definition.

I don’t mean to suggest that this is big a problem. But sometimes it can helpful to differentiate between the most critical shots from the still-important-but-not-as-critical ones.

Recently, Darryl Blackport, the creator of pbpstats.com, unveiled a new filter on his site which provides player and team stats under different “leverage” conditions. By developing his own win probability model (based on features like score margin, time remaining the game, and the pre-game moneyline odds) Blackport has been able to classify each possession of a game as either Low, Medium, High, or Very High Leverage based how much the result of the possession impacts a team’s overall chance of winning (or losing). The higher a possession’s leverage, the bigger the impact it has on the outcome of the game.

For this post, I was just interested in shots that occur in High or Very High Leverage situations. On his site, Blackport lists some common examples of these types of possessions:

High Leverage

• Down 5 points with 7:06 left in the 4th quarter

• Tie game with 6:38 left in the 4th quarter

• Down 4 points with 1:29 left in the 4th quarter

Very High Leverage

• Down 2 points with 1:36 left in the 4th quarter

• Down 1 point with 1:02 left in the 4th quarter

• Down 2 points with 0:41 left in the 4th quarter

Overall, this is similar to the NBA’s definition of clutch, but it’s a little more flexible in deciding what gets counted as a clutch shot and what doesn’t. More importantly, the leverage thresholds on pbpstats.com allow us to differentiate between clutch shots that are taken with several minutes remaining in the game (High Leverage attempts) and clutch shots that ice or decide a game (Very High Leverage attempts).

Lillard has an eFG% of 61 percent in High Leverage situations and 81 percent in Very High Leverage situations. This tells us that while Lillard has been efficient overall in the clutch, what’s really driving his success in crunch time is his shooting efficiency in the most critical moments of close games. Moments like this one:

Since Blackport’s clutch stats are based on his own win probability model, I wanted to see how they compared to Mike Beuoy’s clutch stats on inpredictable.com. Beuoy also uses a win probability model to categorize each moment of a game into either Normal Time, Garbage Time, Clutch Time, or Clutch Squared Time. These are analogous to the Low, Medium, High, and Very High Leverage situations on pbpstats.com.

As expected, Lillard’s clutch numbers on inpredictable.com are just as impressive. His eFG% in Normal Clutch Time is 57 percent, but it’s his shooting efficiency in Clutch Squared moments that stands out. Similar to the Very High Leverage attempts on pbpstats.com, Clutch Squared shots are typically the ones that ice a game or outright win them. To be specific, these are any two point attempt where the result swings the win probability for a team by at least 18 percent or three point attempt where the result swings the win probability for a team by at least 32 percent.

Lillard’s eFG% in Clutch Squared moments is a ridiculous 91 percent. While Lillard’s only taken 11 Clutch Squared shots (you can view all 11 here), to find someone with a higher eFG% in those situations this season, you’d have to set a field goal threshold of three attempts to include Tobias Harris (100 percent eFG%).

So if you leave this post with anything it should be that while Lillard has been great when the game starts to tighten up, he’s been even better when it matters most.

---

I was curious to see if any player in recent NBA history has had a clutch run like the one Lillard is on so I dug into the numbers. To ensure I’m making apples-to-apples comparison (or close enough) I limited my analysis to clutch play (by the NBA’s definition) before the All-Star game since the 1996-97 season, which is as far back as the NBA’s clutch data goes. The chart below shows every players eFG% and field goal attempts in the last five minutes of games when the score is within five points before the All-Star break.

Clearly, Lillard is in a league of his own. No one before him has combined the same level of volume and shooting efficiency in clutch time before the All-Star break. However, this chart also suggests that we should expect some regression for Lillard as the season continues.

But then again, if you had shown me this same chart last month, I would have said the same thing and I’d have been wrong. So if there were ever a player who could break our understanding of what’s possible in the clutch for an entire season it’s Lillard.

Last week, Mike Beuoy of inpredictable.com sent me the data on every player’s career numbers on Clutch Squared attempts. As a reminder, these are the shots that have the highest impact on win probability.

The chart below shows the player with the highest career eFG% since 1996-97 at every attempt number.³

Despite being one of the younger players on the chart, Lillard stands out for his combination of volume (189 attempts) and efficiency (49 percent eFG%). The first player to left of Lillard on the chart to have a higher eFG% is Chris Bosh, who attempted just 134 Clutch Squared shots in his career with an overall eFG% of 52 percent.

Clutch stats are based on small sample sizes which means things can turn at the drop of a hat. So I should emphasize that the most likely outcome is that Lillard starts shooting at around league average in the clutch for the rest of the season, which would bring his astronomical numbers down to more mortal levels. But given what we know about Lillard’s ability to rise to the occasion when the pressure is on, it wouldn’t surprise me if he sets a new standard for clutch play by the end of the season.

1. I’m loathe to do it, but here’s a table with everyone who has taken at least 30 attempts in the last five minutes of a game when the score is within five points.

2. What a typical version of this chart would look like

3. What a typical version of this chart would look like