A statistical deep-dive into Jrue’s playoff shooting

Jrue Holiday’s dropoff has been the subject of a lot of discussion and worry, with a lot of theories about why he does so much worse, but most of these theories are just speculation.

I was curious about why Jrue shot so poorly in the playoffs, so I tried to find evidence for many of these explanations with a bunch of statistical tests, but found that the data doesn’t line up with many popular ideas. In particular, the data seems to go against these theories that I tested:

​

**Reasons I Don’t Think are True**

1. Plain bad luck
2. Jrue is a playoff dropper in general
3. The extra exertion of the playoffs wears down his legs
4. Jrue always shoots worse against good defenses, which are more common in the playoffs
5. Jrue was asked to do too much in the playoffs on offense

​

I honestly have no idea why Jrue’s shooting falls off. I tried pretty hard to come up with a theory I could validate with data, but the one one I couldn’t refute is that something about the Bucks caused Jrue’s shooting to go haywire.

Before going to the Bucks, he shot 39% from three in the playoffs, and I found that in the last three years, bucks players shot 3% worse in the playoffs, compared to a 1% drop for the league as a whole. Some of this average is driven by Jrue himself, but Giannis, Jevon Carter, and Bobby Portis all have very large dropoffs as well, and even Khris Middleton has a 2% dropoff, above the league-wide average.

Maybe it’s Budenholzer’s fault, maybe Bucks practices are too intense in the playoffs, or maybe it’s just a random, meaningless artefact of the data. Still, this, combined with an element of bad luck, could be part of the explanation, although it’s not a very satisfying answer.

​

The rest of this post is just me going through the theories I don’t think are true and explaining why I think that, starting with the possibility it could just be bad luck.

​

# The “Bad Luck” Theory

It seems insane that a 10% drop in shooting percentages could be mostly due to luck, but the sample size is only 250 playoff shots. That’s not actually that much.

Because there are 450 players in the NBA, it seems reasonable that a few of these players could have outlier bad luck. We only notice the playoff dropoff for players with a big drop, so this might be an example of the [base rate fallacy](https://en.wikipedia.org/wiki/Base_rate_fallacy). For example, if there are 5 unlucky playoff droppers and 20 “real” playoff droppers, the true probability that an observed shooting dropoff is not just a random artefact is only 80%, but most people don’t think about the fact that a few droppers will be unlucky, so they just assume that any drop they see has a satisfying explanation.

Could we all just be falling victim to the base rate fallacy? We probably are to some extent, but I don’t think that’s the whole story. The size of Jrue’s dropoff is just too big.

We can grasp how big our 250 shot sample size is with a p-value. Basically, we are asking “if Jrue is not a playoff dropper, what is the probability his shooting will fall off 10% just due to bad luck?” For catch and shoot attempts this value is about 9%, but for open or wide open threes it’s 0.7%. Jrue’s free throw percentage and long midrange percentage also fall in the playoffs.

We don’t want to fall victim to the base rate fallacy, so I want to acknowledge that if every player had a 0.7% chance of being this unlucky, that still means three players in the NBA would drop this much in the playoffs on average just due to bad luck, even if they all had a sample of 250 shots.

There are 15 playoff teams. Let’s say they each have 8 rotation players in the playoffs, so there are 120 rotation players total. Based on the p-value for catch and shoot attempts, if all of them had 250 attempts, we should expect one of them to have a 10% drop in shooting due to bad luck every three-year stretch, just like Jrue.

Still, the preponderance of evidence makes me think that bad luck is a big part of the explanation, but not all of it.

#

# The “Lights too Bright” Theory:

Maybe Jrue is just bad in the playoffs in general. If this is the case, we should expect to see his non-shooting-based stats like turnovers and assists to fall in step with his shooting percentages. Instead his non-shooting stats mostly stay the same or improve.

​

Here, I plotted the change in a bunch of shooting stats (red) and non shooting stats (blue) from the regular season to the playoffs:

​

https://preview.redd.it/p7nhl7vpxgtb1.png?width=614&format=png&auto=webp&s=e1153359c61b2c428cb9dc8601b695893ca14878

# The “Tired Legs” Theory:

Al Horford is old and wore out in the playoffs last year. Maybe Jrue does the same thing. If this is true, we would expect his shooting percentage to decrease as the playoffs go on. Instead, I found this:

​

[The error bars are standard deviations assuming shooting follows a binomial distribution.](https://preview.redd.it/e9x73xbnygtb1.png?width=713&format=png&auto=webp&s=2b7604dd4e520d20b072ad968a20109ce6cbfeec)

Jrue actually shot 37% in the last two rounds of the playoffs on open or wide open shots. I also looked at whether Jrue shoots worse when he plays against a perimeter star that he has to defend. I did this by estimating the correlation between Jrue’s shooting and the highest O-LEBRON of the opponents perimeter players in the regular season. I found no detectable correlation. Jrue’s three point shooting was also better in the second half of games (on slightly lower volume) , which is the opposite of what you would expect if he was wearing down.

Finally, I hand-tracked some of his misses, finding that 48% missed short in the regular season and 46% missed short in the playoffs, essentially the same thing.

#

# The “Good Defenses Bother Him” Theory:

I tested this by looking at Jrue’s shooting against top and bottom defenses in the regular season. Against bottom 10 defenses, he shot 42%, but this dropped to 39% against top 10 defenses. This drop is small enough that is could be caused by random chance (29% p-value), so I wouldn’t read too much into it. As more evidence, here’s a scatter plot of Jrue’s shooting against each team based on their defensive rating, showing no correlation:

​

https://preview.redd.it/rtwxvv433htb1.png?width=609&format=png&auto=webp&s=c56ee9b059c1125bd3723048d872333a19e8c05d

​

# The “Role too Big” Theory:

Jrue plays a similar role in the regular season and playoffs, so this one doesn’t make that much sense to me. One way I tested this idea was to look at his shooting when both Giannis and Middleton were on the floor (i.e. Jrue wasn’t asked to be a 1st option), versus when neither were on the floor. His 3p% went from 44 to 43. While I don’t think this is the reason Jrue’s shooting drops off, he might still shoot better in a smaller role in Boston.

​

​

Putting it all together, I think all these theories might play a slight role, but I didn’t find any evidence that they were the primary reason. After going down this rabbit hole, I’m a little disappointed I couldn’t solve the mystery, but it made me a little more confident that Jrue’s shooting will rebound in Boston, since the issue doesn’t seem to be something like tired legs that’ll only get worse as he ages.

I would expect Jrue’s shooting to regress to the mean upwards in the playoffs, and I would expect more of his shots to be easier catch and shoot attempts. Therefore, this project made me feel a little better about how Jrue will shoot in the playoffs. Since so many of his other indicators in the playoffs are elite (he was the second or third best player on a title team after all), if he even shoots a little better he could be a major contributor to banner 18.

by roblox_master_2002