The Diff: Revisiting ShotScore and how the Cavs are terrible

The Diff is your weekly WFNY look into the amazing world of sports statistics. For a complete log of articles, click this link. Last week, I wrote about the downfall for the Cavs in four terrible home losses. This week, I’m focusing on shooting data again.

Within the next 15 days, this Cleveland Cavaliers roster should undergo yet another significant shift. The past few weeks of speculation have made it perfectly clear that not everyone – front office and/or personnel-wise – will survive this season from hell. While there are storylines for days about what that means for the organization’s future, I wanted to focus – perhaps for one of the final actually meaningful times – on what we’re seeing out on the court from these Cavaliers of late. Today, I’ll be looking at the Cavs’ efficiency from a number of different areas, again highlighting the ShotScore statistic. 

Brief overview of the Cavs season

The Cavs are a bad team. They started the year on a 4-12 run. They’ve now lost seven of their last eight. On the season, they’ve only won one-third of their contests; they are 16-32 entering tonight’s game against Los Angeles. But despite the assumed narrative of a bad team just always being bad throughout, they’ve actually been a very different type of terrible recently. Take a look at this chart:

The split above shows the team’s Offensive, Defensive and Net Ratings before and since Luol Deng’s debut on Friday, Jan. 10 in Utah. As one can tell quickly, the Cavs indeed have been one of the NBA’s four worst teams throughout.

In the 35 games prior to Deng’s arrival, the Cavs were a futile offensive team. This included the failed Andrew Bynum experiment and Kyrie Irving’s early-season struggles. However, they were very solid defensively at No. 13. That was a huge improvement from the three years under Byron Scott.

Since Deng’s debut, the offense now has notably improved: 5.3 more points per 100 possessions is a big deal, even if it still ranks among the NBA’s eight worst offenses. But that very solid defense? Gone. Disappeared. Deceased. Avada kedavra’d, if you prefer Harry Potter-themed metaphors.

The Cavs have suddenly turned into the NBA’s worst defensive team. I’m not quite sure how that’s possible under a Mike Brown unit, but it has happened. They’re even allow 1.5 points more per 100 possessions over No. 29 Milwaukee, who also happens to be the worst overall team in the league.

Some quick plus-minus Cavs data

Before getting into the nitty gritty of shooting data today, I wanted to dive into any possible trends seen in the recent plus-minus data. I’ve shared this type of data several times in the past. Plus-minus data includes a number of caveats: Just because the team is successful when a certain player is on the court, doesn’t really mean that player is the reason why. Also, uneven overlaps can occur that skew one player’s data because of his time with or without another one.

Here is a look at the Cavs players’ on- and-off-court data in these last 13 games followed by just a few highlights:

— The Cavs are terrible without Anderson Varejao. Like, really, really terrible. It’s quite impressive for any rotation player to have a positive on-court Net Rating during a 13-game stretch where the team’s overall Net Rating (-10.1) is that bad. Heck, Andy ranks No. 8 in the NBA overall this season for on-off court efficiency differences. Is he really this good? Probably not. The numbers are crazy though.

— Perhaps related, the bench has been dreadful. What seems more accurate: The Cavs starting lineup is surprisingly mediocre or the Cavs bench is horrendously bad? It’s hard to exactly tell. Varejao’s on-court numbers are perfectly mediocre. Everyone on the bench has really, really bad on-court numbers – most notably, Matthew Dellavedova, Earl Clark, Henry Sims and Anthony Bennett.

— What’s with the bad Kyrie Irving-Tristan Thompson defense? I looked into this a bit more last night, given the fact that Varejao’s on-off court numbers are excellent. Obviously, Irving-Thompson play a decent chunk of their minutes with Varejao in the starting lineup. Here are the gory results: Irving-Thompson have a -33.3 per 48 minutes sans Varejao in these last 13 games. Their defense is most troubling.

Reintroduction to the ShotScore statistic

Now, the fun stuff. Back in mid-October, Grantland’s Kirk Goldsberry made another splash in the NBA analytics world. His article introduced the world to ShotScore, a proprietary statistic emblematic of his signature spatial analysis. Go read that initial article. And also go read his mid-season follow-up on the best and the worst ShotScore players in the NBA.

ShotScore analyzes how players perform in shooting shots from exact areas on the basketball court in comparison to the NBA average from those areas. Goldsberry’s version looks at the exact space on the court with the help of a team of researchers. In order to easily estimate the statistic en masse, I just have exported the shooting zones data from www.nba.com/stats. That’s what I did in my Cavs-based ShotScore article in late October. It’s what I’ll do again today.

As a point of review, here’s how the NBA has performed in every shooting zone over the past several seasons. Additionally, within the 2013-14 data, I’ve broken out guards and other players per the same data source. This will be useful later.

Some league-wide trends: non-restricted area two-point shooting efficiency is down and the ratio of above-the-break three-pointers attempted is up. Nothing more to add there, but just interesting to point out that context.

Now, moving on to the Cavs: They are not a good shooting team. Their .465 efficiency field goal percentage is 29th in the NBA. Bleacher Report’s Jared Dubin had a very detailed article yesterday on the many issues plaguing the team’s offense. Specifically, their restricted area shooting efficiency was on a historically bad pace until Deng’s arrival.

But as mentioned above, the defense has actually been more worrisome of late. It’s been the NBA’s worst over the past three and a half weeks. Opponents shine in the restricted area against the Cavs. And, as I shared in Kirk’s Film Room last week, opponents are shooting drastically more threes in the absence of Andrew Bynum. That’s not a good thing because, as the chart above shows, threes have a much higher expected point value on average.

Thus, here are the detailed shooting zone stats for the Cavs in the already identified pre- and post-Deng arrival splits.

Yikes, the defensive struggles recently. There’s really only one positive thing to say: opponent corner three-point shooting efficiency has gone down. That’s it. Teams are shooting way, way, way better everywhere else. And they’re shooting 3.5% more above-the-break three-pointers, which obviously hurts. Perhaps these numbers are bound to regress to the mean. Hopefully.

On offense, you can note the restricted area improvement. The team has been performing worse from mid-range and corner threes since Deng’s arrival. Technically, the mid-range issues aren’t that surprising: As I’ve noted many times now, Deng is a below average shooter. He’s taking away long shots previously taken by guys like Jarrett Jack and Earl Clark, who actually aren’t half bad.

Diving into the Cavs player ShotScores

Now, I’ll finally be sharing the updated ShotScores for Cavs players this season. Per the suggestion of readers and Fear The Sword’s Sam Vecenie, I’ll be adjusting these for position. I shared the data earlier of shooting zone ratios and efficiencies for guards and other NBA players. That’s what I’ll be using here in this section, as also used in the current mold of All-Star guard-forward positions.

Previously, I simply calculated ShotScores by subtracting relative success or failure in comparison to the overall NBA average. Now, it’s in comparison to either guards or other players. Here’s the full table (this includes Luol Deng’s full season stats):

— Kyrie Irving’s season. He’s a 21-year-old that had a very poor start to the season, especially shooting-wise. Through 20 games, he was shooting 30.4% from three-point range and had a .432 efg. He has vastly improved in his last 25 games: 40.2% and .510, respectively. Those numbers are much closer to his usually elite career levels. But alas, for the season overall, he’s been practically average for NBA guards in ShotScore. That’s not helpful when he has taken that many shots.

— Tristan and Dion in the restricted area. Technically, Thompson is better at .514 compared to Waiters’ .459 mark. But, consider this: guards are usually worse in the paint anyway and Thompson takes a higher ratio of his shots in the restricted area. So that gives Thompson a worse ShotScore. He ranks third-worst and Waiters is sixth-worst in the NBA. Both marks are dreadful. There’s slight hope, however, in Thompson’s .558 shooting in the restricted area since Bynum’s suspension.

— Everyone else. Last year, the ShotScore chart had very good performers in Irving, Wayne Ellington, Shaun Livingston and C.J. Miles. This year? Only Miles has really separated himself from the pack of average. Jarrett Jack is actually second-best (more on this in a moment). The remaining guys, such as Andrew Bynum and Anthony Bennett, have been horrendous shooting-wise. Their sub-replacement level shooting stats have been a huge reason for the team’s overall offensive woes.

Adding even more context to ShotScores

To finish, I’ve got one final chart to share that somewhat brings everything together. When you consider ShotScores, shooting zone ratios are a large factor. For example, 60% shooting in the restricted area is average, but 60% shooting from mid-range is unheard of. Thus, based on a player’s shooting ratios, I can create an estimated efficiency field goal percentage based on NBA averages for guards or big men.

This final chart now shares that statistic, actual efficiency field goal percentages and player’s ShotScores per 100 field goal attempts.

Now, this is why Jack’s ShotScore is higher than Irving’s. Jack has an estimated efficiency field goal percentage of .444. Given his shooting ratios, that’s what an NBA guard would do. For NBA players with 300 field goal attempts, he has the third-lowest esimated efg. Irving’s estimated efg is also slightly below average at .478, compared to the overall average of .489 for an NBA guard. (The average of an NBA forward is .505.)

Which brings me to my final point: This actually isn’t that terrible of a shooting team. They’re not good in the restricted area, but they’re getting better and more respectable with the addition of the slashing-prone Deng. They are a fairly decent team elsewhere in the paint, in mid-range and from three-point range, as was shared previously. The biggest difference? Their shooting ratios.

As many have written before, the Cleveland Cavaliers take the worst shots in the NBA. Per estimated efficiency field goal percentage, only two Cleveland players take better shots than peers at their position. Those two are C.J. Miles and Tyler Zeller. That carries over into the team-wide data. Cleveland’s .487 offensive estimated efficiency field goal percentage ranks 30th.

The NBA’s average efficiency field goal percentage is .498. The Cavs are setting themselves back significantly without even considering their efficiency of shooting these shots. Offensively, this team needs to take more three-pointers. They need to attack the rim more frequently now that they’re regressing to the mean in the restricted area. They need to make it easier for themselves to score points.

On the flip side, the Cavs defense is perfectly average with an opponent estimated efficiency field goal percentage at exactly .498. The bigger issue? Stopping teams from converting those shots. They’re getting worse and worse in the restricted area. Teams are now starting to take way too many three-pointers. These are serious issues. And combined with the offensive woes, it’s why the Cavaliers a just a bad, bad team.