Nylon Calculus Week 22 in Review: George Hill, Joel Embiid’s free throws, and other questions

We’re so close: the NBA regular season is roughly a month away from completion. Most discussion now involves seeding and health for the playoffs — we’re zeroing in on the championship. It’s been a longer season than usual too, and I think a large chunk of the teams and players in the league are ready for the playoffs to begin.

If you’re one of those unlucky fans rooting for a lottery team, you’re probably wincing at all the awful play as teams directly and indirectly tank, and you’re waiting for that 2018 draft in June. I for one am ready for this season to end so we can get to the post-season games — it’s time to get to the serious games. And with that, let’s look back at the last week in basketball.

George Hill

I’ve been keeping tabs on George Hill for a while now. Hill, at his best, is ideal for a LeBron James title team, and for a team with so many issues this season he could be their elixir. Unfortunately, he hasn’t been their magical solution. The Cavaliers have had a rough month, and their prospects for the postseason aren’t looking great — they are not acting like an elite team. Plus, by numerous metrics, this is Hill’s worst season by a large margin, excluding only his rookie season. What’s going on? Will his numbers and impact prove?

The good news is that all his offensive numbers this season and in Cleveland are all well within the range for his career. His assist rate has barely budged, even though he’s playing with a ball dominant player now. He’s taking fewer shots at the rim, but he’s taking more from behind the arc. He’s turnovers are down too in Cleveland. But his usage rate remains pretty low, even for him, and, importantly, the regression to the mean bug finally hit his 3-point percentage: after leading or being among the leaders in 3-point percentage, he’s hit around 34 percent for his new team. That’s killed his efficiency, especially with him getting to the rim less often. Fortunately, that’s below his career average, and he should improve as the season progresses.

Defensively, there’s no glaring sign in the countable stats either, as I’ve discussed before. His teams haven’t performed well this year, but crediting team defense isn’t that simple and single year plus-minus should not be trusted. But video can be a lot more telling. He played well against Toronto, which was a huge win for the Cavaliers, defending both DeMar DeRozan and Kyle Lowry. Here’s an example where DeRozan, who rarely turns the ball over, drives and George Hill pokes the ball away from him and saves it. He drives again on Hill and gets stuffed near the rim. Against Kyle Lowry, you can see how well he guards him in the below clip, tracking him around a pick and contesting a tough step back jumper that somehow goes in after clanging off the front of the rim.

You can see him clean up a mistake from LeBron James here, who fell asleep “guarding” Fred VanVleet, but George Hill noticed that in time and closed out hard on Fred’s 3-pointer. The issue with some of his defensive numbers now, the broad metrics like RPM and BPM, is that is own team is pretty awful overall on defense, and he’s sharing some of that credit because of how those are calculated. I’d expect his results to improve, and they should be better off with him. But they have many other issues to solve.

Ben Simmons’ versatility

The Rookie of the Year race is a good one this season, and while Donovan Mitchell has indeed been fantastic, we can’t underrate what Ben Simmons has done this season. The concerns about his offense a bit overblown, as we already have one giant pseudo-point guard without range success story in Giannis Antetokounmpo, and he’s near league average in efficiency as a rookie guard — that’s rare. His passing is definitely special, but his defense is what separates him from the pack. He has the size of a power forward and the quickness of a guard. His positional versatility is arguably the best in the league already. When people talk about the new position-less NBA, it’s hyperbole — except in the case of Ben Simmons. It’s made him one of the most valuable young players in the league, and it may lead to rookie of the year.

Defense is … ?

Portland has been one of the bigger surprises this season, hanging onto the third seed with one of the better defenses, despite a lack of elite big name defenders. The backcourt of Damian Lillard and CJ McCollum was supposed to lead to a poor defense and a team that would light it up on the scoreboards. But digging deeper, you can see how outside of those two guys they have very little offensive talent, and they have a number of defensive role players like Al-Farouq Aminu. They actually do need both high-scoring guards; otherwise there is very little in the way of creation.

At the pivot, you also have Jusuf Nurkic, and he’s a perplexing figure now. When Denver handed the keys over to Nikola Jokic instead Nurkic, the Nuggets took off and played great basketball. But Nurkic fit well with the Blazers. He sets big screens, teams can’t go small without him trying to take advantage of that inside, and he’s played great defense. But as many have noted, he’s been a ball hog, and his efficiency has been well below average. That directly leads to critics pointing to him when the team struggles. But we might be underrating his defense.

Going back a few years, Kevin Garnett was a mainstay in defensive plus-minus metrics, and there were all sorts of theories about why this was happening. One was his defensive communication or “quarterbacking” the defense. When you have a great defender who can see the entire court, you can cover holes in the defense and pick up plays just as they start. When done well, an NBA defense is miles better. I’m not saying Nurkic is at Garnett’s level or even close. But communication and execution seems to be one of Portland’s strengths right now, and he does deserve some of the credit. Let’s not write him off entirely.

Victor Oladipo: Jekyll and Hyde in Indiana

As I suspected, Oladipo’s numbers have dipped after the All-Star break. I’d call this the Jrue Holiday phenomenon. Holiday made the All-Star team a few years ago as a youngster — I’d bet many people have forgotten that fact, even die-hards — thanks to some great play before the All-Star break. But it wasn’t sustainable, as I like to say, when his shooting percentages absolutely cratered. That’s happening with Oladipo now.

After hitting 38 percent of his 3-pointers before the break, he’s only hit 30 percent after. Overall, his true shooting percentage has been below 50, which is terrible for a first option. I wouldn’t say his improvement should completely be dismissed, but I was worried this would happen. We’ll see how he finishes the season, and what he does next year. The first half of 2018 could be an anomalous stretch for him. Thankfully, his playmaking and defense haven’t collapsed too. And it’s best to look at season totals for future prediction, not just the last few months. Just don’t expect him to be this efficient again, especially with the load he’s taking.

NBA predictions based on actual player minutes

This was pretty cool to see: you don’t often see NBA predictions posted in-season adjusted for playing time and injuries. ESPN actually posted an article with this for playoff seeding predictions. The key takeaways: Toronto and Houston are nearly locked into the first seeds, while Boston and Golden State, due to some recent losses and injuries, are locked into the second seeds. Also, things are looking up for the Nuggets and Jazz due to how well they rate with their full lineups healthy (Utah is a monster with Rudy Gobert and Denver has Paul Millsap back.) Hopefully we get to see more intelligent predictions like that because team metrics — win percentage or point differential — are blunt and miss a lot of information.

Most Valuable Player voting

Annual scheduled rant: MVPs shouldn’t have to come from elite teams. Let’s think about this clearly. If an MVP has to come from a team with at least 55 wins, roughly, then what are we implying about the value of that player? Are they worth like 40 to 50 wins each? (A team of replacement level/borderline G-league players will win around 10 to 15 in a season.) That makes no sense. You can absolutely have a player who’s the most valuable in the league on a team with a 0.500 record; you just need awful teammates. And sometimes that happens.

Put it another way: James Harden would not be able to lift the roster of the, say, Kings to 50 plus wins. There’s also no excuse about separating the diamond from the rough either in those weird situations. We have video from almost every play this season on stats.NBA.com and YouTube, among other sites, as well as a plethora of national games with other viewing options and a nearly bottomless well of statistics with a variety of flavors that’d make Jelly Belly blush. All this is to say, yes, it’s okay to consider Giannis Antetokounmpo for the Most Valuable Player award, even if it’s just to put him second or third. No, his team isn’t great, but that’s not his fault. He’s already doing almost everything humanly possible on the court. He shouldn’t need better teammates to be considered for an individual award.

This is an interesting observation. Free-throw percentage is one of the most stable stats out there, and the only change you typically see is a player getting better as they get older, which rarely happens in a single season. Instead Joel Embiid has gotten worse. Since the All-Star break, for instance, his free-throw percentage has dropped by 10 points, and it was nearly 15 points before his 11-for-11 performance on Friday. You can see the trend in the plot below. He peaked around mid-December, and then took a huge dip around the beginning of March.

There’s another way to look at free-throw percentage variation. It’s rough, and it’s more of a preliminary exploration, but it gets the general idea across. If you plot every player’s free throw percentage for every season by the number of attempts, you can see the general trend in variance. Basically, the variation doesn’t greatly reduce until 100 attempts, and it’s still regressing to the mean past 200. While he has taken nearly 450 attempts this season, you don’t see the large swings in his free-throw percentage until you cut things into smaller samples. But I do see the concern there.

For a final check, I repeated the weighted averages I did for Embiid on his fellow central African, high-scoring center, Hakeem Olajuwon. He had typical swings in free-throw percentage too when you look at a rolling weighted average. In fact, during his sophomore season he had a high of 75 percent when looking at a 10-game rolling average, and a low of 50 percent. That’s similar to what we’re seeing with Embiid, so I wouldn’t be too concerned. His career free-throw percentage is the most important number; focusing on shorter chunks of the season can be misleading.

Here’s what I’d say: don’t ever look at (any) numbers in a vacuum. The context is key. No metric will be perfect, so if you want to use it, learn its strengths and weaknesses and roughly how it works. Kawhi Leonard’s defensive plus-minus was a big topic last year, which we discussed here. After some investigation, it was found opponents were making an unsustainable amount of 3-pointers when he was on the court. There are ways to adjust for this in a plus-minus model, but you won’t be able to adjust for every weakness so for a best practice consider context, context, context.

This is a fun question because I’ve always considered it. Basically, if you have a good offense you’ll have fewer turnovers and missed shots, and that makes life a lot easier for your defense — and vice versa. Let’s start with turnovers because those make up the bulk of the difference. Based on what I’ve found, generally teams are about 14 percent more efficient after steals. Cleveland is one of the worst teams in terms of picking up steals, so they’re about 0.8 steals per 100 possessions worse than average. That translates to about a 0.13 points per 100 possessions loss — not much, but let’s move onto field goals.

Next: Nylon Calculus Week 21 in Review -- Lillard, JaVale McGee, and reader questions

If they had an average field-goal percentage defense, they’d only pick up an 0.8 defensive rebounds per 100 possessions, roughly speaking. The difference between a possession following a rebound versus a made shot is only about 4 percent, so  this isn’t a big change. I’d say just 0.04 points per 100 possessions –yes, that’s tiny. In sum, maybe with boosting their defensive rebound percentage to league average and allowing for some adjustments, their offense is only losing about 0.20 points per 100 possessions.

It’s the same story for their defense. Even with their elite offense, the disparity between how many field goals they make versus the average team is small enough that you’d only see a marginal loss in their defensive rating. Plus, they actually give up near the league average amount of live-ball turnovers, so that’s not an issue. And they’re a below average offensive rebounding team too. On a small scale and in specific scenarios, factoring in efficiency gains after rebounds versus made shots, or made shots versus turnovers, is hugely important. But overall, it doesn’t make a significant difference for Cleveland.