Nylon Calculus Week 5 in Review: Boston’s big week and offensive fouls

The NBA has been stretching the season to reduce the number of back-to-back games and awful travel trips, so there have been fewer games per day, but it doesn’t seem like it. It’s been a busy season with many storylines, and there’s too much to unpack every week.

There’s no way I can reasonably cover everything I’d like to with the kind of depth the topics deserve. But I think that’s how every die-hard fans feels — there’s too much to consume right now. That’s not the worst problem to grapple with, but it’s the one we have to manage.

The Boston Celtics’ win streak and chaos theory

Most people are aware of the chaos theory from Jurassic Park or some of those fractal drawings adorned with a kaleidoscope of colors, but it’s a serious mathematical topic with real-world consequences. Dr. Ian Malcolm in the video below is right though, even if his water example is a little creepy (you know he tried that on undergrads while he taught.) Complex systems can be very hard to predict, and basketball is not an exception. What’s important to know about chaos theory is that its very definition is that it’s mathematics focused on systems that are highly sensitive to initial conditions. That’s what you should remember — and it can be broadly applied to the butterfly effect, water droplets on your hand, the financial system, and the NBA.

For example, let’s look at the Boston Celtics. If you tweak the initial conditions a bit, you could conceivably see vastly different results. Let’s say Jayson Tatum, while practicing over the summer on getting used to the NBA 3-point line, listens to a friend about how to tweak his shot. This throws off Tatum at the start of the season, and given the heavy demands of the league he has little time to fix the issue. He loses confidence and plays worse overall — nothing strange for a rookie. Then let’s say Aron Baynes, an important defender for them, lands on someone’s ankle and is out for a few games. That’s also common.  Finally, in their third game, imagine because Baynes is out and Tatum isn’t playing well, it’s a close game and the Celtics miss a shot right at the buzzer. I think that’s plausible, and no one would state that any shot like that is inevitable. Everything else is the same for the games previously.

What would happen? That would mean the Celtics would be opening the season with three straight losses, and they’d be dealing with Gordon Hayward’s injury and a void in the frontcourt with Aron Baynes. Marcus Morris would still be injured too. Also, with Tatum struggling, Brad Stevens would be scrambling to fill the void on the perimeter, and they’d surely have to press into the services of lesser players. Their offense even with Tatum shooting well has been mediocre, so you can imagine the issues they’d face. They also didn’t have a ton of blowouts, so with a softer defense overall a few of those other games, like the the one point win over Toronto or the recent one against the Warriors. There’s a snowball effect with these games too, as players can ebb and flow due to confidence in their abilities.

It’s entirely possible that with a few small tweaks the Celtics could be standing near 0.500 in the East. Due to their dogged defense and Brad Stevens’ coaching I wouldn’t count on them ever being truly bad in a plausible scenario, but defense is about effort and if players feel hopeless or lose their confidence, things can get dire. If you could “rerun” the season several times with slight variations, I think reasonable people would agree you’d see a large range in the outcome for the Celtics — and yes, most of those would include a healthy Gordon Hayward. But this is still probably one of their rosiest scenarios. The problem is that we only have access to one “simulation” of the league and it’s impossible to truly know what their true baseline is — but let’s enjoy this ride while we can.

All is fair in VORP and WAR

The excellent baseball writer Joe Posnanski penned a response recently to Bill James, who was disappointed with baseball’s individual player metrics (“War”.) Specifically, James was wondering why no final link was made between runs created and wins — because wins in baseball, and sports, are all that matters. You can extend this to the NBA, of course, and all our metric soups out there. But we gotta be careful about linking them directly to wins rather than expected wins. Yes, there are certain effects that happen with point differential translating to wins, and it can explain a little of why, say, the Grizzlies do well in outperforming their expected wins, but the adjustments are usually small. What we need is a thoroughly researched bridge between wins and point differential that’s beyond the basic Pythagoreas formula. It’ll take some work, but that final step is worth it, if not just for completeness. (And I shouldn’t breach the clutch subject, since that one is incredibly noisy and has little to no consistency year to year.)

Return of the CP3

Chris Paul hasn’t been back for very long, and his minutes have been low, but I think we can still note a few key observations. First, very few players this side of LeBron “cyborg” James have played more minutes than James Harden in recent seasons, and Paul’s inclusion will ideally let Harden rest a little bit more. It’s too hard to tell if that’s the case now since it’s only been two games with Paul playing a little more than 20 minutes each — plus they played some weak opponents. But it’s fair to say that Harden won’t be impacted too greatly, at least from these initial observations. He had a monstrous 48-point, 7-assist game against Phoenix. Harden played well in the other game too. If anything, given Paul’s history, we shouldn’t have to worry about Harden’s numbers; it’s Paul who will take the smaller role.

I could only find one Chris Paul assist where the recipient was James Harden, but it’s illustrative of a dangerous play type they have now. Paul feigned a pick-and-roll at the top with Clint Capela, and the defense predictably reacted with heavy interest. But this left Harden’s man, a little turned around after a quick switch before the pick, too far in the middle of the court, clearly concerned about Capela and Paul. Harden fired up an off-balance shot, which he made anyway, but now he has the ability to get cleaner looks by playing next to this offensive star.

The effect cuts the other way too. You can see here during a Harden pick-and-roll where Paul is left completely open in the corner — he’s not known as a 3-point shooter, but those types of shots are easy money for him. They didn’t have too many inventive plays yet involving both guys directly, but I did spot this out-of-bounds play. Paul throws the ball in to Capela, and then darts directly to the rim thanks to a screen by Harden, who’s right next to Capela too. It’s not the trickiest play, but it works so well because the defense is so concerned about Harden. We’ll see what they do in the future with both guys. The potential is there because both guys are exquisite passers and great shooters too, and surrounded by all of Houston’s shooting and some great pick-and-roll big men, they have the best possible playground for their skills.

Evaluating a trade five years later

Whenever a trade happens, I understand the inclination to judge it right away and speculate on its impact. But that does not seem accurate or the preferred method, and it makes me wonder if we could build better trades. We could, for example, evaluate trades five years later, which should be a long enough time span to judge most of the major effects without too much time passing. And what trade happened about five years ago? Well, the infamous James Harden trade was in 2012. I know this trade has been discussed as much as any trade, but you can see how much easier it is to trace the ramifications. Oklahoma City received Kevin Martin and Jeremy Lamb; Martin wasn’t there long but was a good scorer, while the latter never worked out. They also received a couple draft picks that turned into Mitch McGary, who was decent when he got minutes, and starter Steven Adams, an important piece of their defense.

Obviously, Houston received James Harden (and some minor players) in that deal. They clearly won, but in retrospect it looks even worse for OKC. Oklahoma City, and many other organizations, misjudged where the cap was heading, and they put all their eggs in the Serge Ibaka basket only to see him wilt and leave for other NBA cities. But you can do this with other trades, some of which are more interesting with hindsight, like the trade that sent Emeka Okafor and Trevor Ariza to Washington for Rashard Lewis. Okafor only lasted one season before injuries ended his career, and Ariza played well for the Wizards but left for Houston, of all places, after two years. At the other end, Lewis was a pure salary filler who was waived right away. The Hornets used that cap space to … re-sign Eric Gordon, who was a disaster for them. The best laid plans of mice and men are subject to the harshness of reality, and apparently the gravitational pull of the Houston Rockets.

Warriors vs. Celtics

In an instant classic, the two teams with the best record in each conference finally squared off, and the result was exhilarating. The Warriors jumped out to an early lead, and people were already writing the Celtics off, and then the lead evaporated near halftime. However, the Warriors again blitzed their opponent in the third quarter. The difference is the Celtics weathered the storm and actually came back; typically, the Warriors destroy people there. Part of this had to do with foul trouble — both Klay Thompson and Stephen Curry ended the game with five fouls each. But it was impressive nevertheless, and the win streak survived.

I know I’ll receive flak for this from all areas in and around Boston, but their run reminds me of the Atlanta Hawks back in 2015. Stay with me — that was still a good team, and it was fun watching it happen. They won 60 games, and their 19-game win streak was joyous. But it wasn’t sustainable, and they were operating a little over their heads. The Celtics, relying on a bunch of young guys, Marcus Smart, and the duo of Al Horford and Kyrie Irving, are about as good on offense as I expected without Gordon Hayward. But their defense is operating at the peaks you’d see from legendary defenses like the Tim Duncan-era Spurs from over a decade ago or the Kevin Garnett Celtics. No offense to Aron Baynes and Al Horford, but I don’t see the same type of foundation here — they aren’t that good.

Elo

I know there are way too many stats out there to consume, but it’s good to get a little passing overview of some of the uncommon ones for whenever you do encounter them. For instance, on 538’s website, they use Elo ratings for team stats. It’s not something you ever really see with basketball anywhere else. It’s basically about establishing some baseline, then changing it — adding or subtracting based on whether or not you win — based on recency and strength of the opponent.

That recency part was something I was always curious about because it’s usually best to use full season results and the problem that human observers had was recency bias — putting too much emphasis on recent games. Phil Birnbaum found the same issue with their Elo itself. One issue is that the recency adjustment is supposed to deal with changes in the team itself, like an injury, but Elo is so slow to respond that it’ll likely be in inaccurate for most of the time span in question. There are other concerns too, like strength of schedule early in the season, but it’s something to consider when you see their Elo cited.

Embiid’s pantheon game

I’m sure my readership is already aware of Joel Embiid’s monster game against the Lakers, but I have to at least discuss it because it was a milestone. We all like to make narratives out of the randomness that happens in the NBA, and this is the point in the script where Embiid has crossed the threshold and become the legend (we all hope he is.) If you need to pick a specific date for when he truly became a star, this is a decent one. And if you think a game against the Lakers shouldn’t count, then witness the game he had against All-Star 7-footer DeAndre Jordan.

Also, pet peeve here, many people trumpeted the fact that no one else has ever had Embiid’s stat-line of 46 points, 15 rebounds, 7 assists, and 7 blocks. (Well, we’re ignoring seasons before 1974 because blocks didn’t exist — sorry Wilt Chamberlain — but it’s still a large span of time.) But the more specific you go, obviously the more guys you can weed out. I need to find a name for this problem — the arbitrary minimum? — but basically, if you only look at games that hit at least those criteria or exceed them, then you’re biasing the sample set to games that will automatically make the game in question look better. The game defines the minimum, not the mean — we’d ignore a game with only 46 points, 15 rebounds, 7 assists, and “just” 6 blocks even though it’s remarkably similar. And it’d be grouped with superior games only, if any.

Relaxing the filters a bit, there are still only a few similar games in NBA history, featuring two monstrous legends in David Robinson and Shaquille O’Neal and multi-facted big men in Bob Lanier and Alvan Adams. Shaq’s game — 40 and 17 with 8 assists — was pretty mighty, but he did miss a ton of free throws. Alvan actually had arguably the most impressive statline with 47 points, 18 rebounds, and 12 assists. But no one else was raining 3-pointers, and heck, ignore the arbitrary minimum stuff — just enjoy his dominance in the video below.

Trends in offensive fouls

One of the most hated NBA plays is the dreaded charge. Sure, it’s a hustle play, and coaches probably love it, but most fans hate the break in action and fear that a player will get bumped in the air and go down hard. suffering a nasty injury. It’s not a play the NBA would want to see increase, yet from early returns that’s exactly what’s happening, as you can see from the team averages in 2018 compared to 2017. What’s going on? Is this an aberration? And have we seen anything like it before?

Although charges have been tracked publicly since hustle stats were debuted in the 2016 playoffs, you can actually track them back several years using the play-by-play logs. Unfortunately, charges only go back to 2011, but offensive fouls go back to 2006 and, weirdly, 2000 is available too. (Just don’t think charges are a recent invention; they’re almost as old as the sport itself.) I’ve compiled this data before — the scraping is a little different now, but the end result is the same — so this is no major task. However, I should always offer a disclaimer with play-by-play stats: the data is rough, it’s not official, and there’s often missing data, sometimes entire games. The further you go back, the more issues you see, generally. But the 2006 to 2017 range is pretty good, thankfully.

You can see a graph below for the results. It’s all in rate form so season to season comparisons are possible — you don’t want to use totals when there are fewer teams in a season or there’s a lockout. Basically, this is the highest rate of charges we’ve ever seen, but that can only be tracked to 2011. We’ve actually had higher rates of offensive fouls before. Based on what we saw in 2000 and in 2006 and 2007, it’s likely there were high rates through the entire first part of that decade too.

Note: per game means per NBA game, not team game. In other words, if two teams meet, then it counts as one game.

Thankfully, the ratio of charges to offensive fouls is fairly consistent although a little noisy. It’s just a ratio of about one-third. Thus. when the offensive fouls are higher, the charges should be too. Since the rate of offensive fouls now is actually lower than it was in 2006, 2007, or 2000, I’m comfortable in saying we’re not seeing a historically high rate of charges now — it’s been worse. In fact, when you adjust for pace — there are more significantly more possessions now — the higher rate of fouls a decade ago increases. The pace adjustment is modest enough that it doesn’t necessitate another graph. Besides, it’s also weird that offensive fouls are spiking too.

What caused this? This is the confusing part. There’s been no official release from the NBA about emphasizing anything related to a charge. Also, foul rates have actually decreased overall since the early 2000’s. That’s basically why charges were down recently until this season. And nothing’s changed in the 2018 season: the free throw rate is roughly the same as it was in 2017. The offensive foul rate used to roughly follow the overall free throw rate, but this season is indeed the exception.

Next: Nylon Calculus -- James Harden's scoring is unreal

One can craft other theories, like how early season foul rates are so much higher, but I think most of those can be discredited. For example, the charge rate in the first month of last season wasn’t appreciably different from the rest. This initial exploration mapped the waters, but it’s still an unsolved mystery.

At least we can resign ourselves to the fast that these aren’t the worst rates ever seen, if that helps anyone out there.