NBA Week in Review 21: March Sanity

The basketball world is being distracted by March Madness right now as the NBA approaches the end of the regular season. Both camps are fighting about who’s superior as well as over the nature of amateur athletics, but I only care about the NCAA for one reason: it’s a feeder league. This is a long, convoluted means to the NBA draft with Cinderella storylines and a few kids picking up degrees before they try their hand at basketball. You can watch March Madness for the next generation of NBA players. It’s not the ideal proving grounds for the league because the competition is so different, but it’s a prelude to what matters to me and may other NBA diehards: these guys will be important millionaires soon.

Warriors vs. Spurs: Redux

In past articles, I discussed how the Warriors were lucky — their opponent three-point percentages were abnormally low and a few of their guys, even ignoring Stephen Curry, were shooting significantly better than their career norms from behind the arc. Most of that is noise, and it does not hold up over following time segments — meaning, that advantage evaporates the minute you observe it. Thus, in a statistical sense, the Warriors/Spurs rematch was a simple regression to the mean, and the team with the larger season point differential — and the team that wasn’t missing a key cog in Andre Iguodala — won.

Of course, that’s a robotic way to view the league, and there’s a lot of narrative and strategy lurking about in how the Spurs thoroughly defeated the Dubs. Humans are natural story-telling machines who craft messages from disparate information. This can be useful or detrimental, depending on the situation. People are often clouded by their own incessant pareidoliac need to make a narrative about anything, like stating that the Warriors are running out of gas at the end of the season or Curry is not a “true” superstar like Tim Duncan and his magical ride is about to turn back into a pumpkin. The former explanation would hold merit backed up with more information, and the latter is the kind of biased view you’d see from someone who simply does not want to accept that Steph is a great player.

This is a point I’ve returned to ad nauseam, but it’s important and it’s the future of basketball analysis — and analysis everywhere, really. It’s about picking out relevant, consistent patterns. It’s about ignoring the narrative ephemera that distorts the league or the statistical fluff that looks interesting on paper but means nothing. Stephen Curry will likely not shoot 1 for 11 from the three-point line again versus the Spurs, but their game-plan was successful and they might be able to control Golden State’s outside shooting as well as reasonably possible. That’s the kind of synthesis needed for the NBA: not fully relying on what you see but not just relying on the numbers either. The notion that you can “only” use one brand of analysis is a false dichotomy and it’s a distracting argument that leads nowhere.

Westbrook and Versatility

Russell Westbrook has been on a triple double binge lately, and he’s in striking distance of the post-merger single season record of 17 from Magic Johnson. That would actually put him 8th all-time, at least according to this article, which is impressive considering how much easier it was to achieve triple doubles in the past. But it’s not quite about the triple doubles but what they represent. Going into the numbers, what you can learn from one-metric stats is that interaction variables are powerful predictors of team success[1. You can read about interaction variables, like AST*TRB, from another source too.]. In other words, good players more likely have a ton of, say, assists and rebounds in concert together. It’s an interesting finding that has rarely been discussed outside of the APBRmetrics community. Westbrook, consequently, will look outstanding by any metric with interaction stats like that. But there’s a legitimate question about how much those stats are overrating him, if at all, because it’s derived through correlation and it may not apply well to everyone, especially with a guy at the extremes like Westbrook.

Her Majesty’s Bismack Biyombo

The Toronto Raptors grabbed Bismack Biyombo from the Hornets with little commotion, and he’s been a key component in their strong bench. The book on him is pretty simple: he’s the prototypical center with great athleticism and length, possessing shot-blocking and rebounding skills but little else, especially on offense with the ball in his hands. Yet the Raptors have still been better with him on the court during the season, and his long arms are a great basketball gift that makes up for his deficits elsewhere.

For instance, last week Biyombo grabbed 25 rebounds — there have only been five such games including that one during the season so far. When you collect that many boards, rebound stealing is likely, like on this play where he basically takes it from a teammate. He can get pushed around too, but his arms are so long he can still grab the ball over the stouter player. Those arms also allow him to tip the ball to himself above a scrum. He’s an excellent shot-blocker too because of that wingspan, and it allows him to block shots from further away than most players expect. You don’t want him passing or shooting, and he doesn’t want that either, but he’s a good back-up and a big defensive plus.

Point Guard Purity Control

Russell Westbrook still receives some criticism about not being a so-called “pure” point guard — whatever that means — but his image has improved a lot in that respect[2. And why don’t we talk about pure shooting guards or pure small forwards?]. People have begrudgingly accepted his greatness. Since he’s well over 10 assists per game, the point guard concerns are mostly extinguished. His aggressive style of play is a great complement to his passing because he draws so much attention and can create daylight in a split second. Here’s a play where he races down court at a blistering speed and shovels off a pass at the rim for an easy bucket. The defense is entirely focused on him and because he’s moving so fast no one’s even positioned well. Maybe he’s not the traditional definition of a point guard from 50 years ago, but that’s true of almost every position and role on the team. So why does it matter?

Lillard’s Defensive Strength

Damian Lillard has been roundly criticized for his defense, but it’s fair that we point out his assets on that end of the court as well. Going to the offensive fouls drawn leaderboard, he’s fourth in the league by totals after Ersan Ilyasova, Kemba Walker, and Ricky Rubio[3. Per my own research, charges are not more valuable than any other type of offensive foul drawn and, in fact, they were less valuable.]. As I’ve found before, these are uniquely valuable stats to collect, and it’s a great sign he collects so many. Usually smart crafty defenders like Rubio are league leaders, but I wonder if Lillard’s weakness, navigating screens, is paying dividends because he’s creating more offensive fouls from smashing into screeners. Here’s a play where he runs right into a screen and it’s a turnover for the offense. He was ranked pretty high last season as well so he’s not a fluke. He still has problems on defense getting lost and picking up on switches or screens, and he doesn’t do much on the other basic defensive stats like steals, but he’s been drawing offensive fouls and that’s valuable.

Anthony Davis: An Abrupt End

Once again, Anthony Davis’ season was marred by injuries, and this time it has ended a season — a season that was already afflicted by a cornucopia of injuries and set-backs. Many people have already discussed his scary injury history, as he’s played between 61 to 68 games per season. He suffers an inordinate amount of odd injuries, and some are significant. It’s not a good sign when a player can’t stay healthy in his late teen’s and early 20’s — it only gets worse as you age. He’s actually been suffering through shoulder pains for three years and will just now be getting a surgery since he’ll be out anyway with the knee problems. Knees are notoriously troublesome in the sports world, but shoulders are a problem too, as they’re often slow to heal and recur.

Unfortunately, the history of players Anthony’s age who were good enough to play right away is short, so it’s tough to find any comparisons. Here’s a list of players who have played between 60 and 70 games with a decent amount of minutes under the age of 24. Chris Bosh is probably the best prototype here because he was excellent quite early on as well and missed a few games nearly every year even as a young kid. He hasn’t aged terribly at all, but it’s difficult to learn anything about his injury history because the blood clot problem was a (scary) fluke. Luol Deng didn’t age well, but he was heavily used in Chicago and wore down under the strain and equally unlucky spinal tap scare. Mike Miller could be a harbinger of things to come, however, as he missed a few games in his early 20’s and the injuries just piled on as he got older.

Why does this matter? Numerous people have pointed out that if Davis doesn’t make an all-NBA team, he misses out on 24 million dollars due to the Rose rule. He may still get voted onto a team due to the weak center position, but his injury problems are already affecting his career. I am in no way endorsing myself as a medical expert, but he was a late-bloomer physically and perhaps that’s leading to a few more odd injuries than normal. It’s tough on any body, of any size, to play in an NBA, and few people can move like he does. Let’s hope his body can strengthen and he can flourish into his mid-30’s. There’s too much talent to waste.

To Find a Thief: Breaking the Single Game Steals Record

Last week, I calculated the odds for Hassan Whiteside and a couple others breaking the single game blocks record, which I thought was a tough task but possible. Concurrently, I assumed steals would be possible as well because the record is a mere 11, held by Kendall Gill and Larry Kenon. In fact, Brandon Roy got 10 back in 2009 — so 12 shouldn’t be unreachable. But the odds are greater than I thought, and it’s a tougher record than one would think.

The methodology is the same as it was for blocked shots. I’m using a gamma statistical distribution, which works decently but it’s not perfect. There’s an issue with minutes skewing the distribution — that has a large influence on the results and it’s tough to skirt around that issue. But the players who are likely to break this record are starters with consistently high minutes, so much of that problem is mitigated. The key here is selecting a data set with a stable environment where the variation is more “natural” and not of circumstance, like a player getting injured and seeing his steal totals fall or a large reduction in minutes.

After going through the numbers of high steal games, one observation stood out: if you want to break the steals record, you’ll want to play for the 76ers. The team had a disproportionately high number of games where a player had six or more. For instance, Michael Carter-Williams had four in his one-and-a-half seasons with Philadelphia and none since then, including an outburst of nine early in his rookie year. Thaddeus Young has six such games in his career and five of them were with Philly (and in the same season.) Even their center, Nerlens Noel, recently had a game with seven steals over the past week. But their king is Lord Robert Covington, who leads everyone in total games with at least six steals over the last two seasons — he has six and the next highest total is three.

Looking at Covington’s last two seasons — his rookie season had scattered minutes — you can see his distribution of steals in the graph below. It’s a pretty steep dropoff, but he has the flattest curve of virtually everyone. It’s tough to see, but he has a few games past five steals — they’re rare, but at least they show up. Using those numbers, however, he only has a remote chance at breaking the record: 0.0038% per game. Over 126 games (his total in the data set I used) the odds of hitting 12 steals or more would still only be 0.47%.

However, steals seem more difficult than blocks in terms of fitting to a distribution. Perhaps it’s because the highest averages are so much lower or some other reason I have yet to conjure. Also, you may have noticed something about Robert Covington’s graph: it’s bimodal, meaning he has two peaks. I tried filtering out lower minute games, but it was still there. I don’t expect the pattern to continue, but it’s something to track. Maybe he plays in a way where a bimodal distribution just happens.

No other players have a better realistic chance than Covington, unfortunately, due to his age, high steal rate, and team. Steals have been declining for a few seasons now overall. But the record can’t be overtaken by any one player. It’s all about the overwhelming odds of dozens of full-time starters who have hundreds of opportunities over the years to break the record. With enough high-steal players out there, it’s possible, but it may not happen for a long time.

And if it does happen, I’d bet a 76er is involved.