Nylon Calculus December in review: 3-pointers, Trae Young, and lineups

Welcome to the new year. I’m finally back to the NBA writing grind. Yes, I’m covering topics now that I originally started back in December, but as I try to do with most of my writing the message is still valid. The games can flow by but the core truths about the game are still intact, regardless of all the chatter about outside shooting and the extinction of positions or just big men in general.

With the new year, this column is going to focus on longer pieces and ditch weekly articles for monthly recaps hitting a few key topics. And with that, let’s take a look back at December.

The most valuable shot is not the 3-pointer

This is a thought that’d bounced around in my head for a while, and I thought I’d take a swing at addressing it because it is always relevant. With Gregg Popovich’s comments about 3-pointers (including calling it a circus shot) going viral last month and inspiring conversation in multiple places, there’s one common refrain I want to question: the 3-point shot is the “best” in basketball and any team who doesn’t embrace it will thus fall by the wayside. This is not just an exaggeration but it misses the crucial basketball analysis that led to 3-pointers becoming more valuable.

First of all, no, the 3-pointer is not the best shot in basketball. Strictly speaking, you can argue that the dunk is the best shot in basketball. There are usually play-by-play data issues with how play types are recorded, but even with a sizable reduction in field-goal percentage dunks are still top-of-the-line valuable. In fact, all shots at the rim are more valuable than shots behind the arc, especially once you factor in free throws.

Obviously, you can’t force shots at the rim; they’re not easy to generate. But that actually makes the 3-pointer more valuable because, with improved spacing, you can create more shots at the rim. It wasn’t just about a bunch of basketball nerds saying 3 was greater than 2; it was about the actual play on the court being affected by the positive synergy of outside shooting.

Generally speaking, getting a shot at the rim is still better. Those shots are converted at a high percentage, and even when they’re contested, the chances of a shooting foul are much higher. And I haven’t yet discussed offensive boards, which are more likely too when the shot is closer to the basket. Yes, the 3-pointer is still a great weapon, and yes, it’s usually better than a mid-range shot, all other things being equal and ignoring tactical decisions about how to strain some opposing defensive styles. But it’s not the best shot. You need to get to the rim for that.

The Suns and the rest of the west

The Western Conference is highly competitive again, as it has been for decades now, We don’t have teams on track for 70 wins, but we do have an incredibly deep playoff race. It’s strange right now to look the standings and see the Spurs, Jazz, and Rockets wrestling for seeds without homecourt advantage. We have 14 teams over a win percentage of .400, and it’s just poor lonesome Phoenix in the cellar at .150. Has anything like that spread happened before? It’s rare, but it has happened. The last time was in 2006 when the worst team hit 21 wins and the second-to-worst had 33 wins. There was also the Eastern Conference in 1988 when every team but one was at least two wins from the playoffs with the poor Nets at 19 total wins. We don’t have any crushing dominance in the West right now, but we could at least have more meaningful games down the stretch of the season.

Will the Cleveland Cavaliers have a FIRE! … sale?

We’ve got a number of teams competing for the distinction of the worst record in the league, and while the draft odds are no longer what they once were, “tanking” is also about accumulating other future assets by dumping what you currently have. Two of the bottom four teams are already in the midst of rebuilding where their best assets are from recent moves, like the trade for Zach LaVine or the Trevor Ariza signing, that clearly fit into some sort of franchise plan. Then there’s New York, who were devastated by the Kristaps Porzingis injury. Finally, you get to the Cavaliers, who still have a few remnants from their (very recent) contender days. Should we expect a flurry of moves soon?

I imagine that many people inside the Cleveland organization were hopeful the team could still compete for, say, the No. 8 seed. After all, they won the title in 2016, and they still have Kevin Love and the core that competed so well last year without Kyrie Irving. There’s also the psychological effect where few people there would want to say something negative about the team, creating a positive feedback loop that artificially warps the perception of their chances. But the reality should be sinking in now. Even with a healthy Kevin Love, they cannot compete with the league.

Luckily, the Cavaliers have multiple players whose contracts expire by the summer of 2020. Tristan Thompson is by no means an up-and-comer now, JR Smith is a veteran on a rebuilding team with a price tag that goes up to $15.6 million, Jordan Clarkson is a nice $12 million trade piece as well, and then there are the new additions of Matthew Dellavedova and John Henson who also expire in 2020. Were it not for Larry Nance Jr. (who does have trade value) and Love, the team could wipe the entire slate clean for 2021. They could be deal-breakers in free agency the next two summers, but it will take patience.

The final relevant trade target, obviously, is Kevin Love. But as astute readers know, he has a $120 million dollar extension that goes for four years. He’ll be making roughly $30 million dollars a year until 2023. But he’s a 30-year-old power forward who probably won’t make an All-Star game again. What’s the market? The issue isn’t just the production — last year he scored efficiently with his usual solid rebounding — it’s the lack of a want for a player of his ilk. The shift in the frontcourt has meant that power forwards have to cover increasingly more mobile players, and the center or nominal center is the often the sole rim protector because so much of the game has moved to the perimeter.

Love can’t keep up with quicker players, and you don’t want him to be the only defender near the rim. He’s proficient at a position that doesn’t really exist anymore. His contract alone will be tough to move, but with the lack of need for his type of player we could be in for a long wait or a disappointing trade for Cleveland, who may have to gift up something of significant value just to clean the books. The Cavaliers have little talent left on the roster, but they have just enough to be involved in some trades.

However, look at it from this point of view: he’s injured now, but even if he’s on the court, how much value will he provide? If his value is so low now that teams won’t trade for him, then he won’t boost his team’s wins and kill their lottery odds. Being patient is probably the right move here. He could have a rejuvenated season next year, and it only takes one franchise with an itchy trigger finger who wants a proven name to fight for low playoff seeding to sell over fans and management.

Does the 3FG column decide wins?

I have one nit to pick here with Tom Haberstroh’s excellent piece about analytics and the 3-pointer. I don’t think it’s disappointing that the 3-pointer fell behind a few other stats in terms of who “won” the box score — namely, the win percentage of teams that led in a specific box score category. We also have to factor in how often the event occurs. The 3-point shot is only one part of the game; they happen only roughly 11 times a game for a given team. Of course, field-goal percentage is more important. Not only does a field-goal happen more times a game, but 3-pointers are included! If you compare only 3-pointers to 2-pointers, then the 3-pointers actually win by a whisker — that should clue you in on their potency.

The actual best box score stat? Points … if you beat your opponent in the points column, you kinda win (and by kinda I mean by the strict interpretation of the NBA rules.) The NBA is decidedly simple like that, and it’s all about the long, winding paths you can get to those points and outscore your opponent. Sometimes it’s about 3-pointers. Sometimes it’s more about defense. And that’s okay. The analytics movement pointed out how, and more importantly why, 3-pointers were undervalued. It’s our job to keep working on the why in various situations, not to declare victory over everything.

Robert Covington’s defense is real

I just wanted to point out that Robert Covington, long an analytics darling with many metrics pegging him as a near star, has gotten the ultimate test of his defense. He switched to another environment, one that was actually struggling on defense, and the results are clear so far: the Timberwolves have been 6 points per 100 possessions better with him on that end of the court. If you make all the fancy adjustments you want with teammates and stats, he still looks good. Let’s hope this guy gets healthy soon and returns to the court.

Trae Young: bust or bad season?

Trae Young had a lot of hype coming into the season. He got the Stephen Curry comparison due to his proclivity for launching long 3-pointers and size, but so far his season has been a disappointment. It’ s not for lack of opportunity — few rookies have gotten more minutes. It’s just that he’s had a lack of production, and for all the talk of his shooting prowess, he’s shooting terrible.

Let’s look back at one of his higher scoring games back in December against the Cavaliers, which I highly doubt most readers viewed. At first glance on paper, he had a good game, scoring 21 points with 9 assists. You can see how his range is tough to cover here, for instance.  He bombs away from 32-feet because his defender was rolling back to where Alex Len was setting up for a pick, which is a perfectly logical reaction in most instances, even though that may not mean he’s open more often because of his long-range shots. With all that attention on his outside shooting (given his 3-point percentage that’s all reputation), it opens up the court for him to drive or go for open shots inside the line like this floater.

However, for all the revelry of Young’s range, he’s shooting just 28 percent from 25 to 29 feet. The Hawks, already an awful team, somehow manage to play worse with him on the court — one part of a season doesn’t prove anything, but it’s not a great sign. Defense, naturally, is where the Hawks fall apart when he plays. He regularly lets point guards get right to the rim, like Dellavedova here or Terry Rozier in this play. Picks, in general, cause him a lot of trouble; watch him stumble here and go under the pick even though his teammate is clearly there and further inside. Some of that is poor team communication and the blame can be shared, but Young isn’t helping in any appreciable way.

Young is the classic empty-calorie shooter. He doesn’t help win any rebound battles. He picks up assists but mainly because he has the ball frequently, he loses value with his high turnover rate, his free throw rate is too low to save his efficiency, he’s below average finishing at the rim, and his defense is predictably a running disaster. Without a top-tier shooting performance, it will be nearly impossible for him to provide any value on the court. That’s the razor’s edge his career has to walk, and if his 3-point percentage doesn’t sky-rocket upwards it won’t be a very long one.

Choose five: Counting lineup possibilities

I’d like to end with a small dive into some math surrounding a question that’s been nagging at me for a while: how many possible lineups can an NBA team have? With the wealth of lineup data out there, we’re constantly bombarded with stats about how X lineup is the team’s best lineup or how well a certain five-man set is doing. But I don’t think people understand just how many possible lineups there are, and how quickly the stats can be subdivided into small, highly noisy lineup sets. Just how many lineups are possible and how quickly do they build up?

Let’s start with a simple example. Say you’re playing two-on-two, and there are four people on your team. It’s just you, Giannis Antetokounmpo, Pascal Siakam, and Nick Young — it’s a weird day at the YMCA. How many possible lineups are there? Well, you’ve got three that include yourself because you could play with each guy. Then there’s Antetokounmpo and Siakam, Antetokounmpo and Young, and finally Siakam and Young. That’s six total, and it seems pretty easy to manage.

Now let’s say Boban Marjanovic shows up and wants to play on your team. Who’s going to say no? The new lineups all involve Boban, of course, so that’s four more lineups for a total of ten. So yes, even with just more one player, you nearly doubled the combinations. But then what happens when it’s a lineup of, say, three and a team of eight? With some patience, you could write down every possible combination, but that’s tedious. Instead, let’s dive into the math for a direct calculation.

You can see a proof online like this one, but it’s not too hard to conceptualize. Think about just ordering those five above players in any possible order. You can choose any of the five for the first spot, and then the second in line would be chosen out of four because one player was already taken, and the third in line would be chosen out of three, etc. Mathematically, you can write that with a factorial: n!  or n*(n-1)*(n-2)* … 1.

What if they’re in any order but chosen in subsets we’ll denote as r? It’s the same as a factorial but you stop at (n – 1 + 1). Here’s a concrete example tweaked from the proof linked above: how many ways can you order three numbers from the set of 1 to 99? You have 99 choices for the first slot, 98 for the second, and 97 for the third. Going back to the formula, that’s n* … (n – r + 1.) Since (n – r + 1) is (99 – 3 + 1) = 97, it’s 99*98*97. You can rewrite the formula as n!/(n – r)!

How do you account for unique combinations (i.e. order doesn’t matter)? Let’s go back to the simple example of four players for a two-person team. If order matters, like unique positional lineups where being at the point guard slot is different than being the center, then it’s n!/(n – r)! or 4!/(4 – 2)! = 12. If order doesn’t matter, how do you quantify the reduction? Think about this: each set or team can be ordered a certain number of ways, and if order doesn’t matter then the reduction is the number of ways you can order that set. In other words, you divide by r! In this example, a team of two can only be ordered to ways, so once you lose ordering you lose half your combinations, or 12/2 = 6. And that’s the previous answer we found earlier for our two-player team.

The formal way of writing this is nCr  or n!/[ (n – r)!*r! ]. Finding the unique combinations of lineups for an NBA team is now trivial after all that work. You can even use this online calculator or you can even search Google “12C5.” For a 12-man roster, it’s 12!/[ (12 – 5)!*5! ] = 792. And with a roster of 15 that balloons to 3003. Remember why this is happening. At the top of the formula, it starts with 15*14*13 and on and on. Adding even just one more player drastically increases the number of possibilities through all the new paths of combinations created. Just consider this: the 1997 Dallas Mavericks, graced with the most players given minutes in a single season for a team with 27, had 80,730 potential combinations for lineups. Of course, many of those players were not on the roster at the same time, but it’s an illustration of how quickly the numbers can build.

The next time someone cites a lineup stat go check how much that lineup has played and see how many lineups that team actually has with significant minutes. There are hundreds of ways to build lineups, and once you dig into the reserve players and trades it gets even worse. There are thousands of NBA lineups this season, and we have to be careful about the interpretation. Small sample theater reigns as there are too many possible combinations to use for a season. With a little bit of math, it’s easy to understand why.