Nylon Calculus: How to build NBA lineups stronger than the weakest link

Which types of players become liabilities in the playoffs and how can we build NBA lineups that will protect these “weak links” from being exploited?

The playoffs are harder than the regular season. The competition is better. There is more time for scouting and game planning. Coaches sniff out their opponents’ weaknesses and players exploit them ruthlessly. In the postseason, stars play more minutes and benches shorten, defenses lock down and the intensity ratchets up — it’s just a whole different animal.

Each year, some players who were really good during the regular season will end up becoming less effective during the postseason. But which types of players are most likely to become liabilities in the playoffs? And how do we build lineups that hide our players’ limitations and showcase their strengths? How can we become stronger than our weakest link?

I built the Weak Links app to help us answer these questions.

The app asks you to define a player’s archetype based on his position (guard, wing, or big) and his talent grades (for perimeter shooting, playmaking, perimeter defense, and interior defense) and it shows you how often and how well lineups with this type of player have performed. These talent grades — which were developed by Tim (@Tim_NBA) and Jacob (@JacobEGoldstein) at the Bball-Index — use publicly-available stats to provide objective measures of skill designed to be independent of team-specific factors such as coaching, schemes, and teammates. I linked players’ percentile ranks in these talent grades with six years of lineup data from PBPstats.com to find the types of players that have been weak links for their teams during the regular and postseasons from 2013-14 to 2018-19.

In the app you can choose to include or exclude teams that missed the playoffs and, in this post, I’ll opt to exclude them. This will allow us to compare two consistent sets of teams each year and it will ensure that the personnel available to the coaches was more-or-less the same during the playoffs as it was during the regular season (with the obvious caveats of trades, injuries, suspensions, etc.). As a consequence of this approach, the average net rating across all 700,000+ regular-season possessions was +4.6 points per 100 possessions, as playoff-bound lineups tended to feast on weaker teams. Likewise — because the app ignores any lineups that played together for less than five possessions in a given season — the average net rating in the playoffs was also a tick above zero, at +1.9 points per 100 possessions. As such, it’s useful to compare the lineup net ratings that we’ll come up with to the league average, rather than just assuming that anything with a plus-sign in front of it is good.

How can we build better NBA lineups?

The table below shows us how lineups have performed relative to the league average when burdened with one of five different types of possible weak links: a “minus” perimeter scorer, a minus perimeter defender, an immobile rim protector, a non-shooting ball handler, or a spot-up 3-and-D wing.

In both the regular season and in the playoffs, lineups with a minus perimeter scorer fared the worst. I defined this archetype as any guard or wing who graded out as a “D” or “F” in perimeter shooting while also being a “D” or “F” in play-making — including inconsistent spot-up shooters like Mo Harkless, Michael Kidd-Gilchrist, and Luc Mbah a Moute. Lineups with this type of player accounted for 11 percent of possessions during the regular season while posting an average net rating of +1.7 points per 100 possessions, which was -2.9 below the league average for playoff-bound teams. In the postseason, lineups with this type of offensive dead weight became less playable, accounting for only 7 percent of possessions and averaging a lower net rating of -2.8 points per 100 possessions, which was -4.7 below the league average. Because the average opponent will be tougher in the playoffs than in the regular season, we can expect to see a big dropoff for any lineup we evaluate (-2.7 points per 100 possession, on average); but the dropoff for lineups with a minus perimeter scorer was even more dramatic than usual (-4.5 points).

The song was the same for minus perimeter defenders, but the music was a little softer. I defined this archetype as any guard or wing with a “D” or “F” grade in perimeter defense — including guys like Marco Belinelli, Seth Curry, and Damian Lillard. Lineups with matador defenders like these accounted for 34 percent of possessions during the regular season with an average net rating of +2.3 points per 100 possessions, which was -2.3 below the league average for playoff-bound teams. But in the postseason, where opponents tend to hunt mismatches more relentlessly, lineups with minus perimeter defenders played less often, accounting for just 29 percent of possessions and averaging a net rating of +1.0 points per 100 possessions, which was -0.9 below the league average.

The immobile rim protector is another type of player who is sometimes maligned for getting “played off the court” during the postseason. Using the app, I defined this archetype as elite interior defenders — in the top 10 percent of the league — with a “D” or “F” grade in perimeter shooting, playmaking, and perimeter defense including big stiffs like Rudy Gobert, DeAndre Jordan, and Roy Hibbert. Lineups with defensive anchors like these gave up just 103.6 points per 100 possessions during the regular season (less than any other type of lineup that I checked in the app), with an average net rating of +6.2 (+1.6 above the league average for playoff-bound teams). However, the proportion of shots attempted at the rim has dropped by several percentage points from the regular season to the postseason in each of the past two years, undercutting the value of these rim-protection specialists. In the postseason, lineups with immobile rim protectors were still used as often as they were during the regular season, but they performed much worse, averaging just +1.5 points per 100 possessions (-0.4 below the league average).

Non-shooting ball handlers are another potential weak link for playoff teams. In the table, I’m defining this archetype as guards who are in the bottom-50 percent of perimeter shooters and the top-10 percent of playmakers — guys like Ben Simmons, Rajon Rondo, and Russell Westbrook. Lineups with bricky ball handlers like these accounted for 11 percent of possessions during the regular season while producing an average net rating of +4.1 points per 100 possessions, which was -0.5 below the league average for playoff-bound teams. In the postseason, these lineups played just as often (accounting for 10 percent of possessions) and continued to perform close to league average at +1.8 points per 100 possessions.

Jonathan Tjarks’ assertion that “In the NBA playoffs, you’re only as strong as your weakest link” was part of the reason I wanted to make this app. He wrote about the way that spot-up 3-and-D wings like Robert Covington and PJ Tucker were neutralized by smart defenses during the 2018 playoffs.

“Elite teams need offensive threats at every position. The playoffs have shown that it’s too easy to game-plan pure 3-and-D players out of a series.”

In the table, I defined this archetype as wings with an “A” or “B” for perimeter shooting and defense grades and a “D” or “F” for playmaking who were not rim protectors (ie. they were 75th percentile for interior defense grade or worse), which included Covington, Tucker, and DeMarre Carroll. I found that lineups with stationary shooters like these actually played about the same amount in the playoffs as they did in the regular season with comparable levels of success (+0.4 points per 100 possession above the league average net ratings for playoff-bound teams).

OK — so those are some examples of the types of players that have the potential to be a weak link on a playoff team. The fun thing about the app is that we can select archetypes for possible teammates to be slotted in next to our weak link and then evaluate how their complementary skills might protect them from being exploited by the opponents. For example, if we put our spot-up 3-and-D wing on the court alongside four playmakers (ie. players with “A” or “B” play-making grades), the results are really good — with average net ratings of +7.3 and +6.3 points per 100 possessions in the regular and postseason, respectively. In contrast, lineups that combined a spot-up 3-and-D wing with fewer than two playmakers were pretty much unplayable, especially in the playoffs (accounting for less than 3.5 percent of possessions).

Similarly, we saw how a minus perimeter defender has the potential to become a weak link in the playoffs; however, pairing that weak perimeter defender with a strong interior defender can mitigate his negative impact. The app allows us to ask — how good does the interior defender need to be to carry that weak perimeter defender? And how many weak links can one rim protector cover for? I paired an elite interior defender (ie. 90th percentile interior defense grade or better) with different numbers of plus (ie. “A” or “B”), minus (ie. “D” or “F”), or neutral (ie., “C”) perimeter defenders and summarized how well the different combinations played.

Playing an elite rim protector with a bunch of plus-perimeter defenders is generally a recipe for success, whether it’s the regular season or the playoffs. Interestingly, there were three distinct lineup combinations that successfully absorbed one weak-link perimeter defender while achieving playoff net ratings of +5.4 points per 100 possessions or higher. Generally speaking, lineups with two negative perimeter defenders were used much less frequently.

OK, so those results suggest that an elite rim protector can make up for one but probably not two weak perimeter defenders; but what happens if your rim protector is just average and not elite? Can you still get by with a weak-link defender or will he cause your whole defense to break down? Here iss a look at what happened when lineups combined an average interior defender (ie. 75th to 90th percentile interior defense grade) with different numbers of plus (ie. “A” or “B”), minus (ie. “D” or “F”), or neutral (ie., “C”) perimeter defenders.

If you’re lucky enough to have four plus perimeter defenders, an average rim protector is all you need to produce solid net ratings of +6.2 and +6.0 points per 100 possessions in the regular season and in the playoffs, respectively. However, without an elite rim protector, you sacrifice the luxury of being able to cover up any weak-link perimeter defenders. In the previous chart, there was a trio of lineup combinations with a single weak-link perimeter defender that each had a playoff net rating above +5 thanks to the back-up provided by an elite rim protector, but those same combinations each had a net rating below +2 points per 100 possessions when there was an average rim protector patrolling the backline instead.

We can do the same sort of analysis in thinking about how a non-shooting ball handler like Ben Simmons might be able to lead a team deep into the playoffs. What other pieces does he need around him to minimize his limitations? One obvious answer would be to surround him with good shooters. Here is a look at how lineups have fared when pairing an offensive initiator (ie. 75th percentile playmaker grade or better) who was not a good perimeter shooter (ie. “C”, “D”, or “F” perimeter shooting grade) with different numbers of plus (ie. “A” or “B”), minus (ie. “D” or “F”), or neutral (ie., “C”) perimeter shooters.

If you combine a non-shooting guard with two plus perimeter shooters it tends to work pretty well, even in the playoffs (as high as +9.4 points per 100 possessions). If you only have one plus perimeter shooter the results are less impressive (as low as -1.5 points per 100 possessions). And if you don’t have any plus shooters in your lineup, your prospects will be pretty bleak in the playoffs (-5.2 points per 100 possessions).

There’s one last type of player that I snuck into the first table which I haven’t mentioned yet: playmaking bigs. In September, Rob Mahoney wrote that playmaking bigs like Nikola Jokic, Draymond Green, and Marc Gasol are becoming the key to playoff basketball.

“Whether a center can space the floor—and defend in space—is now their defining feature. It has become a shorthand for whether a big will be able to survive in the unforgiving climate of postseason basketball, in which defenses will exploit any player who doesn’t present an immediate threat to score. Shooting at the 4 and 5 offers the simplest form of relief. Passing from those positions, however, unlocks the greatest possible advantage.”

In the first table, playmaking bigs were the only archetype that tended to be a part of above-average regular season lineups (+2.8 points per 100 possessions above league average for playoff-bound teams) which became even MORE effective in the playoffs (+4.7 above average). The chart below shows that surrounding a big playmaker (ie. “A” grades in playmaking and interior defense, like Jokic, Green, and Gasol) with plus perimeter shooters (ie. 75th percentile perimeter shooting grade or better) can maximize their advantages.

Just as we saw for the lineups with non-shooting guards: the more shooting, the better. The difference is that — whereas it was difficult for coaches to surround a non-shooting guard with three plus perimeter shooters — it was much more feasible to combine a playmaking big with three perimeter shooters. For example, lineups with a playmaking big, three plus shooters, and one neutral shooter were pretty common, accounting for 7 percent of possessions in the playoffs and also pretty successful, beating opponents by +7.5 points per 100 possessions. Having two perimeter shooters also worked well, but lineups with just one perimeter shooter weren’t as good and lineups without any perimeter shooters couldn’t even stay on the court.

Do you have your own hypothesis about what types of players seem to turn into pumpkins during the playoffs each year or what types of players seem to become even more helpful when their team reaches the postseason? Maybe you have a theory about lineup combinations that will or won’t help a team win a championship. Try out the Weak Links app and stress test the idea for yourself!