In Search of the “League Average” NBA Player (Part 2: WS/48)

Last week, we looked at PER to try and figure out just what constitutes a “league average” NBA player. Since that stat is normalized to 15 every season, it gave us a cross-section of players across positions and even perceived talent levels (any time you have Brook Lopez, Jayson Tatum, and Terry Rozier on the same list, that’s a wide swath to cut.)

But Tatum especially is nothing even resembling “league average”. He’s become the breakout star on a Celtics team that has made three of the last four Eastern Conference finals. Hardly “league average”.

So let’s try a stat with a more direct correlation with NBA wins and losses, namely Win Shares per 48 minutes.

In an ideal world, a team made up entirely of guys with .100 WS/48 would win exactly 41 out of every 82 games—and it is officially weird that we won’t see another 82-game season until at least next year and possibly never if rumors about the league shortening the season to save player health and eliminate back-to-backs prove to be true and the 72-game seasons of the past couple of years serve as the catalyst. But I digress.

Anyway, same as last week, the methodology here is to go over to Basketball Reference, look up players who are reasonably close to that .100 WS/48 for a reasonable number of minutes (at least 1000 minutes played, so covering most if not all rotation players, even starters who got hurt) in 2019 and 2020, and see if there are any lessons to be found in the data.

Pulling up the list and keeping it between .095 and .105 WS/48, we get a 53-player sample size, which is better than the 20 guys we got last week for PER between 14.8 and 15.2, so it’s clear that unlike PER, WS/48 tends to like to cluster players in the middle rather than punt them well outside the band in the center of the stat’s measuring area.

At the top of the list, we have Jaren Jackson‘s rookie campaign with Memphis in 2019, where he posted .105 WS/48. He nearly matched that in his sophomore season in 2020, posting .103 WS/48. He’s injured this year, so no word on whether his floor is his ceiling (a problem that plagued Kristaps Porzingis in New York and, despite a solid 2020 campaign, seems to have caught up with him in Dallas; three of his four full seasons in the league hover right around .100 as well.)

That’s right…something Mavs fans suspected as Porzingis has regressed hard in this disappointing campaign (the Mavs are 12-14 and stand 11th in the West through games of February 10) seems to be holding true, and Porzingis is no better than a league-average power forward.

But back to the list.

Other names that show up and are particularly noteworthy are Marcus Morris, Tomas Satoransky, Goran Dragic, Derrick Rose, Luka Doncic‘s rookie year—he too has regressed from .207 WS/48 last year to just .147 in 2021, but as a rookie he was league-average—Kevin Love, Russell Westbrook, and Victor Oladipo.

Ain’t it funny how fallen stars tend to look…well, pedestrian as they get old or get hurt.

Perhaps more telling is the guys who show up on the list in both 2019 and 2020. Besides Jackson, there’s Jae Crowder and Kentavious Caldwell-Pope.

On top of that, there’s the four guys who are both league average for PER and for WS/48: there, you see Reggie Jackson, Jalen Brunson, Derrick White, and Tatum’s 2019 season.

So as we get this study over half court and set up near the basket to slam it home, we’re starting to see a pattern developing—Jackson, Brunson, White, and the 2019 version of Tatum (the year before he became an All-Star for the first time and posted 3.4 VORP in 66 games at age 21) are or were fairly pedestrian NBA players in their year. Crowder and KCP are practically the epitome of “he’s just average, not great, not awful, just…average.”

But this also invites the question of just how good “average” is supposed to be. PER offers us no discernible sense of how good a player is supposed to be in order to be considered average other than putting up a certain set of largely offensive counting stats that minute-for-minute make them good fantasy basketball plays.

But WS/48 ties this to a .500 record in actual NBA competition, and as such there are a lot of rotation bench guys who make the list.

Consider the nine guys on the list who started fewer than five games in the season for which they posted the league-average WS/48, and their team’s record that year:

Zach Collins ’19 (POR, 53-29)
Doug McDermott ’19 (IND, 48-34)
Goran Dragic ’20 (MIA, 44-29, made Finals)
Norman Powell ’19 (TOR, 58-24, won title)
Marvin Williams ’20 (traded midseason from CHA to MIL)
Michael Kidd-Gilchrist ’19 (CHA, 39-43)
Alfonzo McKinnie ’19 (GSW, 57-25, made Finals)
Jordan Clarkson ’20 (traded midseason from CLE to UTA)
Alex Caruso ’20 (LAL, 52-19, won title)

Four of these guys played in the NBA Finals as “league-average” bench players. Another four made the playoffs either on the team for which they played all year or for the team that picked them up at the trade deadline. Only one—MKG—was on a team that missed the playoffs, and they finished ninth in the East while finding ways to lose so many close games that if they’d just made a shot here and there (and not had one of the worst clutch players in NBA history in Kemba Walker choking away those games), they could’ve been the 6 seed.

So now we begin to see something. A “league-average player” is a guy who starts on mediocre-to-bad teams (generally) and, if you’ve got someone of his caliber coming off the bench, your starting lineup that guy can’t crack might just be good enough to win the title.

That’s powerful, and it speaks to just how important bench depth is. League-average WS/48 can, for the right guy playing the right role, be a game changer…or it can be a sign that your starting lineup simply isn’t very good.

And general managers need only look at the stats and the minutes to tell which of those is which.

Next week, we’ll look at “league average” VORP, which is going to be a much, much stickier issue to try and sort out, because while there is such a thing as a league-average VORP team (in the simplest possible terms, 10 VORP across an entire roster equals 41 wins), VORP isn’t a per-82-games or a per-48-minutes rate stat for players on Basketball Reference, so this is going to take a bit of digging and a change in methodology.

I’ve got a week to figure it out; that piece is coming next Thursday. Stay tuned, and thanks for reading!