How Deep is the NBA At Each Position?

Recently, there has been a lot of talk on Twitter about the “top 10 or 15 point guards” in the league, and other superlatives that are so oddly specific that they start to veer into “great, you just designated a guy above average, what are we supposed to do, make him an All-Star?”

After all, there are only 30 teams and therefore 30 starting point guards, and while some team’s backups may be better than other team’s starters, the fact remains that saying someone is the 15th-best point guard in the league is essentially saying “this guy is league-average for a starter.” And they’re treating that like the guy’s an All-Star!

We can look down the list of guys at every position, rank them by every stat you can imagine (counting or advanced), and start the debate machine moving, but the bigger question is how far down the list of best guys at any given position do you have to go before the debate is less about who’s an All-Star and more a debate about “is a team getting fleeced paying this guy the veteran minimum and eating a salary slot?”

So let’s go ahead, take guys at each position, and in essence ask “who’s the best guy who sucks?” to determine how far down a Top list you need to go—10, 15, 30, whatever—before the list becomes a case of scraping the barrel.

Now, because we’re using Basketball Reference for this, which necessarily requires broader definitions of position than the more restrictive definition people use to describe “point guard” or “shooting guard” (and indeed, for most players, a simple breakdown of their minutes shows there is rarely such a thing as a guy who only plays one “traditional” position), that’s going to color our view a bit.

But let’s say, if only for the sake of argument, that excepting a necessary overlap between shooting guard and small forward and another one between power forward and center, that there are 60 starters at each of the front/backcourt Basketball Reference positions and 30 at center.

Now then, let’s do this by VORP, because in my testing, it created a nice set of intuitive data to work with and works great as a catch-all because it’s a cumulative stat.

Value Over Replacement Player is basketball’s answer to baseball’s WAR (Wins Above Replacement) and makes a fine starting point. Anyone below zero is bad enough that he belongs in the G-League, anyone above 2 is at least a solid starter, higher than that and you’re into All-Star and MVP territory (for point of reference, LeBron James led the league in VORP last year with 8.9.)

Guards

The top 5 guards: James Harden, Damian Lillard, Russell Westbrook, Stephen Curry, and Kyrie Irving.

See? Pretty good stat for “that guy is good”, isn’t it?

Continuing down the list, Ben Simmons is sixth, Marcus Smart is 21st, and Lonzo Ball is 32nd.

Of 249 guys in the sample, Collin Sexton is the worst (-2.0), followed by Jamal Crawford, Elie Okobo, Eric Gordon, and Antonio Blakeney. Exactly 100 players are below zero.

Who’s the first guy we find who brings up “he’s not that good, what’s his name doing on the list?” Well, there’s Darren Collison at 18th, Chris Paul (who’s nobody’s idea of a great player in 2019) in 25th, and Ball at 32nd. Any would be a good cutoff, but if we apologize to Smart and say the “top 20 guards” are going to be guys who are good by VORP, that leaves De’Aaron Fox as our cutoff. Not a bad place to cut the list.

Forwards

Similarly, at forward, if we run the same calculation, our top five guys are Giannis Antetokounmpo, Anthony Davis, Paul George, Blake Griffin, and LeBron James. Kevin Durant is sixth, Montrezl Harrell (yes, really) is seventh (making his Sixth Man of the Year case in the process), and the rest of the top 11 are Kawhi Leonard, Pascal Siakam, Jimmy Butler, and Domantas Sabonis (the other huge candidate for 6MOY.)

Where’s the dropoff on this one? Well, 18th (Willie Cauley-Stein), 19th (Kevon Looney), and 28th (Draymond Green) are all plausible cutoffs; all of those guys are either benefiting from VORP counting their stats in a way that doesn’t square with what’s going on down on the court or in Green’s case having a demonstrably rotten year but still getting the benefit of the doubt from the advanced stat.

Cut it off at 20 and you land on LaMarcus Aldridge…yikes. Let’s just say forward goes no more than 17 (Tobias Harris) deep.

Centers

Finally, the centers, and here things get a little weird, in part because of the questionable worth of an outside-shooting big man compared to a rim-dominant player.

The top 5? Nikola Jokic, Rudy Gobert, Karl-Anthony Towns, Nikola Vucevic, and Jusuf Nurkic, respectively.

Joel Embiid is eighth. Myles Turner is 11th. And Brook Lopez is 12th. The stat doesn’t like three-point shooting centers unless they do something else exceptionally well (consider Jokic and his very wide lead in assists among centers.)

There are 60 centers in the league by Basketball Reference’s definition. All but three (Channing Frye, Dwight Howard, and Greg Monroe) have a positive or neutral VORP.

But if we look at the list, the true stars are in the top 4 and the guys with a case are in or floating around the top 10 (only DeAndre Jordan, in 10th, raises real questions.)

Conclusions

The moral of the story here is that if a guy is a top-50 player in the NBA, he’s pretty good. And furthermore, you can pick 10 guys from each of the five cardinal positions, build your Top 50 that way with an eye toward balance, and get…well, the starting lineups for 10 playoff teams that would have reasonable expectations of making the second round (assuming a good 4-5 matchup in each conference.)

So to answer the question asked at the top of the show here, how deep is the NBA at each position?

Well, from the looks of it, “top 10” is a perfectly good superlative to work from, so let’s stick with that. Top 15 if you want “could plausibly start on a playoff team.” Below that, things get…questionable, especially at center.

So keep at it with those lists. They make great Twitter fodder and it’s usually pretty easy to find a stat that isn’t a blatant cherry pick if you want to support your arguments.