clock menu more-arrow no yes mobile

Filed under:

How to compare players across eras

Since Chris Russo, Stephen A. Smith and JJ Redick showed us what not to do

NBA: Playoffs-New Orleans Pelicans at Phoenix Suns
Is Chris Paul a transcendent PG?
Mark J. Rebilas-USA TODAY Sports

In theory, I should be writing about the Wizards season or the playoffs. But I’m stuck on this exchange on whatever show this, in which literally everyone seemed in competition to make aggressively stupid comments.

Okay, was Bob Cousy, as Chris Russo asserted, the greatest point guard ever? Nah. Is Chris Paul “not transcendent” as Russo contends? Hell no — as Stephen A. Smith understates Paul is one of the top five PGs ever.

Should Cousy be dismissed or downgraded because he never shot over 40%, as JJ Redick pointed out, or because he was “defended by plumbers and firemen” as Redick said and Smith concurred? Also no.

Comparing players across eras can be tricky because the contexts often radically over time. For example, I recently had a conversation with someone who compared Jordan Poole to Gilbert Arenas at comparable stages of their career. The numbers were similar, but the comp didn’t work because the contexts were different.

To illustrate: Poole’s offensive rating this season was 112. In Arenas’ second season, his ortg was 108. Hey, Poole’s more efficient. He’s even better than Arenas. Except, league average this season was 112.0. In the Arenas comp season, league average was 103.6.

The usual way that we debate players from different areas is as dumb we debate just about everything. The Time Machine exercise is pointless. Time travel Cousy directly from the 1950s to 2021-22, and he doesn’t make the league. Time travel Chris Paul to the 1950s, and he destroys the league. Hell, time travel Kyle Kuzma to the 1950s, and he dominates because there just weren’t 6-10 guys with handles and the ability to shoot with range.

Another popular method is marginally less dumb, but still in the dumb ballpark: it’s the baby time travel theorem. It goes something like: “If Bill Russell was born in 2003 and was entering this year’s NBA draft, he’d be 6-10 and 245 pounds and with guard handles and three-point range and elite defense and...” You get the point.

There are better ways, though. Consider and evaluate players within the context of their own environments. Find ways to measure relative dominance and then compare those levels. So, back to Cousy for a moment — he may have been defended by plumbers and firemen, and the league he played in was not fully integrated so he didn’t compete against all the best players — but there was nothing he could do about that.

What Cousy could do is play in his own era against the competition of that era. Same for Chris Paul and Bill Russell and everyone else who ever lived. And what we can do is try gauge how each player fared against the competition of his era, and then compare the levels of relative dominance.

One way of doing that is to look at accolades and awards. Cousy, for example, was all-league 12 times and league MVP once. He was the assists champ eight times. That’s an impressive resume. In comparison, Paul has been All-NBA 10 times (he probably gets his 11th this season), and led the league in steals six times and assists four times.

Then you have to get into other factors like that Cousy played in a league that had just eight teams and two rounds in playoffs. So maybe his level of competition was lower? On the other hand, Paul played a post-expansion era when there used to be complaints that the league was too diluted. And, there rules get changed, coaching evolves, training and fitness practices improve...

It would be easy to get stuck.

This leads me to my final approach: Era Translated Stats. To grossly simplify, let’s say we have two players who each average 20 points per game. Now, let’s say Player A gets his 20 per game in a league where the average team scores 110 points per game. And Player B gets his 20 in a league where the average team scores 90. Which is the better scorer? Player B, of course, because he’s scoring a greater portion of all points scored.

One of the things I found interesting when I started running era translations was that player types repeated. For example, in 1961-62, Bill Russell averaged 18.9 points, 23.6 rebounds and 4.2 assists per game. Translate those numbers to the 2021-22 environment, and they come to 17.6 points, 14.7 rebounds and 4.3 assists. Those are basically the numbers of a better passing Rudy Gobert (15.6 points, 14.7 rebounds, 1.1 assists). In previous seasons, an era translated Russell looked like Dwight Howard, Ben Wallace or Hakeem Olajuwon.

Oscar Robertson’s famed triple-double season (1961-62) — 30.8 points, 12.5 rebounds, 11.4 assists — translates to 28.7 points, 7.8 rebounds and 11.7 assists this season. That’s basically James Harden plus some scoring. Or Trae Young plus some rebounds. Or Luka Doncic plus some rebounds and assists.

The translations work in either direction. The Big O was often compared with Magic Johnson, and in Magic’s best season (1988-89) he averaged 22.5 points, 7.9 rebounds and 12.8 assists. Translated to 1961-62: 24.5 points, 12.8 rebounds and 12.0 assists.

Joel Embiid’s 30.6/11.7/4.2 stat line translates to 32.9/18.8/4.1 in 1961-62. That would have landed him second in scoring (behind Wilt Chamberlains 50.4 average) and fourth in rebounds per game (behind Chamberlain, Russell and Walt Bellamy).

Weirdly, the translation can also adjust shooting for era. In Cousy’s MVP season, he shot 39.7% during a season where the league shot 38.5%. That’s equivalent to shooting 47.5% this season when the league shot 46.1%. It’s kinda crazy to imagine a dominant player shooting 37.5% for his career and never shooting better than 40% for a season, but...Cousy was average to a little below average at shooting when compared to the league he played in. The league didn’t crack the 40% barrier until the 1959-60 season — Cousy’s 10th in the NBA.

I’ll close with a few more not-exactly-random translations:

  • Wilt Chamberlain — In 1961-62, Chamberlain averaged 50.4 points, 25.7 rebounds and 2.4 assists. (He also played every minute of every game and scored 100 points in a single game, which (by using my era translator) I determined was probably not the single best scoring night in NBA history.) Translated to the 2021-22 season: 46.9 points, 16.0 rebounds and 2.5 assists.
  • Shaquille O’Neal — In 1999-00, Shaq averaged 29.7 points, 13.6 rebounds and 3.8 assists. Translated to 1961-62: 36.2 points, 22.6 rebounds, 4.1 assists.
  • Michael Jordan — 1990-91: 31.5 points, 6.0 rebounds, 5.5 assists. Translated to 1961-62: 35.2 points, 9.9 rebounds, 5.3 assists.
  • Russell Westbrook — 2016-17 triple-double season (his best): 31.6 points, 10.7 rebounds, 10.4 assists. Translated to 1961-62: 35.6 points, 17.6 rebounds, 11.0 assists. Compare with Robertson’s triple-double average that season: 30.8 points, 12.5 rebounds, 11.4 assists.

I’d be remiss if I didn’t include some Wizards/Bullets.

  • Bradley Beal — best season, 2020-21: 31.3 points, 4.7 rebounds, 4.4 assists per game. Translated to 1961-62: 33.2 points, 7.6 rebounds, 4.3 assists. A player with those numbers that season would have been the league’s second leading scorer behind Chamberlain.
  • John Wall — best season, 2016-17: 23.1 points, 4.2 rebounds, 10.7 assists per game. Translated to 1961-62: 26.0 points, 6.9 rebounds, 11.3 assists. Those numbers would have ranked 7th in scoring, 26th in rebounding and 2nd in assists (behind Oscar Robertson).
  • Wes Unseld — best season may have been his 1968-69 rookie year when he joined Wilt Chamberlain as the only players to win MVP and Rookie of the Year in the same season. His averages: 13.8 points, 18.2 rebounds, 2.6 assists. Translated to 2021-22: 13.6 points, 14.2 rebounds, 2.8 assists. Closest comp: Gobert.
  • Elvin Hayes — best season, 1976-77: 23.7 points, 12.5 rebounds, 1.9 assists. Translated to 2021-22: 24.6 points, 11.8 rebounds, 2.0 assists. Closest analog might be Karl-Anthony Towns...if Towns grabbed more rebounds. Jarrett Allen plus scoring? Anthony Davis plus rebounding?
  • Gilbert Arenas — best season, 2006-07: 28.4 points, 4.6 rebounds, 6.0 assists. Translated to 1961-62: 34.2 points, 8.0 rebounds, 6.7 assists. That season, a Lakers guard averaged 30.8 points, 7.9 rebounds and 5.4 assists. That guard: Jerry West. Translate Arenas to 2021-22, and 1) the numbers are kinda mind-blowing, and 2) we can add the rest of the stat line. So, the translated to 2021-22 numbers: 31.8 points, 5.0 rebounds, 6.9 assists, 2.0 steals, 2.2 turnovers per game. Where things get kinda mind-blowing: in 2006-07, Arenas had 7.9 three-point attempts per game. The average team attempted 16.9 threes per game. His three-point attempts translated to this season: 16.5 per game. This season, Stephen Curry led the league in per game three-point attempts: with 11.7. That’s the league record.

Who else interests you? Drop the name in the comments, and I’ll run a few more through the translator over the next few days.