NBA’s 75 greatest players of all time

[SouthernDukie]

Coincidentally/relatedly, I think this is a similar issue when assessing where to place Bill Russell among the all-time greats. Obviously nobody won more championships than Russell. But that was, to some extent, a function of his teammates (he entered the league in a year when Boston got the top TWO draft picks in Russell and fellow HoFer Heinsohn, not to mention the return of some injured stars and one of the greatest guards ever in Cousy; as he aged, the Celtics added Sam Jones and Havlicek to provide support as well) and a true mastermind at the helm (Auerbach) along with his own terrific talents. I think there were better individual players than Russell, but there clearly weren't better teams than those Celtics teams.

If you substitute Lakers for Celtics, and James Worthy for Bill Russell, you can make the same point.

[luvdahops]

One worthy candidate I have yet to see mentioned is Sidney Moncrief. He was arguably the top 2-way player in the NBA in the early to mid 80s, winning 2 DPOY awards and making 5 All-NBA Teams (1 First Team, 4 Second Team), while anchoring some very good Milwaukee teams that could never quite get over the hump against Boston and Philly. Moncrief's career declined from there due to recurring knee and foot injuries (not unlike Walton), but at his peak he was among the game's most respected all-around talents. Worth noting that, as a defender, Moncrief was considered by both Larry Bird and Michael Jordan to be among the very toughest they ever faced. He was also a key cog for some entertaining Arkansas squads in the late 70s.

I would personally put Moncrief in ahead of Maravich (who never won in either college or the NBA), Monroe and Wilkens, among backcourt players.

[Steven43]

Eh, I disagree. I think it's a pretty darn good measure. If you aggregate win shares and win shares per 48 minutes, you get a pretty darn good "who's who" of NBA history.

Here are the guys who are top-30 in both win shares and win shares per 48 minutes in NBA/ABA history:
Kareem: #1 in WS; #7 in WS/48 (in part due to playing well past his prime)
Wilt: #2; #3
LeBron: #3; #6
Karl Malone: #4, #22
Jordan: #5, #1
Stockton: #6, #18
Duncan: #7, #17
Nowitzki: #8, #27
Chris Paul: #10, #5
Robertson: #11, #20
Shaq: #12, #19
David Robinson: #13, #2
Barkley: #14, #13
Russell: #19, #28
West: #20, #15
Magic: #21, #8
Durant: #23, #12
Bird: #25, #23
Schayes: #29, #29
Harden: #30, #9

Obviously no single metric (or two) is perfect, but this seems to capture the list of candidates for greatest ever pretty well to me.

More specifically, I might or might not use it as a purely ordinal rating of players. But I would feel pretty comfortable looking at the list of all-time win shares and all-time win shares/48 minutes lists as a pretty good barometer for "should this guy be on the top-75 list?" criteria.

In terms of capturing an individual player's greatness, I think it's a pretty good measure.

I agree with you that it can be somewhat helpful in helping choose the best 75 or 100 players in league history. But the WS/48 is not particularly useful at helping choose the best 10.

Chris Paul #5? David Robinson #2? James Harden #9? Larry Bird was way better than all three of those guys and he’s #23? Nope.

[CDu]

If you substitute Lakers for Celtics, and James Worthy for Bill Russell, you can make the same point.

Yep, a similarity for sure. Russell was much better than Worthy of course, but conceptually a similar point. Worthy is another guy who probably shouldn't be on the top-75 list, although he certainly had some phenomenal playoff moments including a Finals MVP award.

[Acymetric]

I agree with you that it can be somewhat helpful in helping choose the best 75 or 100 players in league history. But the WS/48 is not particularly useful at helping choose the best 10.

Did someone in this thread use WinShares to make a case for their preferred top ten and I missed it? Seems like it is being used in exactly the way you say can be helpful.

[Steven43]

One worthy candidate I have yet to see mentioned is Sidney Moncrief. He was arguably the top 2-way player in the NBA in the early to mid 80s, winning 2 DPOY awards and making 5 All-NBA Teams (1 First Team, 4 Second Team), while anchoring some very good Milwaukee teams that could never quite get over the hump against Boston and Philly. Moncrief's career declined from there due to recurring knee and foot injuries (not unlike Walton), but at his peak he was among the game's most respected all-around talents. Worth noting that, as a defender, Moncrief was considered by both Larry Bird and Michael Jordan to be among the very toughest they ever faced. He was also a key cog for some entertaining Arkansas squads in the late 70s.

I would personally put Moncrief in ahead of Maravich (who never won in either college or the NBA), Monroe and Wilkens, among backcourt players.

Great catch! Moncrief has slipped through the cracks of time. And yeah, I remember Larry Bird giving him major respect.

[Steven43]

Did someone in this thread use WinShares to make a case for their preferred top ten and I missed it? Seems like it is being used in exactly the way you say can be helpful.

You’re right: I just thought of it as a secondary point while I was looking at CDu’s WS list.

And thanks so much for feeling the need to point it out! You are always so helpful. 😉

[CDu]

I agree with you that it can be somewhat helpful in helping choose the best 75 or 100 players in league history. But the WS/48 is not particularly useful at helping choose the best 10.

Chris Paul #5? David Robinson #2? James Harden #9? Larry Bird was way better than all three of those guys and he’s #23? Nope.

Nobody is suggesting using WS/48 alone to decide the best 10. I think you may have missed the point of that post.

Bird is a pretty unique circumstance. He got hurt relatively early in his career, and wound up playing a bunch of his career as a lesser version of his greatest self. So his greatness isn't as well captured in WS or WS/48. Had he not had the back problems, he'd have ended up much higher on both those lists.

But I think Paul and Harden are appropriately on that list. They are truly great players. Unfortunately for them, they generally played with inferior teammates or simply at the wrong time: Paul played with really bad teams for most of his career; Harden played on a super-teams just at the wrong time - too early and too briefly in his career in OKC, and when Irving was injured and making bad public health choices now). And they both suffered from being in the Miami/Cleveland and Golden State era of the NBA. It's not unlike the story for Robinson, who was a truly transcendent player stuck with bad teammates in the Jordan era for most of his career until the end with Duncan (who played the same position as him).

Now, I certainly don't think Robinson is the second best player of all time. But I do think he's highly underrated, and I think his win shares ranking is probably pretty close. He was an absolutely otherworldly player in the 90s. A 9-time All-NBA selection at center despite playing in an era of amazing centers (Ewing, Olajuwon, O'Neal, Duncan).

[FellowTraveler]

I don’t think this Win Shares thing helps very much in determining the greatest players in NBA history. I appreciate you talking the time, but….

I think measures like Win Shares and VORP that, while imperfect, have been developed with some level of analytical rigor are extremely helpful in making such determinations, as is setting clear and consistent criteria for how one assesses players rather than cherry-picking different measures for different players based on whims or preferred outcomes. Other people prefer to rely on subjective contemporaneous assessments by fellow players, news media, or fans at large. Still others prefer simply going by their own "I saw him play and I know how good he was" metric. YMMV.

I agree with you that it can be somewhat helpful in helping choose the best 75 or 100 players in league history. But the WS/48 is not particularly useful at helping choose the best 10.

Chris Paul #5? David Robinson #2? James Harden #9? Larry Bird was way better than all three of those guys and he’s #23? Nope.

No metric or subjective assessment is perfect, of course, and there are certainly valid arguments that a given metric over (or under) values a given player. Rudy Gobert's career WS/48, for example, is 10th-best all-time, and I think one could make a perfectly reasonable argument that this overstates his value and obscures his limitations, particularly w/r/t defending smaller, faster players who shoot well.

But I wouldn't be so quick to dismiss the usefulness of a metric because a couple datapoints don't match our expectations -- particularly doing so out of hand, without thinking about why or how it may do so -- or whether the metric is really an outlier, or perhaps it's our expectations that are the outlier.

I'll take Chris Paul, since he's first on your list of objections. And the thing about Chris Paul is that he is an absolutely elite, top-tier, inner-circle all-time great, and that’s true pretty much any way you assess his career. Other than, I guess, "I watch him play and I don’t agree." Which is certainly an opinion people can have; they don't need my permission for it. But it doesn't leave much room for discussion.

On offense, Paul gets baskets for himself and others, and he does both efficiently. On defense, he forces turnovers and does so without gambling too much.

Advanced stats? Chris Paul is 7th all time in VORP, 5th in Box plus/minus (top 10 in Offensive BPM, top 20 in Defensive BPM), 5th in WS/48, 10th in WS, 9th in PER.

Traditional stats? He’s averaged 18 ppg and 9.4 apg with only 2.4 turnovers per game for 1,100 games. Magic averaged a couple more assists again but also 1.5 more turnovers — and played nearly 200 fewer games. Oscar Robertson exceeds Paul’s PPG/APG numbers, but we don’t have turnover data for him. No other point guard in NBA history comes close to matching Paul’s combination of scoring, playmaking, and efficiency. Plus he led the league in steals per game six times.

(Paul’s ability to limit turnovers really can’t be overstated. The only player with a higher career assist percentage than Paul is John Stockton. Paul’s has put up a career assist percentage of 45.3 against a turnover percentage of only 13.3. Stockton? 50.2 and 20.8 -- slightly better assist percentage and *far* worse turnover percentage. The gap between Paul and everyone else when it comes to assisting teammates without turning the ball over is enormous.)

Subjective accolades? Paul is a 4-time first team all-NBA, plus 5 times on the second team and another third team, for a total of 10 all-NBA appearances. 11-time All-Star, 7-times first team all-defense (plus 2 second team appearances). Five top-5 finishes in MVP balloting; 9 top-10 finishes; 10 top-15 finishes.

The advanced stats tell us Chris Paul is one of the very best players ever. Traditional stats agree. Contemporaneous subjective assessments agree.

This really isn’t a situation in which his WS/48 numbers are some kind of fluke that can be disregarded as an indication of a flawed metric. He is perhaps the best point guard ever to play the game, elite on both sides of the ball. And that assessment holds whether you prefer advanced stats, traditional stats, or subjective assessments and accolades like all-NBA and All-Star honors.

Paul isn’t just a pretty-good-for-a-long-time guy, either: He has a very high peak. His top-5 seasons in Win Shares are better than Kobe Bryant’s top 5, better than Magic Johnson’s, better than even Larry Bird’s.

Nobody is suggesting using WS/48 alone to decide the best 10. I think you may have missed the point of that post.

Yep. Exactly. Nobody thinks WS/48 is the sole or ultimate determination of a player’s accomplishments or abilities. I included this rate stat alongside total WS count (and VORP) to provide another dimension on which we can assess the players under discussion; it can help illuminate the greatness of some players whose raw totals are limited by shorter (or in-progress) careers. Some people value a high peak over sustained Very Goodness (I tend towards this view) and WS/48 is one way of highlighting some players who may have been truly excellent over a shorter period of time. (It isn’t the best way, but it was considerably easier than, say, calculating 5-year-peok WS totals.)

[Steven43]

Paul isn’t just a pretty-good-for-a-long-time guy, either: He has a very high peak. His top-5 seasons in Win Shares are better than Kobe Bryant’s top 5, better than Magic Johnson’s, better than even Larry Bird’s.

This point is exactly the reason that Win Shares have limited usefulness in separating the very best players in history from one another. There isn’t a basketball historian alive who would try to argue that Chris Paul is on the same level historically as Larry Bird and Magic Johnson (or Kobe Bryant, for that matter).

Still, you make a pretty good case that Paul might well be among the Top 25-30 players of all time.