I
posted earlier about how LeBron's peak ranks in terms of WS/48 and PER. In both cases, he has the best, second-best, third-best, and fourth-best season of any player still eligible for the list. I also ranked the best seasons by MVP shares, and found that LeBron has the #1 season overall, three of the top 10, and four of the top 20. Bird is the only other player still eligible with more than one season in the top 20 by this metric. Especially with the discussion of alternatives to WS and PER going around, I wanted to look at the top seasons by from one more perspective, this time by RAPM.
I'll be using the RAPM chronology sheet Doc MJ put together
here, which scales to make values comparable across years. That sheet ends in 2012, cutting off one GOAT-level peak year from Bron and another league-leading year, so I copied the numbers from J.E.'s site into excel and calculated the number of standard deviations above average for those years, too. I'm not 100% sure the numbers I used are calculated the exact same way as the source data, so I've denoted the years as such. I'm also a little distrustful of the 97-98 data; it seems to have a higher variance (Mookie Blaylock is at +3.71 SDs, the 10th-highest single-season score on the list), and being the first year in the data, I'm not sure if there's any prior-informed component helping its accuracy, so I've noted those years as well. Here are the top seasons of all time by this metric:
Code: Select all
LeBron 09-10 4.62
Shaq 97-98 4.31 (see above)
Garnett 03-04 4.26
LeBron 08-09 4.15
Mourning 97-98 4.00 (see above)
LeBron 12-13 3.97 (see above)
Wade 09-10 3.96
Garnett 02-03 3.89
Dirk 10-11 3.87
Duncan 06-07 3.83(For the curious, the other +3 seasons in the 13 and 14 seasons I calculated are 13 Paul +3.22, 13 Durant +3.14, and 14 LeBron +3.01)
Once again, we see a similar story. LeBron has three of the top 6 years of the 98-14 era, and if we don't trust the 97-98 data (Doc, if you can clarify where exactly this comes from and if it is as trustworthy as future years with more of a prior, I'd appreciate that), that moves up to 3 of the top 4 years over the 99-14 period. This includes the vast majority of Shaq, Duncan, and KG's careers, (as well as other elite players like Dirk, Wade, Paul, and Nash), and LeBron just blows them away.
LeBron's best season is .31 standard deviations better than Shaq's best (1.02 without 97-98), .36 better than KG's, and .79 better than Duncan's.
LeBron's second-best season is .45 standard deviations better than Shaq's second-best (.90 without 97-98), .26 better than KG's, and 1.06 better than Duncan's.
LeBron's third-best season is .72 standard deviations better than Shaq's third-best (.75 without 97-98), .37 better than KG's, and .89 better than Duncan's.
LeBron's fourth-best season is .04 standard deviations worse than Shaq's (.16 better without 97-98), .13 worse than KG's, and .37 better than Duncan's.
LeBron's fifth-best season is .01 standard deviations worse than Shaq's (.08 better without 97-98), .21 worse than KG's, and .26 better than Duncan's.
The sum of LeBron's 5 best seasons is 18.93, compared to 18.28 for Garnett, 17.41 for Shaq (16.03 without 97-98), and 15.57 for Duncan.
A few more things to keep in mind:
-Each standard deviation is worth roughly 2.5 points, and we're talking about the most extreme top end of the curve. ElGee has posted his championship odds curve a few times, but what's important is just to remember that it's hyperlinear, with the slope continually increasing as you get higher and higher. The extra point or two that peak and near-peak LeBron is worth compared to even (near)-peak Shaq/KG/Duncan comes at the steepest part of the graph, where even every half point makes a huge difference in championship odds.
-RAPM can be influenced by randomness a fair amount. You'll often see results that don't make intuitive sense, and that's partly because there's a lot of chance involved. Sometimes you play well and your teammates just happen to miss shots when you're on the floor. Since it evaluates the performance of your whole team rather than just your individual stats, an individual player has a much lower degree of control over their results, and it's harder to dominate. Even if someone was far and away the best of their era, you wouldn't expect to see them go down the line and have every top season like LeBron does with box score statistics.
-RAPM is purely per-minute or per-possession. That means there's no bonus given for staying healthy or for playing lots of minutes, which LeBron has done every year of his career, unlike Shaq and Duncan. To go back to ElGee's championship odds graph, this is the very steepest part at the extreme rightmost part of the single highest line on the graph, and for most of the competition seasons, not only are you moving left in terms of impact, you're moving to a lower line in terms of health as well.
-The benefits of versatility aren't totally accounted for in RAPM. Suppose you're a +6 player who can only play one position and one role. The team already has, or can sign, a +2 player who plays the same role as you, but because that's a poor fit, they use the same resources to get a +1 guy who complements you. You then play with the +1 guy, get a +7 result, and the equation looks at your lineups and says: you took +1 surroundings to +7; you're a +6 guy. What it can't measure is that really those surroundings could've been +2, and really your impact was more like +5, taking +2 to +7. All the combinations that don't work with you, but which obviously enough don't work that the GM or coach avoids them, don't end up hurting your RAPM score, even if they do hurt your team.
This versatility is a point in Duncan's favor, who can play alongside a PF or a C, a Diaw or a Splitter. It's a point in Bird's favor, as his size and shooting ability lets them match up in unique ways on offense and play multiple positions. But nobody exemplifies versatility like LeBron, who can play initiate your offense as a PG on one end and help anchor your defense as a PF on the other; who's got arguably the best post game in the league and who hits spot-up threes as well as Curry or Korver; who can lead a 66-win team as the primary ball-handler and then do it again four years later alongside one of the most ball-dominant guards the league has ever seen; who is the embodiment of the wave of positionless basketball that is taking over the league.
-RAPM, like box-score statistics, has no accounting for clutch performance. It only accounts for point differential, and not a player's ability to affect when and where those points come. Usually, that assumption is a safe one. LeBron's penetration might be the single most potent self-created offensive option ever seen, though, and while going to it every time down might wear down his body too much, it allows him to absolutely dominate in the clutch. In several different seasons, we've seen Lebron-led teams play at a level roughly equivalent to a +35-45 MOV team in clutch seasons, with LeBron averaging a statline in the neighborhood of 50-15-10 on .600 TS% per 48 clutch minutes. (These numbers are so unbelievable I understand if you want to look them up yourself. Here's
08-09,
09-10, and
12-13. That's a +45, +37, and +37 team net rating per 48 minutes, respectively.
). The numbers simply defy comprehension, but they're yet another indication that a metric like RAPM, or like PER and WS/48 that just looks at how a player impacts point differential is selling LeBron short. He dominates the competition there, and yet he's probably even better than the metric makes him seem.
The one knock about LeBron is in terms of longevity. And yet I get the feeling that the problem is not that he hasn't accomplished enough in his career, but just that he's done it all before age 29, and we feel like there's even more to come. If this was the portfolio of a player whose career was already over and we were looking back on it, though, I don't think we're half as concerned about it. Michael Jordan played 11 full seasons with the Bulls. He won #1 easily, and nobody was even thinking of anyone besides Russell and a couple of Kareem voters over him. Duncan and Shaq didn't get any attention at all. Why? Because Jordan reached a level of dominance over the league where he changed the entire landscape, where the entire championship picture went through him, and no other player was at close to his level. Jordan maintained that level for 8 years or so with three more just below it, compared to about 6 for LeBron with four more just below it, but there's really not a huge difference there. It's enough to knock LeBron out of consideration for #1, sure, but for #5, he should be right there.
Magic and Bird finished at #4 and #6 in 2011, ahead of Shaq and Hakeem. Duncan has added a few post-prime seasons since them, but that was enough for them to surpass the entire careers of the other two. And they've got roughly the same number of total career MP as Lebron, fewer prime years, and a far less dominant peak. They alternated years winning MVPs and generally seemed to be each other's equals for most of their careers. LeBron took players like '09 Wade, Paul, and Kobe, or '13 Durant, and won the MVP over them with 97 - 99% of the vote. He combines Magic and Bird's offensive genius and incredible passing with the efficient volume scoring of a Durant or a Barkley and the defensive ability of a Pippen or Iguodala. I've got LeBron comfortably ahead of Magic and Bird, and they were good enough to be ahead of every other player available last time. I could see an argument being made to move Shaq or Duncan above those two if you value longevity, but nobody would make the argument to move them above Jordan for that reason, and that's closer to the kind of company LeBron is in right now. Nobody else has had a sustained period of unquestioned dominance over the league for more than a couple years, and LeBron's working on 5 or 6 and counting.
