DraymondGold wrote:Voting PostNow we get to the next tier. Here are some of the likely candidates, ordered chronologically:
-Wade, Kobe, Dirk, Chris Paul, Durant, Kawhi, Giannis (all of these players except Dirk/Paul were listed in the last thread, and Dirk's starting to get mentioned in this thread)
There's a few more candidates who may sneak into the conversation, depending on what people value (Nash, Harden, Shai), although I expect a few others to go before they really get steam. For Nash, people presumably have concerns about the defense. For Harden, presumably it's scalability/resilience. For Shai, presumably it's sample size / resilience. Shai's regular season suggests he may be truly special, and a similar regular,ar season with a slightly better playoffs next year (with a bit more experience to help empower that) might have us retroactively raising 2025 Shai. At the same time, plenty of people have had great single regular seasons without ever sustaining longer runs, while others never put together a dominant playoff run. When evaluating players, I'd rather underestimate them to start and wait for the larger sample, than preemptively reward them for what ends up being something unsustainable/noisy.
To get a rough estimate of value,
let's look at EPM (looking at volume rather than rate using their volume stat, Estimated Wins). I’ve also looked at nbaRAPM across a range of sample sizes, AuPM, On/off, and a little WOWY — there’s broad trends for some of these guys, and some disagreement. As we get further down the list, the players get closer together, so the error bars will eventually make it hard for any player to clearly stand out. I do still think there’s some trends in the impact data though. Let’s focus on EPM, since it’s measured to be the most accurate stat in the modern era, although other stats show somewhat similar trends.
3-year RS Average EPM Estimated Wins:
13-15 Paul +17.9 [08-10 Paul +18.6]
06-08 Kobe +17.7 [07-09 +16.9]
15-17 Kawhi +16
06-09 Wade +15.7 (skipping 08 as it’s an injury year)
16-18 Durant +15.2 [14-16 +15.8]
19-21 Giannis +14 [15.2 in 82-game pace]
06-08 Nash +14.8
10-12 Dirk +13.3 [05-07 +19]
3-year PS Average EPM per Game:
16-19 Kawhi 7.3 (6.6 accounting for missed games)
14-16 Paul +5.9 (5.2 accounting for missed games)
16-18 Durant +4.8 (4.6 accounting for missed games)
= 09-12 Dirk +4.8
~= 08-10 Kobe 4.7
= 19-21 Giannis 4.7 (4.4 accounting for missed games)
05-07 Nash +2.8
05-07 Wade 2.4
[I don't have access to playoff Estimated Wins, so have averaged and converted to per game to get a volume stat for the playoffs]
So looking at each Player’s impact rankings: In EPM (and across other impact metrics like multi-year nbaRAPM), Chris Paul seems like the clear most valuable player in the regular season, and longer samples seem to favor Clippers Chris Paul specifically. Paul is still among the more impactful in the playoffs when playing, but some players are comparable. Paul also gets injured in numerous playoffs, and people may have varying scalability/resilience concerns.
Kawhi seems like the most impactful playoff player during his peak (noting that small playoff samples have more noise), although like Paul he suffers from untimely playoff injuries. His regular season is still impactful, among the top few.
Kobe surprisingly has the 2nd most Estimated Wins in the regular season and is right in the middle of this group in the playoffs, despite his reputation by for being relatively poorer in impact metrics. Looking at EPM per 100, Kobe’s closer to last. Turns out Kobe has quite the season volume advantage over the competition — he plays 82 games in both 08/09, averaged 79 games played across 06-10, and did so at 39 minutes per game (helped by a slightly slower pace). So a lesser per possession impact is partially compensated by playing more possessions per game and particularly being more durable each season. That health also shows up in the playoffs, where Paul and Kawhi and others struggle.
Wade looks better than Kobe per possession, but worse than Kobe per season (as above, Wade‘s impact advantage in rate stats are less than Kobe’s durability/volume advantage) and Wade looks last in multi-year playoff runs. Now Wade has the 2nd best playoff run of these players in 06 behind 17 Kawhi, but he also looks much worse in other (albeit quite small and occasionally injured) samples. In multi-year runs this pulls him down. Wade has a steeper decline outside short sample peaks. This is presumably partially driven by injuries, but also raises a concern that his best playoff runs in 06/09 are boosted a touch by small sample noise, or whether he could scale his impact with more talented teammates.
Durant isn’t the most impactful, but it also seems like Durant > Dirk and Giannis in this stat. Arguments to take him higher would probably focus on scalability — he had great playoff impact on arguably the best team ever. Arguments against him would say the situation was easier as teams focused on Curry, and that Durant didn’t show as much resilience in other scenarios. Durant seems more durable than Paul, Kawhi, Wade, or Giannis (he doesn’t seem to wear down under long playoff runs or over successive years), but still has more durability compared to Dirk and Kobe.
Giannis has less impact than one would expect given his reputation. He is partially held back but having his playoff peak and regular season peak in different years (which does not seem to be caused by injury or dramatic teammate changes like Curry, and regardless Giannis doesn’t have quite as much of an impact advantage relative to the completion like Curry). I also have resilience and scalability concerns with him, for reasons I described in an earlier post.
Dirk is also held back by having a regular season and playoff peak in different years. His run in 05-07 would be the best regular season run of anyone here, but if we force the run to include 2011, he’s last in the regular season. This seems like a classic case of player evolution (I would argue like Giannis). When a player is younger, they have more athletic stanina and motor to try in the regular season. As they get older, they develop more counters and gain more experience, and thus get more resilient in the playoffs, at the cost of regular season effort. Different players peak at different points in these arcs, and this development can make it interesting to choose a single stretch as their peak. I agree with the data that Dirk was more impactful in the regular season while he was younger, and better in the playoffs as he was older. I think this evolution to maintain impact over the years helps give him one of the better primes of this group, and gives him one of the better careers, but I do think it holds him back a bit with his peak.
One theme when comparing this tier: most player has great impact in some sample/stat, but are inconsistent in other scenarios. -Health: Paul, Kawhi, Wade, Giannis, and Durant all have injuries to differing extents. Wade has general durability issues throughout his prime. Paul, older Kawhi (starting in late 2017), and Giannis seem to wear down over longer playoff runs. Durant doesn’t wear down during his peak, but is still somewhat susceptible to month-scale injuries.
-Stamina/motor: Kawhi, Giannis, and Dirk seem to have their regular season and playoff peak in different seasons, in part due to stamina/motor. One of my bigger issues for Kawhi is his lack of stamina; his best regular season is 2016, while his best playoffs ignoring injury is probably 2017. I do think there’s an offensive development and shift from speed to strength that helps unlock the playoff impact in 2017, but I also think a lack of motor means he can’t combine his defensive and offensive peak at once, and I do think some coasting is required for the playoff rise. Giannis lost some of his athletic motor of 2019/2020 as he developed the experience and counters that helped his 21/22 runs. Dirk had a similar development to Giannis, albeit with a larger separation between his regular season and playoff peak.
Kobe doesn’t have as many health issues, or have as much of a separated regular season and playoff peak due to motor/development. However, I do still see a lack of motor limiting his regular season defense given the offensive load and the lack of coasting/load management. It would have been interesting to see if a more load managed Kobe had more in the tank defensively, or looked better per possession.
I think my top group will be Chris Paul, Kawhi, Durant, and Kobe. They’re roughly the most impactful according to EPM (again with Kobe making up a per possession disadvantage with more minutes/games played). I like much of this group from a scalability / team building perspective compared to say Wade or Giannis or Dirk. Wade’s lack of shooting and ball dominance makes it hard to maximize offensive impact with another ball dominant star. Giannis’ insistence to play on-ball rather than focus on more traditional big skills (screens, rolls, lobs, midrange counters) has made his chemistry with Lillard a bit disappointing. Dirk is pretty great offensively, but is a bit limiting to a team defensively given his lack of impact at his position. Team-wise, Kawhi, Durant, and Kobe all were key players on teams that were clearly more impactful, which supports the idea that their more scalable playstyle can lead to more dominant top-end teams.
1. 2017 Kawhi LeonardOf these four guys, I could be swayed multiple ways. At the moment, I’m leaning 17 Kawhi. I do think Kawhi still has motor issues limiting how well he could combine his peak offense/defense and regular/post season, and I do think he was already showing some durability concerns (he had already missed 2nd round games before going down to Zaza). But I don’t think he was as injury-prone or load-management-reliant as he would become, and from a probabilistic perspective, he was just as likely to make it through in a 2019-style run in 2017 as he was going down to the ankle injury.
And when he wasn’t injured, he might be the best playoff player of the bunch, and the impact data seems to support that. An all-time playoff scoring threat, with shooting that could be catch and shoot off-ball, fantastic resilient midrange that could come in isolation and off the dribble, and some basket pressure. This three-level diverse scoring package, combined with active off-ball action and a robust handle, drove some incredible playoff offenses. His lack of playmaking and lesser activity off-ball meant he wasn’t quite the offensive player of KD or Paul or Kobe. But he was still highly impactful, and could fit well alongside other ball-dominant playmakers. And he was the best of the bunch defensively. Not quite 2016 or 2014 levels, with the heavier build and offensive load, but still a clear positive with his hands in isolation and in passing lanes, and with solid rebounding.
It’s a tough choice, and I could definitely be swayed otherwise. The treatment of the injury seems like the key question here — by some criteria, the injury would discount this year altogether. One could also argue that the regular season specifically had too little impact to be chosen here (the +1.6 on/off is pretty pathetic, even with bad shooting luck; although his 3-year on/off is +6 with an on rating of +11.5, which is more respectful, and his Goldstein RAPM is 7th in the league that season which isn’t bad). One could argue that Kawhi’s defense had fallen off by 2017, or that his lack of motor/IQ/playmaking justify picking someone else. But even so, he seems like the most valuable playoff performer when healthy, which I find compelling.
2. 2015 Chris Paul (> 2014 Chris Paul)
His regular season impact metrics are the clear best of this group. He’s one of the best playmakers and passers ever — with a massive advantage in playmaking over Wade, Kobe, Dirk, Durant, Kawhi, and Giannis. Then we add that Paul had strong playoff scoring (22 Pts/75 at +6 rTS% in 2014; +23 Pts/75 at +10 rTS% in 2015), with near all-time midrange and positive 3 point spacing, and give him the best IQ of this group, and Paul looks to be contending for the best offensive player here ( I’d hear arguments for Kobe/Dirk/Durant too). Then add in that Paul is more consistent in his defensive value than Kobe and Durant, and it becomes clear why he seems like the most impactful player of the bunch.
I do think he doesn’t level up his game as much as other players here in the playoffs, but it’s the absolute value (not the relative change) that’s important. He starts at a higher place in the regular season than anyone here, so take a little off the top from him or give a little on the top to someone else… and Paul’s still up there competing to be the most impactful.
I have health concerns for him, like Kawhi. For greatest peaks, I’m most interested in figuring out which player was the best at their best, which gives some reasoning to down weighting peaks (especially ones from unlikely/freak accidents). At the same time, if players show consistent injury concerns in the playoffs e.g. from wearing down under the strain, then that is a concern. With Paul, he was reasonably healthy in the 2014 playoffs and performed great (Thinking Basketball’s recent podcast talked about his impressive defense on both Curry and Durant). In 2015, I think he was a little better overall. Although he did go down in the playoffs, he also played 82 games, which is unusual for a lot of these guys — a bit more load management from a bit more forward-looking coach than Doc Rivers might have helped him remain healthy for a full playoff run.
HM: 2008 and 2009 Kobe, 2016 and 2017 Durant.
Open to arguments for both players and somewhat to others as well.
3. 2008 Kobe Bryant (> 2009)
As above, this seems like the tier where limited stamina/motor cause players’ offensive and defensive peaks to not be simultaneous (Kobe, Kawhi), players’ regular season and postseason peaks to not be simultaneous (Dirk, Giannis), or durability causes players to wear down over continuous runs (Wade, Paul, Kawhi, Giannis). The players that don’t have these issues as severely seem slightly less impactful in a plus minus sense per possession (Kobe, Durant).
With Kobe, I do think his per possession (particularly regular season) impact was lesser than some of his competitors, particularly because he seemed to be a neutral defender on average in the regular season. But one thing this project has made me appreciate is Kobe’s durability and volume advantage compared to his contemporaries. He actually played far more possessions per season, primarily due to his high percentage of games played (although he also had a respectably high minute load). In peak seasons, this games-played volume advantage seems to be greater than his per-possession impact disadvantage, to the extent where something like his total EPM volume over peak seasons looks to be among the best in this group.
Now we’re talking about a player who’s a neutral-ish defender on average over the regular season. The recent Thinking Basketball peaks podcast talked about how Kobe would seem like an outright negative defender during an unimportant regular season game against middling teams, then then it up for important matchups, then turn it up more for the playoffs. Now I’m not saying he’s the best defender of the group in 08/09. But I do think that change in defensive value based on effort is enough to boost him up a bit if he had more stamina.
In other words, I think if he had more stamina to play full-effort defense, I see that as enough of a boost to shrink some of his per-possession impact disadvantage. And while he didn’t have the motor/stamina to maintain such effort while playing 80 regular season games per season. And when the competition didn’t have the stamina to play so many games, used load management to help boost their per possession impact, or didn’t have the durability and would wear down under longer runs like Kobe had, then it feels wrong to punish Kobe. All else being roughly equal, I’d rather someone be available to play but coast defensively in less important regular season games than skip games or wear down and not be able to play.
Kobe’s at the intersection of being neutral to positive in both resilience (he retains value as the competition gets better) and scalability (he retains value as his teammates get better). That’s pretty important from a championship odds perspective. I’ve already touched on his defensive improvements in the playoffs (which admittedly come partially from flawed stamina than having a resilient defensive playstyle). But he’s also one of the most resilient shot makers ever. This combines to show some signal for resilience when comparing regular season and playoff data. Taking 10-year on/off in the RS, PS:
These are noisy numbers, and for long prime durations. Rotations change on the playoffs, and minutes go up, so I wouldn’t take the raw change too literally. Context like playing through injury or missing a playoffs in a certain year can also cause things to change. But relative to the competition, Kobe’s clearly in the upper echelon of playoff improvers.
In terms of ceiling raising, Kobe’s showed some great chemistry with his all star teammates in all of his championship runs. He was always too focused on scoring, but his blend of on-ball and off-ball action allows him to fit next to other ball dominant players. And he was always fairly cerebral, capable of good passes, and running more complex offensive schemes. Impressively, the 2009 Lakers are a top 20/25 team where, which is very respectable to do in your multi-year peak. I see 2008 as a slightly better version of 2009 Kobe, even if the 2009 team achieved more. So I went with 2008, although I see them as quite similar players, and I’m open to the argument that 2009 is better in the playoffs if anyone wants to make it.
4. 2016 Kevin Durant (> 2017 > 2014)
My post is getting long, so I’ll be briefer. Not quite an impact king, but better regular season performance than Dirk around his postseason peak, and more consistent playoffs than Wade (who had great runs in 06/09 but struggled in adjacent years, potentially due to health/durability issues) or Giannis. Not as great of a floor raiser due to limitations on his handle/playmaking, but definitely the most scalable of the remaining players in this tier. Fits great with surrounding talent offensively with his all-time shooting, spacing, hybrid on/off-ball scoring style, and defensive versatility. I like his ceiling raising offensively more than Wade, and his defensive versatility and team building more than Dirk. He is somewhat injury prone, but I do think he has better stamina/motor than some of the competition, and I don’t think he’s as injury prone under high load as a few other players either.
His team level performance is great
Two teams in the top 10. Of course, those teams were led by Curry, but putting that aside, he had the 2016 Thunder in the top 30.
