Top 25 peaks of the 2001-25: #9-#10 Spots

[jalengreen]

Yeah my entire point is that saying that something could have changed things is not the same as arguing that it *did* change the outcome. Some would say that there's enough randomness in play for you to swing a 15 point loss to a win just by re-simulating a game without changing any pre-game variables at all. But that's not meaningful in any way. A position of confidence that the outcome changes *is* meaningful and does suggest a very lofty view of Draymond Green's peak impact.

[lessthanjake]

I think it’s worth providing some contrast between Draymond and a different sidekick that I’m much more confident would’ve done well leading a team.

I noted in prior posts that from 2015-2025 minus 2020, the Warriors had just a -4.21 net rating with Draymond on and Steph and Durant off, in minutes where the other team actually had 4 or 5 starters on the floor (it’s -2.68 if we add playoffs minutes in). Even in his peak period of 2015-2017, the Warriors had a -2.97 net rating in those minutes. Basically, in minutes where he was leading the team against the opposing team’s best players, the results were not good. And that’s a worrisome signal when the question is how he’d do if he actually was leading a team all the time, since most of your minutes as your team’s best player will be against these sorts of starter-heavy lineups. Meanwhile, being able to dominate opposing teams’ bench units is not something that will come up much when you’re your team’s best player, since teams won’t often put lots of those guys out against you.

In contrast to this, over the course of all of Manu Ginobili’s seasons playing with Duncan, the Spurs had a +8.60 net rating with Ginobili on, Duncan off, and the opposing team having 4 or 5 starters on the court. It was +5.15 in the playoffs (I want to calculate with RS+PS combined, but PBPstats keeps crashing when I try it). Ginobili was at his best as a player in 2005-2008, and in those years the Spurs had a +9.64 net rating with Ginobili out there without Duncan and facing starter-heavy units. So the signal here makes me confident that Ginobili can lead a team to be very successful against the types of lineups he’d typically face as the best player on a team.

Granted, I wouldn’t look at these and just simplistically say that the difference between Ginobili’s numbers and Draymond’s numbers represents a precise difference in their impact in these scenarios. The samples here aren’t super small, but they’re still subject to some noise, and while the Warriors had some good supporting players over the years, I do think the Spurs supporting cast generally was better. Even so, I find the discrepancy between these two in this data to be pretty compelling. And I think that that discrepancy fits with what my own expectations would be, since Draymond feels much more like a guy who is better as a complementary piece than as the main guy, while Ginobili feels like someone who could be either.

[One_and_Done]

Spoiler:

1. Kawhi (2017)
2. KD (2014 I guess?)
3. SGA (2025)
4. TBD

My vote for Kawhi has been discussed before at length. At his peak he was aptly nicknamed Robo-Jordan, because he was a better version of MJ. His two way impact was insane, and I can’t see any logic to punishing him for a random injury in the WCFs (when we are voting in guys for seasons who didn’t even see them progress that far). No criteria which rewards a guy for playing fewer games makes much sense to me. If he’d missed large chunks of the season maybe, but he played enough.

Kawhi also stands out as a guy who has no real weaknesses. He gives you a better version of Jordan on offense, and a better version of Pippen on defence. That’s an insane combination. I saw someone suggest Kobe has better numbers than Kawhi. No, he doesn’t, nor is it even close. Here’s 17 Kawhi v.s 09 Kobe (who was advanced as the best Kobe year by that person).

Kawhi RS 17: 39/9/5 per 100, on 610 TS%, while playing DPOY level D
Kobe RS 09: 38/7/7 per 100, on 561 TS%, while being an ok defensive player at best.

Kawhi PS 17: 40/12/5 per 100, on 672 TS%, while continuing to play at a DPOY level on D
Kobe PS 09: 39/7/7 per 100, on 564 TS%, while being at best solid on D

Kobe has no argument whatever against Kawhi. Kawhi was clearly more impactful on both ends. Then you look at Kobe’s terrible floor raising, often detrimental attributes, and we’re a long way off Kobe getting a look in.

The next three names are tougher. I considered a number of candidates, particularly KD, Luka, SGA, Harden, Nash, CP3, T-Mac, Dirk, AD and J.Butler. I’ve ultimately gone with KD and SGA for 2 and 3, but will need to mull on vote #4 for a bit longer. I will likely go with Dirk just because he has more traction, but I could easily see myself voting for CP3, Nash, Harden, Luka, or Butler.

We have ample evidence KD gave a huge impact on winning, even for non-stacked teams.

In 2014 for example, the Thunder were 25-11 in the games Westbrook missed, thanks to KD.

His Brooklyn time is a bit of a mess to assess, because of all the stuff that happened involving availability of guys, but we can see in 21 the team was 23-12 with KD, and only 25-24 without him. Similarly, the Nets in 22 were 36-19 with him, and only 8-19 without him. We also saw KD carry the Nets in the 21 playoffs, almost past the Bucks, with Kyrie and Harden both going down with injuries. If KDs toe isn’t on the line, the Nets likely win the championship this year largely on the back of KD.

In KD’s Phoenix tenure, despite being past his prime, the win-loss still holds up well for KD. From 23 to 25 the Suns were 85-60 with him, and 15-30 without him. The contrast was stark.

But hey, some computer formulas that, by their nature, are unreliable at accurately measuring value don’t agree, so I guess forget all that other stuff.

As for SGA last season, he didn’t have the best playoffs, but he also had a historic RS and led a very injured team to 68 wins then a title. I’m satisfied he’s done enough to get the nod.

I’ll follow the debate as it unfolds before settling on my #4.

Put Kobe at 4

At this point it’s somewhat redundant to explain how Kobe is overrated on offense, but I’ll give it another crack.

Let’s look at just 2009 CP3 v.s 09 Kobe as an example. Paul’s per 100 numbers are particularly impressive this year; 32/8/16 on 599 TS%, with a 124 Ortg, and justifiably finishing 6th in DPOY voting

Now, if we took a simplistic approach, and assumed assists are worth 2.4 points each, then prima facie CP3 added 70+ pp100 to his team, compared to 54pp100 for Kobe. This is a simplistic approach obviously, but it’s intended to try and demonstrate how much more there is to generating offense than “ppg!”.

In reality CP3 was generating even more points of course, because of things like hockey assists, and because he’s scoring at better efficiency, or because he’s putting his team mates in their sweet spots so they can shoot at better efficiency, etc, but even by a flat numbers comp was CP3 waaayyy ahead. CP3 even rebounds more per 100 than Kobe in 09. He looks better than Kobe in everything really. The reason CP3’s teams have better offensive ratings is because setting guys up for easy and efficient baskets is often more important than actually scoring at so-so efficiency.

[Top10alltime]

Spoiler:

1. Kawhi (2017)
2. KD (2014 I guess?)
3. SGA (2025)
4. TBD

My vote for Kawhi has been discussed before at length. At his peak he was aptly nicknamed Robo-Jordan, because he was a better version of MJ. His two way impact was insane, and I can’t see any logic to punishing him for a random injury in the WCFs (when we are voting in guys for seasons who didn’t even see them progress that far). No criteria which rewards a guy for playing fewer games makes much sense to me. If he’d missed large chunks of the season maybe, but he played enough.

Kawhi also stands out as a guy who has no real weaknesses. He gives you a better version of Jordan on offense, and a better version of Pippen on defence. That’s an insane combination. I saw someone suggest Kobe has better numbers than Kawhi. No, he doesn’t, nor is it even close. Here’s 17 Kawhi v.s 09 Kobe (who was advanced as the best Kobe year by that person).

Kawhi RS 17: 39/9/5 per 100, on 610 TS%, while playing DPOY level D
Kobe RS 09: 38/7/7 per 100, on 561 TS%, while being an ok defensive player at best.

Kawhi PS 17: 40/12/5 per 100, on 672 TS%, while continuing to play at a DPOY level on D
Kobe PS 09: 39/7/7 per 100, on 564 TS%, while being at best solid on D

Kobe has no argument whatever against Kawhi. Kawhi was clearly more impactful on both ends. Then you look at Kobe’s terrible floor raising, often detrimental attributes, and we’re a long way off Kobe getting a look in.

The next three names are tougher. I considered a number of candidates, particularly KD, Luka, SGA, Harden, Nash, CP3, T-Mac, Dirk, AD and J.Butler. I’ve ultimately gone with KD and SGA for 2 and 3, but will need to mull on vote #4 for a bit longer. I will likely go with Dirk just because he has more traction, but I could easily see myself voting for CP3, Nash, Harden, Luka, or Butler.

We have ample evidence KD gave a huge impact on winning, even for non-stacked teams.

In 2014 for example, the Thunder were 25-11 in the games Westbrook missed, thanks to KD.

His Brooklyn time is a bit of a mess to assess, because of all the stuff that happened involving availability of guys, but we can see in 21 the team was 23-12 with KD, and only 25-24 without him. Similarly, the Nets in 22 were 36-19 with him, and only 8-19 without him. We also saw KD carry the Nets in the 21 playoffs, almost past the Bucks, with Kyrie and Harden both going down with injuries. If KDs toe isn’t on the line, the Nets likely win the championship this year largely on the back of KD.

In KD’s Phoenix tenure, despite being past his prime, the win-loss still holds up well for KD. From 23 to 25 the Suns were 85-60 with him, and 15-30 without him. The contrast was stark.

But hey, some computer formulas that, by their nature, are unreliable at accurately measuring value don’t agree, so I guess forget all that other stuff.

As for SGA last season, he didn’t have the best playoffs, but he also had a historic RS and led a very injured team to 68 wins then a title. I’m satisfied he’s done enough to get the nod.

I’ll follow the debate as it unfolds before settling on my #4.

Put Kobe at 4

At this point it’s somewhat redundant to explain how Kobe is overrated on offense, but I’ll give it another crack.

Let’s look at just 2009 CP3 v.s 09 Kobe as an example. Paul’s per 100 numbers are particularly impressive this year; 32/8/16 on 599 TS%, with a 124 Ortg, and justifiably finishing 6th in DPOY voting

Now, if we took a simplistic approach, and assumed assists are worth 2.4 points each, then prima facie CP3 added 70+ pp100 to his team, compared to 54pp100 for Kobe. This is a simplistic approach obviously, but it’s intended to try and demonstrate how much more there is to generating offense than “ppg!”.

In reality CP3 was generating even more points of course, because of things like hockey assists, and because he’s scoring at better efficiency, or because he’s putting his team mates in their sweet spots so they can shoot at better efficiency, etc, but even by a flat numbers comp was CP3 waaayyy ahead. CP3 even rebounds more per 100 than Kobe in 09. He looks better than Kobe in everything really. The reason CP3’s teams have better offensive ratings is because setting guys up for easy and efficient baskets is often more important than actually scoring at so-so efficiency.

[70sFan]

Spoiler:

1. Kawhi (2017)
2. KD (2014 I guess?)
3. SGA (2025)
4. TBD

My vote for Kawhi has been discussed before at length. At his peak he was aptly nicknamed Robo-Jordan, because he was a better version of MJ. His two way impact was insane, and I can’t see any logic to punishing him for a random injury in the WCFs (when we are voting in guys for seasons who didn’t even see them progress that far). No criteria which rewards a guy for playing fewer games makes much sense to me. If he’d missed large chunks of the season maybe, but he played enough.

Kawhi also stands out as a guy who has no real weaknesses. He gives you a better version of Jordan on offense, and a better version of Pippen on defence. That’s an insane combination. I saw someone suggest Kobe has better numbers than Kawhi. No, he doesn’t, nor is it even close. Here’s 17 Kawhi v.s 09 Kobe (who was advanced as the best Kobe year by that person).

Kawhi RS 17: 39/9/5 per 100, on 610 TS%, while playing DPOY level D
Kobe RS 09: 38/7/7 per 100, on 561 TS%, while being an ok defensive player at best.

Kawhi PS 17: 40/12/5 per 100, on 672 TS%, while continuing to play at a DPOY level on D
Kobe PS 09: 39/7/7 per 100, on 564 TS%, while being at best solid on D

Kobe has no argument whatever against Kawhi. Kawhi was clearly more impactful on both ends. Then you look at Kobe’s terrible floor raising, often detrimental attributes, and we’re a long way off Kobe getting a look in.

The next three names are tougher. I considered a number of candidates, particularly KD, Luka, SGA, Harden, Nash, CP3, T-Mac, Dirk, AD and J.Butler. I’ve ultimately gone with KD and SGA for 2 and 3, but will need to mull on vote #4 for a bit longer. I will likely go with Dirk just because he has more traction, but I could easily see myself voting for CP3, Nash, Harden, Luka, or Butler.

We have ample evidence KD gave a huge impact on winning, even for non-stacked teams.

In 2014 for example, the Thunder were 25-11 in the games Westbrook missed, thanks to KD.

His Brooklyn time is a bit of a mess to assess, because of all the stuff that happened involving availability of guys, but we can see in 21 the team was 23-12 with KD, and only 25-24 without him. Similarly, the Nets in 22 were 36-19 with him, and only 8-19 without him. We also saw KD carry the Nets in the 21 playoffs, almost past the Bucks, with Kyrie and Harden both going down with injuries. If KDs toe isn’t on the line, the Nets likely win the championship this year largely on the back of KD.

In KD’s Phoenix tenure, despite being past his prime, the win-loss still holds up well for KD. From 23 to 25 the Suns were 85-60 with him, and 15-30 without him. The contrast was stark.

But hey, some computer formulas that, by their nature, are unreliable at accurately measuring value don’t agree, so I guess forget all that other stuff.

As for SGA last season, he didn’t have the best playoffs, but he also had a historic RS and led a very injured team to 68 wins then a title. I’m satisfied he’s done enough to get the nod.

I’ll follow the debate as it unfolds before settling on my #4.

Put Kobe at 4

At this point it’s somewhat redundant to explain how Kobe is overrated on offense, but I’ll give it another crack.

Let’s look at just 2009 CP3 v.s 09 Kobe as an example. Paul’s per 100 numbers are particularly impressive this year; 32/8/16 on 599 TS%, with a 124 Ortg, and justifiably finishing 6th in DPOY voting

Now, if we took a simplistic approach, and assumed assists are worth 2.4 points each, then prima facie CP3 added 70+ pp100 to his team, compared to 54pp100 for Kobe. This is a simplistic approach obviously, but it’s intended to try and demonstrate how much more there is to generating offense than “ppg!”.

In reality CP3 was generating even more points of course, because of things like hockey assists, and because he’s scoring at better efficiency, or because he’s putting his team mates in their sweet spots so they can shoot at better efficiency, etc, but even by a flat numbers comp was CP3 waaayyy ahead. CP3 even rebounds more per 100 than Kobe in 09. He looks better than Kobe in everything really. The reason CP3’s teams have better offensive ratings is because setting guys up for easy and efficient baskets is often more important than actually scoring at so-so efficiency.

[One_and_Done]

I should have said 'the reason CP3's teams have generally had a higher Ortg', rather than in 09.

[70sFan]

I should have said 'the reason CP3's teams have generally had a higher Ortg', rather than in 09.

None of the CP3 Clippers teams have better ORtg than 2008 and 2009 Lakers.

[Caneman786]

What kinda bot farm is going on up there?

EDIT: I think I understand now. Very poor formatting!

[LA Bird]

Players aren’t individually tanking (they have their own future contracts to care about first and foremost), but I do agree that the level of effort and focus will tend to be lower, so those years will tend to undersell their potential impact. I wouldn’t say that makes it meaningless—it’s just significant context that we should keep in mind. And, even accounting for that context, it is still striking that the Warriors literally did worse in the games Draymond played than the games he didn’t play. You use the 2008 Heat as an example, but at least there the Heat had a -7.56 net rating in the games Wade played and a -12.06 net rating in the games Wade missed. So they actually did significantly worse in the games Wade missed. And they had a -6.40 net rating with Wade on the court and a -12.06 net rating with him off the court (weird that both the above-listed Wade-out net ratings are exactly the same, but they are). I’d say that’s a pretty different looking data point than what we have for Draymond. While I wouldn’t say that that truly accurately represents exactly how much impact Wade could have, it actually does still indicate him having a good bit of impact, while the data for Draymond does not.

You're still missing the point. Me bringing up 08 Wade isn't to compare his impact to Draymond's. It's to show how the numbers for tanking seasons are irrelevant. We just had Wade's peak voted in last round - how many people said anything at all about 08, a season right next to his two best years? Nobody. No one said anything about 08 Wade's impact numbers or lack of wins because no one cares about tanking seasons. And the reality is, if Draymond wasn't one of the most polarizing players with a whole swathes of people praying on his downfall (including some GSW Curry fans), nobody would care about his numbers on a tanking team either.

If you want to say that we also have a 1933-minute sample of time in the prior 5 years, in which Draymond was on the court without Steph or Durant, and the Warriors had a +2.54 net rating in those minutes, then I’d say that’s a very significant data point as well. We could even take a larger sample of 2015-2025 minus 2020, with an assortment of supporting casts, and the Warriors had a +2.56 net rating in 4332 minutes with Draymond on and Steph & Durant off. But a really important thing to realize about the Draymond minutes without Steph or Durant is that those net ratings are often being farmed against opposing benches. If we filtered that above-mentioned +2.56 number down to just minutes in which 4 or 5 opposing starters are on the court, suddenly the net rating with Draymond on and Steph/Durant off is -4.21 in 1506 minutes. And it’s still a -1.00 net rating in 2292 minutes if we filter it down to minutes in which at least 3 opposing starters are on the court.

Why are you using a 11 year sample in a peak project when it's obvious those later seasons are going to drag down his peak numbers? Are we going to be looking at 2025 lineup numbers too when discussing peak Westbrook's impact? And I don't see how 3 starters or fewer on Draymond's side against 4 starters or more for the opponent is somehow a fair comparison. For example, that would theoretically even include minutes with Draymond as the lone starter against all 5 opponent starters. I don't know what's the source for these stats so I can't check it myself but equalizing the number of starters seems like the obvious first step. Second is to separate Durant years from the ones before because he changed the team dynamic considerably. So, instead of an 11 year WOWY with varying number of starters, this is what you should look at:

15/16 Draymond + 3 starters (excluding Curry) vs 4 opponent starters
17/18/19 Draymond + 2 starters (excluding Curry/Durant) vs 3 opponent starters

I am guessing he fares pretty well in both which is why you didn't take this obvious approach instead.

This seems like pretty important context to those numbers, because if Draymond were actually leading a team then he’d almost certainly be spending a significantly higher proportion of his minutes against opposing starters than he did in the non-Steph/Durant minutes. His minutes against mostly bench players make up almost half the non-Steph/Durant minutes, while they’d probably be more like 20% or so of his minutes if he were leading a team. So this is good reason to believe that the net rating in those non-Steph/Durant minutes is skewed upwards from what it would look like if Draymond was leading a team.

If Draymond was actually leading a team, he would have two other stars in place of Curry/Durant, not an empty hole on the roster that takes up majority of the cap space. Many of his non-Curry/Durant minutes are played against mostly bench players because he was playing with mostly bench players himself.... which is what happens when you take out 2 of the 5 starters.

[jalengreen]

I think 2008 Wade is a pretty good counter in the context of this conversation. Would expect more people to bring that up if they cared that much about a non-peak version of a player on a tanking team.

And I'll be honest, I don't see much reason to care about the "2020 Warriors were worse with Draymond playing" data point. Peak Draymond is obviously not a negative impact player. So observing a negative impact signal on the 2020 Warriors is all the more reason for me to either not care about the signal or not care about the 2020 Warriors, frankly.

Though out of curiosity I looked it up and 2020 Draymond:

+0.3 EPM
+2.2 xRAPM
+0.47 LEBRON

So to be clear, more holistic signals do point to him being a positive on that dreadful team.

I'll admit that if I have a tanking team that I'd like to be elevated towards respectability as much as possible, I would not call Draymond Green's number to lead that charge. I also think that has absolutely nothing to do with championship odds (well, it makes him a better tanker which is good for championship odds, but ignoring that...). Like:

2020 Draymond

On: -7.1 , Off: -9.9 , Diff: +2.8

2008 Wade

On: -6.9 , Off: -12.1 , Diff: +5.2

How much more value really is there here, if you do want to take it at face value? Meh

Finally I'll say that the "starter state" is a really cool feature on pbpstats and it's neat to see the data, and it definitely looks very impressive for Ginobili. But RAPM and its derivatives already take account of who is on the floor in a more comprehensive way, and 29-year RAPM has Ginobili and Draymond at +7.1 and +6.9 respectively (9th and 12th all-time).

[Joao Saraiva]

I'll just vote Monday but... #1 will be Kobe.

[One_and_Done]

Another guys who needs more traction is Harden. He's not as good as Curry, but he's not that much worse either. On paper Harden blows Kobe away. There's plenty of evidence he gives teams more lift too.

[ReggiesKnicks]

I don't see any value in assessing 2020 Draymond Green as an example of what it may look like if he were on your team as your best player. The Reality is, Draymond Green is incredibly valuable in the NBA. Top 10-15 defender of all-time, elite passer for a PF/C, High BBIQ, Can push pace with his handle, good in half-court sets as a decision maker. And, the kicker of course, is that since some people love to zero-in on single seasons here since "what happened happened (I am looking at you 2006 Dwyane Wade voters), Draymond was a 39% 3P shooter and best player of a 73-win team during the NBA Finals.

You want to talk about players and their "NBA Finals Performance"? 17/10/6 while leading your team in efficiency, assists, rebounds, steals and blocks is rare and historically dominant.

[ReggiesKnicks]

Another guys who needs more traction is Harden. He's not as good as Curry, but he's not that much worse either. On paper Harden blows Kobe away. There's plenty of evidence he gives teams more lift too.

This is demonstrably false.

[ReggiesKnicks]

^^Very good point on Draymond ON Curry OFF lineups being largely against bench units and correcting for that. It's a pretty big deal to point out.

Also, the 2016 playoff sample is far too small to make any kind of definitive conclusions nor do I think the Warriors necessarily win the Portland series if Curry doesn't come back. Even with him in Game 4 and Game 5, they won by 7 points in OT and 4 points, respectively. They were -20 in those two games without Steph on the court. In 2025, we saw the Dubs without Curry win Game 1 at Minnesota and everyone started jumping on how the Warriors didn't need Curry. I pleaded caution and in fact predicted that the Wolves would still win the series after Game 1. And as we know the Wolves ended up sweeping four straight by 24, 5, 7, and 11 points. With small samples, we underestimate how dramatically the pendulum can swing in the other direction.

Hm? Minnesota immediately swung to being favored in the series despite losing Game 1 (an incredible rarity). And that was before it was clear that Curry wouldn't be returning. I just checked the 2024-25 discussion thread, only person here who said the Warriors didn't need Curry was EmpireFalls, who LTJ identified at the time as clearly trying to reverse jinx.

Yeah, it is even difficult to remember this given it was 4-5 months ago. Here is another simple example to show why Djoker is wrong.

Minnesota Odds to win series before game 1: -178
Minnesota Odds to win series after game 1: -210

For those who know nothing about gambling, a negative number represents how much you need to bet in order to win $100.

[lessthanjake]

Players aren’t individually tanking (they have their own future contracts to care about first and foremost), but I do agree that the level of effort and focus will tend to be lower, so those years will tend to undersell their potential impact. I wouldn’t say that makes it meaningless—it’s just significant context that we should keep in mind. And, even accounting for that context, it is still striking that the Warriors literally did worse in the games Draymond played than the games he didn’t play. You use the 2008 Heat as an example, but at least there the Heat had a -7.56 net rating in the games Wade played and a -12.06 net rating in the games Wade missed. So they actually did significantly worse in the games Wade missed. And they had a -6.40 net rating with Wade on the court and a -12.06 net rating with him off the court (weird that both the above-listed Wade-out net ratings are exactly the same, but they are). I’d say that’s a pretty different looking data point than what we have for Draymond. While I wouldn’t say that that truly accurately represents exactly how much impact Wade could have, it actually does still indicate him having a good bit of impact, while the data for Draymond does not.

You're still missing the point. Me bringing up 08 Wade isn't to compare his impact to Draymond's. It's to show how the numbers for tanking seasons are irrelevant. We just had Wade's peak voted in last round - how many people said anything at all about 08, a season right next to his two best years? Nobody. No one said anything about 08 Wade's impact numbers or lack of wins because no one cares about tanking seasons. And the reality is, if Draymond wasn't one of the most polarizing players with a whole swathes of people praying on his downfall (including some GSW Curry fans), nobody would care about his numbers on a tanking team either.

People don’t bring up 2008 Wade in a discussion about what Wade could do leading a team because we have a whole lot of other data points about what Wade could do leading a team—including a bunch of seasons in which he did so, one of which resulted in a title. With Draymond, we do not have a bunch of seasons of him leading a team, so the one we do have naturally takes on much more importance despite it being a tanking season for his team. A flawed piece of data can and will hold more importance when there is a dearth of other data. I am quite confident that if Wade had spent his entire career with LeBron except for 2008, then people would discuss 2008 a lot.

If you want to say that we also have a 1933-minute sample of time in the prior 5 years, in which Draymond was on the court without Steph or Durant, and the Warriors had a +2.54 net rating in those minutes, then I’d say that’s a very significant data point as well. We could even take a larger sample of 2015-2025 minus 2020, with an assortment of supporting casts, and the Warriors had a +2.56 net rating in 4332 minutes with Draymond on and Steph & Durant off. But a really important thing to realize about the Draymond minutes without Steph or Durant is that those net ratings are often being farmed against opposing benches. If we filtered that above-mentioned +2.56 number down to just minutes in which 4 or 5 opposing starters are on the court, suddenly the net rating with Draymond on and Steph/Durant off is -4.21 in 1506 minutes. And it’s still a -1.00 net rating in 2292 minutes if we filter it down to minutes in which at least 3 opposing starters are on the court.

Why are you using a 11 year sample in a peak project when it's obvious those later seasons are going to drag down his peak numbers? Are we going to be looking at 2025 lineup numbers too when discussing peak Westbrook's impact? And I don't see how 3 starters or fewer on Draymond's side against 4 starters or more for the opponent is somehow a fair comparison. For example, that would theoretically even include minutes with Draymond as the lone starter against all 5 opponent starters. I don't know what's the source for these stats so I can't check it myself but equalizing the number of starters seems like the obvious first step. Second is to separate Durant years from the ones before because he changed the team dynamic considerably. So, instead of an 11 year WOWY with varying number of starters, this is what you should look at:

15/16 Draymond + 3 starters (excluding Curry) vs 4 opponent starters
17/18/19 Draymond + 2 starters (excluding Curry/Durant) vs 3 opponent starters

I am guessing he fares pretty well in both which is why you didn't take this obvious approach instead.

I “didn’t take this obvious approach instead” because I never pulled that data, which I never did because it’s evident that zeroing in on very specific starter-state scenarios in specific years would lead to tiny samples. As it happens, though, in the scenario you describe in the 2015 & 2016 seasons, the Warriors had a -32.57 net rating in those minutes, and in the 2017-2019 scenario you described, the Warriors had a -1.45 net rating in those minutes. If you add in the playoffs and combine these, you get a +1.36 net rating. The sample sizes for those very specific scenarios are extremely small though (even the biggest sample listed above is a grand total of 111 minutes). The fact that very specific starter-state scenarios in specific years are really tiny samples is why I opted to pull larger samples in the first instance.

Please note that I’ve also given a sample for Draymond without Steph/Durant against 4 or 5 starters from 2015-17 (i.e. Draymond’s peak years) in a later post that wasn’t responding to you, and it did not look good either. So I actually pulled both larger-sample data from Draymond’s entire career and data that zeroed in on Draymond’s best years, and they were pretty consistent with each other and didn’t look great for Draymond.

This seems like pretty important context to those numbers, because if Draymond were actually leading a team then he’d almost certainly be spending a significantly higher proportion of his minutes against opposing starters than he did in the non-Steph/Durant minutes. His minutes against mostly bench players make up almost half the non-Steph/Durant minutes, while they’d probably be more like 20% or so of his minutes if he were leading a team. So this is good reason to believe that the net rating in those non-Steph/Durant minutes is skewed upwards from what it would look like if Draymond was leading a team.

If Draymond was actually leading a team, he would have two other stars in place of Curry/Durant, not an empty hole on the roster that takes up majority of the cap space. Many of his non-Curry/Durant minutes are played against mostly bench players because he was playing with mostly bench players himself.... which is what happens when you take out 2 of the 5 starters.

No, this entire discussion sprung from One_and_Done talking about what would happen if Draymond had a team with “average” players. Giving Draymond the dynasty Warriors but just with other stars replacing Curry and Durant is pretty clearly not on point for such a discussion. And yeah, that does mean that these scenarios without Steph/Durant effectively involve a team with “an empty hole on the roster” but I actually do think that the Warriors without those guys roughly approximates a team of “average” players (the team still has Klay and often some solid role players—including Iguodala, who I will note was himself a bench player but not a step down from a starter). Which is what the discussion was explicitly about. I’d also note that I provided similar data for Ginobili without Duncan, and it looked amazing for him, so it’s very possible to do quite well in these “empty hole” minutes.

Finally I'll say that the "starter state" is a really cool feature on pbpstats and it's neat to see the data, and it definitely looks very impressive for Ginobili. But RAPM and its derivatives already take account of who is on the floor in a more comprehensive way, and 29-year RAPM has Ginobili and Draymond at +7.1 and +6.9 respectively (9th and 12th all-time).

Yes, that’s true. Except that this actually gets more specifically at a particular type of scenario that is quite important to an assessment of how Draymond would do leading a team. RAPM accounts for all situations the player played in, so the fact that Draymond farmed opposing bench units will naturally boost his RAPM quite a lot, and would have just as much weight as what happens when Draymond is facing opposing starters. If we want to project how Draymond would do leading a team, we should put less importance on what would happen against bench units and more importance on what would happen against opposing starters. RAPM doesn’t actually help us figure that out, while this kind of analysis gives us information directed at that specific inquiry.

[ReggiesKnicks]

11. Shai Gilgeous-Alexander 2024-25 (‘25 > ‘24). I feel in ‘26 he’ll have a great playoffs, with a lesser regular season, and future projects will have ‘26 as his peak due to the superior playoff performance. It’s always tricky with younger players, as we don’t have a larger body of evidence for continued excellence. But Shai ‘25 really was that good.

Just to be clear:

What SGA might do in 2026 should have no bearing on this project.

SGA could be great in 2026 and it may help up better reflect on his 2024 and 2025 runs and give us additional context, but we shouldn't be assuming SGA will be better in the 2026 post-season and use that as a way to prop up 2025 SGA.

We have two consecutive playoff runs which included 33 games where SGA wasn't nearly as impressive in the regular season as the post-season.

The reality is we need to assess young players for what they have done, not what they could do. If this is going to be included in your argument, you should consider Luka "best shape in his life" Doncic here as well, right?

[ReggiesKnicks]

Honestly, for me it’s really hard to choose a year where someone went out in the first or second round—which is like every one of his best years (except 2018 if we think that’s a candidate for one of his best years). There’s just not a lot there in years like that for me to evaluate postseason quality, nor is there the team achievement that I think is a significant component of “greatness.”

Hmm...Do you think there is a big or noticeable difference between 2013 and 2017 CP3? I don't.

If we look at his totality of post-seasons in this stretch, it includes 42 games. Here are his numbers from those 42 games:

9.6 BPM, .240 WS/48, 60.3 TS%

He was really good in this stretch, but ultimately CP3 has too many weird injuries which prevents us from truly considering him Top 10, even though at his best he was closer to the Top 6 than he was to Top 10 in this project.

[lessthanjake]

I don't see any value in assessing 2020 Draymond Green as an example of what it may look like if he were on your team as your best player. The Reality is, Draymond Green is incredibly valuable in the NBA. Top 10-15 defender of all-time, elite passer for a PF/C, High BBIQ, Can push pace with his handle, good in half-court sets as a decision maker. And, the kicker of course, is that since some people love to zero-in on single seasons here since "what happened happened (I am looking at you 2006 Dwyane Wade voters), Draymond was a 39% 3P shooter and best player of a 73-win team during the NBA Finals.

You want to talk about players and their "NBA Finals Performance"? 17/10/6 while leading your team in efficiency, assists, rebounds, steals and blocks is rare and historically dominant.

Yep, and I plan on voting for Draymond in this project at some point, and that’s in part because for my votes I mostly just care about what happened, rather than about hypotheticals. But the discussion about Draymond was inherently about a hypothetical—i.e. what Draymond would do leading a fairly average team. For purposes of my votes, I don’t really care much about the answer to that question, but that doesn’t mean others don’t care about that (evidently they do, because the hypothetical was brought up by someone!), nor does it mean I can’t discuss it. As I’ve explicitly said before in this project while discussing Garnett/Wade with DoctorMJ, my voting approach takes a relatively narrow lens but that does not mean that I don’t find broader discussion of things interesting, even if it’s about things that won’t really factor much into my vote.

[lessthanjake]

Honestly, for me it’s really hard to choose a year where someone went out in the first or second round—which is like every one of his best years (except 2018 if we think that’s a candidate for one of his best years). There’s just not a lot there in years like that for me to evaluate postseason quality, nor is there the team achievement that I think is a significant component of “greatness.”

Hmm...Do you think there is a big or noticeable difference between 2013 and 2017 CP3? I don't.

If we look at his totality of post-seasons in this stretch, it includes 42 games. Here are his numbers from those 42 games:

9.6 BPM, .240 WS/48, 60.3 TS%

He was really good in this stretch, but ultimately CP3 has too many weird injuries which prevents us from truly considering him Top 10, even though at his best he was closer to the Top 6 than he was to Top 10 in this project.

Yeah, I don’t necessarily disagree with this. And I hear your point that I have more to evaluate postseason quality if I expand out my consideration to several years. It’s definitely a valid point. But a significant part of the issue here is that I don’t really know how he would hold up in a larger playoff sample *all in one year.* Chris Paul seems to often break down in the playoffs. That’s a major concern, and cobbling together his stats from various short playoff runs doesn’t necessarily tell me what his postseason quality would be over a deep run. Of course, thinking about what would happen in a deeper run is arguably delving into the land of hypotheticals so I’m a little wary of thinking too much about that, but ultimately the playoffs are about winning a title. So I do actually want to be confident that a player could maintain high quality/health for an entire playoff run, and the only way to have that kind of confidence about a player like Chris Paul is to actually see it happen in reality. We never got that, so I can’t be all that confident about it for any particular year. Of course, another factor is just that team success matters a good bit to me in an assessment of the greatness of a player’s year. On that front, playing well in some early rounds is of course no substitute for a deep run, and that’s true regardless of whether the player in question is one for whom breaking down in the playoffs is a big concern.