DraymondGold wrote:Finally getting the chance to follow back up on the rate metrics vs per-season metrics. Apologies for the delay! I figured I'd reply to this in the new thread, since Curry is on pace to be in the top two players with Magic. Magic seems like the favorite to win (and I may end up voting for Magic first), but there still are arguments one could make for Curry. I'll make two of them here.
Per 100 Possession metrics vs Total Season metrics for CurryOhayoKD wrote:Uh...how?
Is the rate he is outscoring them per possession likely to increase?
Are his replacements going to look better if they less time against opposing starters?
I recall 2020 Giannis barely playing in the fourth quarter. Does not seem to have been a positive influence on his impact to not be in those circumstances anymore.
Fair enough to point out that Steph *could* maybe theoretically maintain a similar rate with more time played… but at that point it is a question of confidence.
Presenting these lower minutes as some sort of disadvantage seems like a stretch.
Uh… how? Read the rest of the post! Lol
-Is his per season value going to go up if he plays more fourth quarters? Yes.
-Is the rate he is outscoring them per possession likely to increase? I provided reasoning that it won’t decrease, and has a chance to increase.
-Are his replacements going to look better if they less time against opposing starters? These are adjusted metrics, not raw ones, so they adjust for opponents on the court. Regardless, the point is that Curry is resting in blowouts, when the opposing starters would also be resting. In which case if opponents would have any affect, Curry would be playing more minutes against non starters anyway.
You recall Giannis resting a bunch of fourth quarters in 2020. That’s… not really true.
Number of Fourth Quarters Rested for some recent stars:
Garnett 2001–04: 4, 6, 2, 3 —> Average: 3.75 games a season
Giannis 2019–22: 5, 6, 6, 4 —> Average: 5.25 games a season
LeBron 2009–13: 14, 5, 8, 7, 9 —> Average: 8.6 games a season
Jokic 2020–23: 9, 15, 11, 10 —> Average: 11.25 games a season
Paul 2012–15: 7, 16, 10, 16 —> Average: 12.25
Curry 2015–18: 20, 19, 25, 18 —> Average: 20.5 games a season
So Curry rested 16.75 more fourth quarters *per season* than Garnett did! And nobody’s at all close. Not Giannis on his dominant teams (he rests 6 in 2020), certainly not Garnett who’s up for vote now.
And remember, Curry’s not just skipping more fourth quarters, his team is also dominant enough that he’s playing less in the fourth quarters he doesn’t miss (see previous post).
The rate comment was sort of supplemental to my point. But I wasn’t very clear about my point, so let me be more explicit:
-Jake provided (a wall of) stats that favor Curry over Garnett. Some of which were rate stats.
-You argued these rate stats overrate Curry and underrate Garnett because Garnett played more minutes/possessions than Curry, which makes up for the gap in the rate stats.
-I’m arguing that Curry’s teams were so dominant (in no small part because of Curry) that
1) Peak Curry missed far more fourth quarters entirely than any other star in this era [and provided clear evidence for this to be the case]
2) In the fourth quarters peak Curry did play, that he played fewer minutes than expected, which is again because his teams are so dominant [and provided clear evidence for this to be the case]
3) Therefore: some of the gap in per-season volume is from Curry’s missed fourth quarters.
Side Note: Curry played in a much faster era than Garnett. So Curry played more possessions per minute than Garnett. So Curry would be making up even more of the possession gap than the minute gap if he played in the 4th quarter. And the impact stats are usually per 100 possessions, not per 36 minutes.
4) We should only expect Curry’s total season value in these metrics to go up if he played more the fourth quarter, which he would have if he played on any other team that was less dominant.
One counter argument is that if he played more minutes, he would get more tired, so his per-100 rate would go proportionally, such that his per-season value remains roughly the same. This doesn’t seem to work in this situation. Why?
4i: First, there’s the obvious reason that’s he’s playing more. You’d expect volume to go up as minutes / possessions go up
4ii: Second, he’d be playing more at the end of the game. It’s not like he’d be playing more in the 2nd quarter, which might tire him out in the 3rd or 4th. We’d be adding minutes to the 4th quarter, so we could reasonably expect his performance in Quarters 1-3 to be fairly unchanged.
4iii: Third, he’d be playing more particularly in the games that were blowout victories for the Warriors, which correlate strongly with Curry’s better games with positive plus/minus. So he’d be playing more during his better games, when he’s hot, or when he’s already posting all-time level plus minus. Therefore we should expect awarding him more minutes in the 4th should increase his total season value.
Now here’s where we get into personal criteria. As above, people can have different criteria. I can understand penalizing someone (or at least not being willing to project value) for missed time due to injury, or sudden retirement, or something like that. But, me personally, I have a much harder time penalizing someone because their team was so good that they were in so many blowouts that they didn’t have to play that much.
It’s a very rare situation. It’s rare for someone to be on such a good team that they play notably fewer minutes, fewer possessions, than other stars of their caliber in their era. But if this is the case for a player, and if that player looks better in rate stats but slightly behind in total season value due to playing less possessions because they were in so many blowouts, I’m less interested in rating a player lower for that.
I’ve made the case that Curry fits in this situation. As such, I’m more willing to *slightly* mentally curve up his total-season value if he were in another situation, one where his Warriors weren’t so dominant that he got to rest in more 4th quarters than any other star in the modern era.
Curry's Career WOWY vs GarnettOhayoKD wrote: Ohayo, for someone who values WOWY so highly,
How highly though? If we were to look at my post arguing Kareem peaked higher than Micheal...
https://forums.realgm.com/boards/viewtopic.php?p=107462472#p107462472...we can see WOWY, as you are using it, is merely a small chunk of the whole of my case.
As far as data goes,
most of the post is directed towards years where
there is no direct WOWY. And this is for
Kareem, a player where we have very limited alternatives in terms of assessing impact. I have maintained that WOWY has use
even when we have access to data like RAPM. If you thought that meant I rank players according to regular-season WOWY averages(and we barely even have WOWY for KG's best regular-seasons), that is on you.
I think real-world stuff is especially useful to compare outliers(2016, 2004, 2009 ect), or to examine
why something is happening in the artifical-stuff(Duncan staggering minutes with Drob's bad replacements, Spurs not really affected by Manu's absence, ect), but on the flip-side, lineup-adjustment makes things less noisy and is useful for establishing a
baseline over
longer time-frames. FWIW, I do rank RS Steph's highs pretty highly, above the best years for players like Hakeem and Jordan. But the same is true for KG. So unless you are willing to make the case years like 2016 were on another level compared to KG's 2004 or 2003, we get into how they
generally look(KG carries an overwhelming advantage in both of the extended rapm sets we have), how much of their value can be tied to situation(very strongly favors Garnett), and what they and their casts do in the playoffs(2004 beats out 2016 pretty handily on that both fronts imo).
Considering all of that, I'm left thinking KG was the better player. It is
then we get to the longevity, and there, a reasonably close comparison turns into a lopsided one.
And as "low" as I am on Steph, I am still likely to vote him 11th or 12th. Ahead of Bird and possibly ahead of Kobe. But he is not the best "impact" candidate currently on the board. Nor is he the most successful. And I certainly do not see why skill-set analysis would put him ahead of an all-time two-way big and an all-time helio. Two archetypes that have generally established higher floors
and ceilings than the likes of MJ, Steph or Bird. Steph is probably the pinnacle of his archetype, and I imagine he would just look better and better the more you teleport him back, but in an era-relative comparison, I do not see his case for 9.
Kg’s already voted in (and I was a voter for him), so no big disagreement on KG, but I do disagree on the explanation here for KG, as presented.
For someone who says “The main advantage of WOWY is that you can see what truly happens when a player is removed from a team.”, for someone who characterizes WOWY as the true measure of wholistic global impact, this portrayal of Steph seems just flat out incorrect.
Steph clearly has better WOWY than KG. *For his career*.
Here’s all the raw single season WOWY data for KG and Curry: [spoiler]in csv format:
KG,with MoV,games with, wihtout MoV,games without,raw WOWY,sample size (smaller sample of games),Season WOWY,
1996,-5.13,80,-14.5,2,9.37,2,749.6,
1997,-0.47,77,-17,5,16.53,5,1272.81,
1998,0.71,82,,0,0.71,0,58.22,
1999,0.57,47,-2.67,3,3.24,3,152.28,
2000,2.63,81,-6,1,8.63,1,699.03,
2001,1.62,81,-20,1,21.62,1,1751.22,
2002,3.21,81,17,1,-13.79,1,-1116.99,
2003,2.07,82,,0,,0,563.8866667, Use weighted average of 2001–2008 for total season WOWY during the missing 2003–05 years: 6.876666667
2004,5.49,82,,0,,0,563.8866667, Use weighted average of 2001–2008 for total season WOWY during the missing 2003–05 years: 6.876666667
2005,1.45,82,,0,,0,563.8866667, Use weighted average of 2001–2008 for total season WOWY during the missing 2003–05 years: 6.876666667
2006,-1.41,76,-7.83,6,6.42,6,487.92,
2007,-2.45,76,-19.17,6,16.72,6,1270.72,
2008,10.48,71,8.82,11,1.66,11,117.86,
2009,9.14,57,3.8,25,5.34,25,304.38,
2010,4.39,69,-0.23,13,4.62,13,318.78,
2011,5.38,71,5.27,11,0.11,11,7.81,
2012,3.13,60,-3.67,6,6.8,6,408,
2013,0.16,68,-2.23,14,2.39,14,162.52,
2014,-1.72,54,-4.46,28,2.74,28,147.96,
2015,-3.02,42,-2.73,40,-0.29,40,-12.18,
2015,0.8,5,-9.42,77,10.22,5,51.1,
2016,-2.61,38,-4.34,44,1.73,38,65.74,
,,,,,,,,
Curry,with,games with,wihtout,games without,raw WOWY,sample size (smaller sample of games),Season WOWY,
2010,-3.43,80,-10.5,2,7.07,2,565.6,
2011,-1.96,74,-5.75,8,3.79,8,280.46,
2012,-0.54,26,-5.28,40,4.74,26,123.24,
2013,1.04,78,-2,4,3.04,4,237.12,
2014,5.21,78,-3,4,8.21,4,640.38,
2015,10.56,80,-8.5,2,19.06,2,1524.8,
2016,11.35,79,-5,3,16.35,3,1291.65,
2017,12.05,79,0.67,3,11.38,3,899.02,
2018,9.71,51,-0.16,31,9.87,31,503.37,
2019,8.61,69,-4.92,13,13.53,13,933.57,
2020,-11,5,-8.52,60,-2.48,5,-12.4,
2021,3.06,63,-13,9,16.06,9,1011.78,
2022,7.42,64,-1.17,18,8.59,18,549.76,
2023,3.05,56,-0.88,26,3.93,26,220.08,
We can look at their full career weighted average WOWY, weighting by the smaller with/without sample (which is the dominant source for noise):
KG’s Career avg WOWY: +3.4 (+4979.3 total career WOWY)
Curry’s Career avg WOWY: +7.8 (+6900.3 total career WOWY)
Career Average WOWY: Curry >> KGKG Career WOWY (avg WOWY * total games): +4979.3
Curry’s Career WOWY (avg WOWY * total games) +6900.3
Total Career WOWY, method 1: Curry >> KGWe can sum their WOWY each season (individual season WOWY * games in that season) to calculate a ballpark full career WOWY, method 2:
KG’s Career WOWY (summing up each season's WOWY): +8538.9
Curry’s Career WOWY (summing up each season'w ): +8768.4
Total Career WOWY, method 2: Curry > KG*Note that we have to interpolate for 2003–2005, since KG doesn’t have any missed games in those years. I use a weighted average of 2000–02 and 2005–2008 for those years for this method. You could improve his performance up to 15% in these years before KG finally surpasses Curry. While 2003 and 2004 are the best years in this stretch, 2005 is also the worst year in this stretch, making it less likely these three years on the whole are 15% better than the surrounding years. Even if so, this method doesn't give any extra benefit for having a better prime or peak (like most people do), which Curry just might have...
So what if we just look at 10-year Primes?
1999–2008 KG total WOWY: +4796.9
2000–2009 KG total WOWY: +4630
2013–2022 Curry total WOWY: +6564.8
2014–2023 Curry total WOWY: +5610.3
10-year Prime WOWY: Curry >> KGNote that the career value numbers and even prime numbers do *not* curve shortened seasons (1999, 2012, 2020, 2021) to full-length seasons. Curry would benefit more than KG if we did this.
Note that this does *not* apply corrections for diminishing returns on better teams, which would benefit Curry more than KG. When we do account for diminishing returns, recall that KG is 4th all time (above LeBron, Shaq, Bird, Hakeem, Magic, Duncan, Russell, Kobe, Jordan, Wilt, Kareem) in the single-season prime WOWY ranking by Thinking Basketball…. suggesting prime Curry is higher than those players too in prime WOWY.
And note that this does *not* include multi-season WOWY (e.g. comparing 2020 vs 2021 Warriors with or without Curry, without KD or Klay to bias the numbers), where Curry also *clearly* surpasses KG, per my previous post:
I mean if we look at their large-sample multi-year data, ranked by how valuable the samples are:
*Alternate Value: 2019–20 Warriors: Total change: +12.97 [Alternate years: Alternate Value using 2019 instead of 2021, subtracting 2021 Durant’s 1.97 WOWY. Klay’s 17–19 WOWY / 22 WOWY are both too noisy to use and at same time as other injuries.]
-2018 Warriors: 9.71 with, -0.16 without. Total change: +9.87 [Injury year] (51 games)
*Alternate Value: 2007–08 Celtics: Total change: +9.30 [Teammate Adjustment: Alternate Value subtracting 07 Ray Allen’s 3.7 WOWY]
-2020–21 Warriors: 0.4 with, -8.52 without. Total change: +8.92 [Injury year]
-2022–23 Warriors: 5.38 with, -1.0 without. Total: 6.38 [Injury year]
-2015–16 Timberwolves: -2.21 with, -7.57 without. Total change: +5.36 [Traded, joining Timberwolves]
-2012 Warriors: -0.54 with, -5.28 without. Total change: +4.74 [Injury year]
-2008–09 Celtics: 9.88 with, 5.33 without. Total change: +4.55 [Injury year]
-2013–14 Celtics: -0.62 with, -4.97 without. Total change: +4.35 [Traded, leaving Celtics]
*Alternate Value: 2009–10 Celtics: Total change: +3.98 [Alternate years: Alternate value using 2010 instead of 2008]
-2007–08 Timberwolves: -3.16 with, -6.26 without. Total change: +3.1 [Traded, leaving Timberwolves]
-1995–96 Timberwolves: -5.14 with, -8.22 without. Total change: +3.08 [Rookie year]
-2016 Timberwolves: -2.61 with, -4.34 without. Total change: +1.73 [Injury year]
-2009–10 Warriors: -3.28 with, -3.8 without. Total change: +0.52 [Rookie year]
-2015–16 Nets: -3.02 with, -2.73 without. Total change: -0.29 [Traded, leaving Nets]
-2016–17 Timberwolves: -2.61 with, -0.64 without. Total change: -1.97 [Retirement]
-2013–14 Nets: -1.57 with, 1.25 without. Total change: -2.82 [Traded, joining Nets]
Considering how KG is 4th in the prime raw WOWY ranking (single-season data), and has the longevity advantage... and Curry *still* ends up on top for career WOWY value, I'd say that's highly compelling, especially in conjunction with all the other impact metrics that portray him as clearly a top 10 player for primes and many even for career value.
There’s obviously other stats too, but I’m really not getting how someone who values WOWY more than most posters here (obviously including other data/criteria too ofc) doesn’t see Curry’s GOAT level WOWY as cause to rank him higher….
