EDIT: I messed up the model parameters in my post last night; the decline in probability vs. minutes is approximately linear, but the model output is of course still in logit form, corrected numbers below.I know you've tapped out but those are very tricky questions to ask and answer as you are well aware. My response would be that we can and should be asking if coaching decisions contributed to those outcomes. What we should not do is conclude that they either were or were not because its almost impossible to prove.
Back to MPG, I have a bias here and its that these people are generally young, healthy and they are built for hooping. Fatigue is a real issue for NBA players but it really feels like some of the complaints about the C's workload here aren't based on anything other than personal preference. Per your post and the link I shared, its clear that fatigue/game load contribute to injuries but how much is an open question. Maybe we can at least wait for a bad outcome before critiquing workload.
I looked at play-by-play data for every game from the 2019-2020 season and calculated the impact of minutes played and shot type on shot success. For every additional minute played in a game, success rate goes down by (using a binomial model but within the range of observed minutes it's almost perfectly linear):
2-pt hook:
2-pt jump:
2-pt layup:
3-pt jump:
That doesn't look like a lot, but if at the beginning of the game you're a
Anybody who watches the NBA with regularity can see this with their own eyes without having to do hours of data analytics. Overtime puts unusual physical stress on the body (Scanlan et al. 2019), and mental fatigue reduces basketball performance (Cao et al. 2022). One need only look at the 4 overtime Portland-Denver game to see how much player performance suffers due to fatigue, and what a huge advantage a team would have to put fresh legs in; and it's why CJ McCollum (60 minutes) and Dame Lillard (58) missed every shot they took in the final OT and Seth Curry (19 minutes) and Rodney Hood (23) won it - because Mike Malone was too stubborn to take his starters out, despite Jokic (65 minutes) repeatedly turning the ball over and missing free throws and Jamal Murray (55 minutes) building a brick wall. Or look closer to home at Jimmy Butler's ill-fated 3 at the end of the Conference Finals last year. If you go back and read the threads, you'll see the general consensus was he no longer had the legs to drive to the basket and the 3 was his only chance... despite being a, uh, 23% shooter from 3 last year, and playing all 48 minutes which, by my model, would make him more like a 16.5% shooter from deep.
The Panglossian "all coaches know better because of their internal data analytics team" nonsense would be fine if we could all be posting on the Sons of Red Auerbach, but nobody on here is a coach and nobody has access to the kinds of data they do. The eye test is perfectly fine for this kind of thing - the Celtics looked tired in OT last night and a pair of fresh legs almost certainly would have helped. Whether it goes over well in the locker room to take some starters out in favor of the deep bench; or whether it's worth giving starters the experience of playing in tight games late on lots of minutes - those are fine questions. But people shouldn't have to write a dissertation to post casual observations from the games on this board.
edited to add: Of the Celtics' five losses, three have been in overtime. They've lost all three overtime games they've played in. The starters have played every minute of those OTs (*except last night, when Grant played all of OT instead of White, although Grant had played more minutes up to that point). I'll spare you further details, but suffice to say that last night the Heat's starters all had significantly fewer minutes going into OT.
Last edited: