How much do you value playoffs on vs on/off?

[ceiling raiser]

Been thinking about this recently.

Shortened rotations, smaller npossessions, more multicollinearity, fewer opponent lineups faced, and unusual strategies make on/off very tricky.

The off minutes themselves are incredibly erratic, usually blowouts, odd bench lineups, or some other anomaly.

I do think, on the other hand, on has a lot of value, for both offense and defense.

High level thoughts on the topic?

[LA Bird]

The sample size and lineup multicollinarity issues for off minutes in the playoffs apply for on minutes too. On itself doesn't really say much because we usually have a pretty good idea of how dominant teams are already by their winning margins. It's the on/off part which provides more insight because we don't pay much attention to those few off minutes. So I would still value on/off more than just on. One playoff run with a massively positive or negative on/off probably doesn't mean much but if it's a consistent pattern over multiple postseasons then yes, it does matter.

[eminence]

A big On rating (or net rating to look at pre-'97 groups) over a run is a nice team achievement that speaks well of all the core guys. But to differentiate credit amongst the group I tend to lean pretty heavily on RS priors. There may be exceptions where credit should be distributed significantly differently than the RS, but more often than not I find it pretty similar (Jamal Murray currently may be one of the strongest examples actually, but there aren't many).

[Bad Gatorade]

Raw On/Off? Not that much, even in a huge sample, because there's usually not enough to differentiate it from the regular season in most cases. I'm more inclined to think about it if I see a huge difference, and even then, it's probably more of a tweak to a regular season prior than a definitive indicator of value.

I do believe players can get worse in the playoffs, although I think it's wildly exaggerated. I do think players can be better thanks to putting in more consistent effort/being less reckless/experimental. However, there's a lot of random variation that we need to accept too.

Single season results are almost entirely useless - have a look at the variance in Al Horford's playoff On/Off ratings. However, even a larger sample (e.g. George Hill being +17.6 On/Off from 2012-2018 in 82 games) still produces some wacky ratings. Not that George Hill wasn't a good player in his heyday, but rather, I think we sometimes look for conclusions a bit too soon based off a fairly temperamental data set.

On ratings, IMO, just tend to face far too many contextual factors, but again, seeing an extended pattern that deviates far from the regular season value is going to catch my interest. Not sure it influences my final conclusion a crazy amount (although it may!) but it's still worth looking at.

[eminence]

Raw On/Off? Not that much, even in a huge sample, because there's usually not enough to differentiate it from the regular season in most cases. I'm more inclined to think about it if I see a huge difference, and even then, it's probably more of a tweak to a regular season prior than a definitive indicator of value.

I do believe players can get worse in the playoffs, although I think it's wildly exaggerated. I do think players can be better thanks to putting in more consistent effort/being less reckless/experimental. However, there's a lot of random variation that we need to accept too.

Single season results are almost entirely useless - have a look at the variance in Al Horford's playoff On/Off ratings. However, even a larger sample (e.g. George Hill being +17.6 On/Off from 2012-2018 in 82 games) still produces some wacky ratings. Not that George Hill wasn't a good player in his heyday, but rather, I think we sometimes look for conclusions a bit too soon based off a fairly temperamental data set.

On ratings, IMO, just tend to face far too many contextual factors, but again, seeing an extended pattern that deviates far from the regular season value is going to catch my interest. Not sure it influences my final conclusion a crazy amount (although it may!) but it's still worth looking at.

Referencing first paragraph - Can you think of any guys with huge differences over largish samples? I mentioned Murray earlier, it's not a huge sample, but I think it may be clear enough to give it some weight (helped by showing up strongly in the box-score). Dirk has a pretty negative signal in just the on/off area. Is there anyone else who comes to mind?

[Doctor MJ]

Been thinking about this recently.

Shortened rotations, smaller npossessions, more multicollinearity, fewer opponent lineups faced, and unusual strategies make on/off very tricky.

The off minutes themselves are incredibly erratic, usually blowouts, odd bench lineups, or some other anomaly.

I do think, on the other hand, on has a lot of value, for both offense and defense.

High level thoughts on the topic?

I think it's important to look at playoff on/off data, but often there's not enough there for me to feel like I can weigh it that much.

On data is better on this front, and I use it more, but it is a situation where I tend to see that dividing line between positive and negative to be a bigger deal than the extremes around it. If you have primacy on your team, and your team is winning with you on the floor, hard for me to be too critical.

I do like looking at things like OnWin (games with positive +/-) and guys with positive +/- in losing series, but a lot of times this is more geared toward supporting players than the lead star.

[Blame Rasho]

I will copy and paste a discussion I remember and what on / off means at times.

I remember Popovich going on an amusing semi-rant about this many years ago. Basically how all the nerds in the organization were badgering him to give Matt Bonner as much PT as possible because their lineup data was off the charts when he was on the court. And yet from his perspective, all he could see was him getting his ass kicked on defense while making some 3s here and there and never passing so he said he eventually threw all that stuff in the trash and told them to stay out of his office. And then they'd try to point out that Manu also rated incredibly high in order to bolster their argument, and he'd be like yeah, but I don't need some guy with 10 degrees from Harvard to tell me that.

Yea, and a guy like Bonner might have been off the charts because Pop used him selectively and then took him off the court if things start getting bad. If he played him more and in more unfavorable situations, then his lineup data would probably have started to look less pleasing.

Exactly, you have to put context in a lot of stats. You don’t put Matt Bonner to play 38 mins a game just because the team looks good in the 15 mins he is in the game. The sample size matter along with who on the other side of the court is playing. Manu was a per min beast but you couldn’t have him playing hard for 35 mins every game.

You have to put context in any stat esp this one.

[rk2023]

I prefer using the “on” number more with working with smaller samples.

[J-Wolves]

James Harden (Elimination Playoff Games)

19 elimination games : 123-303 @ 40.5% (43-147 3pt) @ 29.2%. 86 turnovers. -117 (+/-)
Only 14 turnovers in his first 6 elimination games.
Last 13 elimination games (handling the ball more) - 72 turnovers (5.5 per game) since 2015.

2023 Celtics 3-11 (1-5 3pt), 5 turnovers. -30
2022 Heat 4-9 (3-7 3pt), 4 turnovers. -16
2021 Bucks 5-17 (2-12 3pt), 4 turnovers. -2
2020 Lakers 12-20 (2-8 3pt), 6 turnovers. -29
2020 Thunder 4-15 (1-9 3pt), 4 turnovers. +9
2019 Warriors 11-25 (6-15 3pt), 6 turnovers. -10
2018 Warriors 12-29 (2-13 3pt), 5 turnovers. -13
2017 Spurs 2-11 (2-9 3pt), 6 turnovers. -28
2016 Warriors 12-23 (3-7 3pt), 7 turnovers. -27
2015 Warriors 2-11 (0-3 3pt), 12 turnovers. -4
2015 Clippers 7-20 (2-7 3pt), 7 turnovers. +12
2015 Clippers 5-20 (2-6 3pt), 1 turnover. -21
2015 Clippers 9-20 (1-8 3pt), 5 turnovers. +23

2014 Blazers 9-15 (4-6 3pt), 4 turnovers. +7
2013 Thunder 7-22 (4-10 3pt), 4 turnovers. +3
2012 Heat 5-11 (3-8 3pt), 3 turnovers. -17
2011 Mavs 7-11 (1-5 3pt), 2 turnovers. +3
2011 Grizzlies 6-10 (4-8 3pt), 1 turnover. +24
2010 Lakers 1-3 (0-1 3pt), 0 turnovers. -1