Steals + Offensive Boards - Turnovers

[tsherkin]

It leaves out a lot though, just counting and adding together box score numbers. Forcing missed shots, forced turnovers, tough D that leads to late shots etc.
iDK, I think some of the ousted plus minus impact stats out there probably are more useful. LEBRON, RAPTOR, EPM, RAPM, DARKO etc. these are all a bit different, but I think give you a better idea how a player affects the game.

I think he is trying to isolate possession impact a little more directly.

Right. Yeah I get that. And I see some value in that. But I think these counting stats - steals, OB, Turnovers have so much to do with other factors that don’t show up in the box score.
Maybe some value in it and added possession is certainly important…I just don’t know about it as an individual stat.

Im with you, it just feels like a bar napkin summary, which has its utility.

There is enough tracking data kn NBA.com to add charges, loose balls, contested shots and all that these days as well. Scrape that, do some kind of summary stat, etc. Could be useful.

[165bows]

It's effectively the net extra possessions a player generates. I kind of think this is an important number to know whether positive or negative. You can't find this number anywhere though and no one speaks on it.

We focus on TS% which is scoring efficiency per possession. But we don't look at net possessions generated. I mean you can even create a TS% equation that accounted for net possessions:
Points / (2x (FGA - TO + STL + OREB + 0.44xFTA).


Should this be a commonly discussed stat?
Has this lack of emphasis impacted the love for role players that are big net possession gainers?

I miss the old points per possession metric. I don’t remember the exact calculation but it’s simple, just combined possessions used via FGs, FTs, and TO vs points. It’s similar to TS but also more comprehensive. Idk why it’s not used anymore.

I think the one you described is good but has the same issue as stocks, it’s vulnerable to guys that sell out good team play for those outcomes. Which is tough to tell by box score numbers.

[cupcakesnake]

Decent LEBRON, PIPM

I think he's not s favourable locker room guy but otherwise he's a decent C option I think.

Poor vet presence but is a vet. Not much playoff experience, performs worse broadly in playoffs than RS. Though he was great for Lakers in playoffs, they probably should have kept him.

There is a reason he hasnt been a 30mpg player in half a decade: he isnt worth it. Rebounding and rim defense with no scoring skills, no range, limited ability to defend in space. His utility is quite limited. He was decent for 5 games 3 years ago for LA, but like again, 21.mpg. Taking 6 FGA/g. 23.2% TOV (trash, obviously). Lots of OREB, wasnt blocking shots...


Eeeeeh. He was OVERrated for a while. Now he is just... there. Filling minutes. He isnt trash, I guess, but definitely not a high-impact player.

He was definitely overrated, now I think he's underrated. Like he should be paid 5-10m

I think the sneaky thing with Drummond is that his backline defense was always very poor. He's always been clueless and slow to react. At his athletic peak, he had really ridiculously quick hands and very quick feet. He'd hedge a screen every now and then and look like he teleported. Alas, the physical speed needs to be paired with the mental speed I guess. I used to wonder how he'd far in a more switchy scheme where his strengths were more emphasized.

You don't mind a meh backline defender at center when it's all-NBA level offense, but since Drummond doesn't bring that, it's not a started level player.

Drummond does very real helpful things with his offensive rebounding and quick hands. It's just not enough to out weigh the bad. He looks better in boxscore derived metrics because of the rebounding, but he hasn't been a consistent positive on the court.

[cupcakesnake]

The guys to look at here are Drummond (+9.5 per 100), iHart (+6.3), Paul Reed (+6), Isaiah Jackson (+5.5), Amen Thompson (+4.8), Jonathan Isaac (+4.8), Josh Okogie (+4.4), Larry Nance Jr. (+4.3).

There are a bunch of other bigs you could add to this list, because offensive rebounds are so much more plentiful than steals or turnovers. So I only listed the guys averaging at least 2 steals per 100.

[Braggins]

not sure it's fair in general for PGs, who are more prone to TO's due to a lot of passing and too small for offensive rebounds.

Right, not meaning it to put down certain roles. I'd expect playmakers to be net negative in possession gained.

Maybe one could work assists in to possession usage efficiency with TS%.

I wonder if there is any merit to doing a version where you deduct a players assist/turnover ratio from their turnover stat?

stl + oreb - (tov - ast/tov)

Its kind of interesting, but I haven't thought it through super thoroughly and there might be holes in it.

[zimpy27]

not sure it's fair in general for PGs, who are more prone to TO's due to a lot of passing and too small for offensive rebounds.

Right, not meaning it to put down certain roles. I'd expect playmakers to be net negative in possession gained.

Maybe one could work assists in to possession usage efficiency with TS%.

I wonder if there is any merit to doing a version where you deduct a players assist/turnover ratio from their turnover stat?

stl + oreb - (tov - ast/tov)

Its kind of interesting, but I haven't thought it through super thoroughly and there might be holes in it.

That's an interesting thought. I'll calculate a list of this stuff later in the week and repost it.

I do find assist numbers to be the most complicated stat. It's really an impact of scoring efficiency but it's very difficult to judge the impact it makes. Like if someone scored 2 points and they were assisted, how much of that 2 points is owed to the assister vs the scorer. We seem to double count, like both got 2 points but that's clearly wrong.

Net possession gain and scoring efficiency can be put together with PTS, FGA, FTA, OREB, STL, TO.

DREB and BLK are also a little complicated.

I agree with others that impact stats do a better job than box score to assessing a players value. But one thing I don't like about it is how variable those scores can be from season to season. Far less stable than typical box scores which suggests that it's still the team and the role affecting player impact.

[oldncreaky]

There is a reason he hasnt been a 30mpg player in half a decade: he isnt worth it. Rebounding and rim defense with no scoring skills, no range, limited ability to defend in space. His utility is quite limited. He was decent for 5 games 3 years ago for LA, but like again, 21.mpg. Taking 6 FGA/g. 23.2% TOV (trash, obviously). Lots of OREB, wasnt blocking shots...


Eeeeeh. He was OVERrated for a while. Now he is just... there. Filling minutes. He isnt trash, I guess, but definitely not a high-impact player.

He was definitely overrated, now I think he's underrated. Like he should be paid 5-10m

I think the sneaky thing with Drummond is that his backline defense was always very poor. He's always been clueless and slow to react. At his athletic peak, he had really ridiculously quick hands and very quick feet. He'd hedge a screen every now and then and look like he teleported. Alas, the physical speed needs to be paired with the mental speed I guess. I used to wonder how he'd far in a more switchy scheme where his strengths were more emphasized.

You don't mind a meh backline defender at center when it's all-NBA level offense, but since Drummond doesn't bring that, it's not a started level player.

Drummond does very real helpful things with his offensive rebounding and quick hands. It's just not enough to out weigh the bad. He looks better in boxscore derived metrics because of the rebounding, but he hasn't been a consistent positive on the court.

Part of the reason Drummond is so atrocious as a back line defender is he prioritises rebounds over everything else.

One habit Drummond showed in his last couple of seasons with Detroit that just drove me nuts: he would often focus on positioning for the defensive rebound even when an opposing player was driving to the hoop. I even remember him doing this a couple of times when it was his man who was attacking the basket, and he'd move away to the other side of the rim for a possible rebound. Good news: fewer fouls, better DReb%. Bad news: layup line for the other team.

Drummond is the first player I think of when asked which NBA player has under-performed their physical talents.

[NZB2323]

I looked up this stat for these guys for these seasons:

Hakeem 1985 - +3.7
Moses 1979 - +4.2
Rodman 1992 - +5.5
Karl Malone 1990 + 0.6
Westbrook 2017 - -2.1

[zimpy27]

The guys to look at here are Drummond (+9.5 per 100), iHart (+6.3), Paul Reed (+6), Isaiah Jackson (+5.5), Amen Thompson (+4.8), Jonathan Isaac (+4.8), Josh Okogie (+4.4), Larry Nance Jr. (+4.3).

There are a bunch of other bigs you could add to this list, because offensive rebounds are so much more plentiful than steals or turnovers. So I only listed the guys averaging at least 2 steals per 100.

This is a list of guys who score higher in impact metrics than people would imagine. Drummonds offensive rebounding is huge. I struggle to see how a guy getting you 10 extra shots per 100 is not going to be impactful. He must suck at everything else.

[madskillz8]

Think about a hypothetical player who:

- attacks for offensive boards at every single possession and averages 5.0+ per game
(but this results in lots of 4 on 5 situations and transition plays for the opponent especially after long rebounds following 3pts)

- gambles for steals all the time and averages 3.0+ per game
(giving up many points after failed gambles)

- and basically no offensive skills and in most possessions he's not a part of offensive plays, barely touching the ball (<0.5 TOs)
(making his team easier to defend)

This player could easily lead the league in this stat.

[John Murdoch]

Montrez

[nikster]

Right, not meaning it to put down certain roles. I'd expect playmakers to be net negative in possession gained.

Maybe one could work assists in to possession usage efficiency with TS%.

I wonder if there is any merit to doing a version where you deduct a players assist/turnover ratio from their turnover stat?

stl + oreb - (tov - ast/tov)

Its kind of interesting, but I haven't thought it through super thoroughly and there might be holes in it.

That's an interesting thought. I'll calculate a list of this stuff later in the week and repost it.

I do find assist numbers to be the most complicated stat. It's really an impact of scoring efficiency but it's very difficult to judge the impact it makes. Like if someone scored 2 points and they were assisted, how much of that 2 points is owed to the assister vs the scorer. We seem to double count, like both got 2 points but that's clearly wrong.

Net possession gain and scoring efficiency can be put together with PTS, FGA, FTA, OREB, STL, TO.

DREB and BLK are also a little complicated.

I agree with others that impact stats do a better job than box score to assessing a players value. But one thing I don't like about it is how variable those scores can be from season to season. Far less stable than typical box scores which suggests that it's still the team and the role affecting player impact.

I don't know if there's a point in adjusting for assists. A high volume of ball handling leads to turnovers, and that can also come from taking on a heavy scoring load.

I think it's better to leave the stats as is and just consider roles when comparing too players from the list.

[zimpy27]

Think about a hypothetical player who:

- attacks for offensive boards at every single possession and averages 5.0+ per game
(but this results in lots of 4 on 5 situations and transition plays for the opponent especially after long rebounds following 3pts)

- gambles for steals all the time and averages 3.0+ per game
(giving up many points after failed gambles)

- and basically no offensive skills and in most possessions he's not a part of offensive plays, barely touching the ball (<0.5 TOs)
(making his team easier to defend)

This player could easily lead the league in this stat.

And yet even at worst he'd still probably a postiive impact stat guy where a team perfroms better with him on the court than off. This is because there are so many decent iso scorers in the NBA that can score fairly well without much spacing.

Vanderbilt might be an example?

[toooskies]

Right, not meaning it to put down certain roles. I'd expect playmakers to be net negative in possession gained.

Maybe one could work assists in to possession usage efficiency with TS%.

I wonder if there is any merit to doing a version where you deduct a players assist/turnover ratio from their turnover stat?

stl + oreb - (tov - ast/tov)

Its kind of interesting, but I haven't thought it through super thoroughly and there might be holes in it.

That's an interesting thought. I'll calculate a list of this stuff later in the week and repost it.

I do find assist numbers to be the most complicated stat. It's really an impact of scoring efficiency but it's very difficult to judge the impact it makes. Like if someone scored 2 points and they were assisted, how much of that 2 points is owed to the assister vs the scorer. We seem to double count, like both got 2 points but that's clearly wrong.

Net possession gain and scoring efficiency can be put together with PTS, FGA, FTA, OREB, STL, TO.

DREB and BLK are also a little complicated.

I agree with others that impact stats do a better job than box score to assessing a players value. But one thing I don't like about it is how variable those scores can be from season to season. Far less stable than typical box scores which suggests that it's still the team and the role affecting player impact.

The thing is, player performance isn't stable. We are not dealing with random number generators.