Rebounding in the 60s

[Mike G]

Well, there are 2 factors at work:
- All missed FT are counted as rebounds, even if they weren't rebounded.
- These (and other) "team rebounds" are added in to total team rebounds.

I wonder why b-r.com doesn't just re-add total player rebounds and list these as team totals, since that's the modern practice.

[AYC]

But those are the "official" team numbers from those years, the same numbers I have in the basketball almanac I got 20 years ago. They just shouldn't be used in advanced metrics

[mtamada]

Well, there are 2 factors at work:
- All missed FT are counted as rebounds, even if they weren't rebounded.
- These (and other) "team rebounds" are added in to total team rebounds.

I wonder why b-r.com doesn't just re-add total player rebounds and list these as team totals, since that's the modern practice.

I haven't been following this thread closely, but I suspect that the source of the confusion is exactly what MikeG says. If someone were to create a Hoopstats 101 FAQ, it should contain those three sentences that MikeG wrote, although I'd modify the last one to say that the listing of a team's players' stats should include not just the rebounds of Pau, Kobe, etc. but also a line for "team rebounds". And then the team total (and maybe also have a subtotal which is the sum of individuals' rebounds). Just like boxscores these days tell you how many turnovers each player had in a game -- and also tell you how many "team turnovers" there were, so that those numbers add up to the team's totals.

[Mike G]

You're right, Mike, that there is still such a quantity as total team rebounds, which is individual + 'team' rebounds. And this describes how well a team has rebounded relative to opponents.

But when we're looking at a player's rebounds as a percentage of all rebounds, then it seems total individual rebounds, by team and opposing team, is/are the relevant denominator(s).

Unless a team is particularly adept (or incompetent) at creating team rebounds by causing the opponent to lose the rebound out of bounds, or some such thing, then a player's rebounds can properly be judged relative to how many missed shots were actually rebounded.

Currently, b-r.com shows Portland rebounds for 2010-11 as 3226. This is also the total for their players. There are no 'team rebounds' included.

So I wonder why they don't also list team total rebounds for all eras as simply the sum of their players' rebounds. Everyone who gets into stats from the '60s runs into this issue. It's perplexing because it's so unexpected and unexplained. If there's a good reason to perpetuate this inconsistency, I'd like to hear it.

[AYC]

Mike G, why does it make more sense to use the sum of player rebounds in the denominator, rather than the total number of missed FGA? As I said before, every missed FGA is a rebound opportunity. The total number of possible rebounds is based on missed FGA.

PS I think the same logic should apply to evaluating assists: the number of possible assists in a game is determined by the number of total FGA, not made FG.

[AYC]

PPS One other question. How does the league today treat rebounds of missed free throws when the miss is a live-ball situation? I assume they aren't counted, since the league rpg is lower than missed FGA/g alone. But what about offensive rebounds of missed FTs? Surely, those ought to be recorded if they aren't already....

[Mike G]

PPS One other question. How does the league today treat rebounds of missed free throws when the miss is a live-ball situation? I assume they aren't counted, since the league rpg is lower than missed FGA/g alone. But what about offensive rebounds of missed FTs? Surely, those ought to be recorded if they aren't already....

[I didn't see these last 2 posts earlier. ]
Of course a missed live-ball FT can be rebounded, and the rebound counts as an OReb or a DReb.
A certain number of missed FG are not rebounded by either team: the ball goes out of bounds (incl. blocked out of bounds), the quarter expires with a shot in the air, the shot clock expires on an air ball, the ball gets stuck on the rim/backboard ...
Some of these are not 'possible rebounds'; but a FGA is recorded.

PS I think the same logic should apply to evaluating assists: the number of possible assists in a game is determined by the number of total FGA, not made FG.

How about (FGA + TO) as the denominator?

[AYC]

Yeah, since I posted that I've come to the realization that it's just not workable to use missed FGA alone to evaluate rebounding.

Using FGA +TO does make sense for evaluating assists. The only problem is that we don't have TOV numbers before '78. I'm working on a "dummy" version of an advanced metric that only uses stats that have been available since the start of the shot-clock era: pts, rbs, asts, missed fga and fta. I'm using TSA/g instead of "pace", and I want to use DWS instead of steals and blocks, which aren't so great for measuring D anyway. The one thing I'm bothered about leaving out is TOVs....

[Mike G]

At one point, it seems we agreed that FGX could be used to denote FGA missed.
Then there are 2FG and 2FGX, 3FG and 3FGX.
A lot of confusion can be avoided by maintaining these as separate events.

[AYC]

OK. Don't forget I'm new here....

[EvanZ]

Yeah, since I posted that I've come to the realization that it's just not workable to use missed FGA alone to evaluate rebounding.

Using FGA +TO does make sense for evaluating assists. The only problem is that we don't have TOV numbers before '78. I'm working on a "dummy" version of an advanced metric that only uses stats that have been available since the start of the shot-clock era: pts, rbs, asts, missed fga and fta. I'm using TSA/g instead of "pace", and I want to use DWS instead of steals and blocks, which aren't so great for measuring D anyway. The one thing I'm bothered about leaving out is TOVs....

What about fouls? FGA + TOV + PF (or 0.44*FTA)

possessions = FGA + TOV + 0.44*FTA - OREB, so isn't what you are looking for just possessions + OREB?

[Mike G]

Assists are never given for FT, but only for FG.
If you make 10 good passes, resulting in 4-5 FG and 8-10 FT, is your Ast% --
4/5 = .800 , or
4/(5 + .44*10) = 4/9.4 = .425 ?

Assisted FT would certainly be an improvement in assist-counting.
But you can't just include FTA in the denominator, without the coinciding positive in the numerator.

[AYC]

Right, Mike. Maybe Evan was suggesting I use fouls in the dummy metric I mentioned? I feel like a lot of metrics ignore personal fouls, and that's what I plan on doing....

[EvanZ]

Right, Mike. Maybe Evan was suggesting I use fouls in the dummy metric I mentioned? I feel like a lot of metrics ignore personal fouls, and that's what I plan on doing....

Yeah, I wasn't sure why you were using TOV either. Obviously, assists aren't given for those. It seemed to me like you just wanted the total possible offensive plays for your metric, which is what I think I gave.

[AYC]

I was talking about how to adjust for differences in the way assists have been recorded over the years. I don't want to use TOVs or "possessions" for the same reason I don't want to use "pace", blocks or steals. I'm sticking to pts, rbs, asts and tsa because we have that data going back to the fifties. Not very scientific, but I think it will be useful for my purposes at least.