APBRmetrics

Hi Folks,

I'm a regular commenter at BBR, but this is my first time here. I have some questions/comments regarding the crazy rebound numbers in the 1960s. When you look at the year-to-year league stats, it becomes apparent that pace alone doesn't explain the gaudy rebound numbers from that era. Here are the league-wide team averages for RPG, missed FGA, missed FTA, and total missed shot attempts for 1968 and '69:

1968: 66.2 rpg, 55.8 mfga, 10.4 mfta, 66.2 total misses
1969: 56.9 rpg, 55.4 mfga, 10.0 mfta, 65.4 total misses

As you can see, the number of missed shots was almost identical in these two seasons, yet the lg rpg saw a huge drop of 9.3 per game. Starting in '69, and every year since, the lg rpg has been roughly equivalent to mfga alone. But from 1958-68, the lg rpg very closely matched the number of total misses:

1958: 71.7 rpg, 72.5 total misses
1959: 70.0 rpg, 70.8 tot "
1960: 73.5 rpg, 73.7 tot
1961: 73.3 rpg, 74.0 tot
1962: 71.4 rpg, 71.9 tot
1963: 66.7 rpg, 66.4 tot
1964: 65.9 rpg, 65.9 tot
1965: 67.3 rpg, 67.3 tot
1966: 68.2 rpg, 68.2 tot
1967: 67.3 rpg, 67.3 tot
1968: 66.2 rpg, 66.2 tot

In 1957, the lg rpg was 62.4 per game, 9.3 rpg lower than 1958. Clearly the league recorded rebounds differently during that 11 year span. Are advanced statheads aware of all this? Is this factored into advanced metrics that evaluate players from that era?

Doesn't this raise some interesting questions? After all, if you miss the first of two attempts, that's a dead ball situation; were guys like Russell and Wilt really getting credit for rebs in those situations? Maybe their stats are even more inflated than we thought... unless Orbs off missed FTs just happened to go through the roof at that time. Or did the league move to a one-and-one situation for those years? But I've seen footage of sixties games where a fouled shooter got two shots even if he missed the first (it's easy enough to find on youtube). I've also seen 60s footage where a FT shooter benefited from a 3-to-get-2 rule, though I'm not sure exactly what years that rule was in effect. If I'm missing something crucial here, please let me know.

AYC

Maybe the concept of team rebounds is somehow involved or they changed when they were rewarded. I dunno.

I found this on changes thru time to free throw rules:

http://actionpcsports.yuku.com/topic/96 ... 55-Present

There was a section on rules at nba.com before but I couldn't find it with a brief search.

You could also post your questions here
http://www.apbr.org/forum/

Some folks there with active interest in NBA history.

I don't think this has to do with team rebounds, because those are still recorded. Clearly the league stopped recording some rebounds after 1968 that had previously been counted. Anyway, if we evaluate rebounders from that time based on missed FGA alone, we would be overstating their production.

Thanks for the link. I didn't realize the 3-for2 rule was still in effect in the early 80s. But wasn't there a time when you only got 2 FTA when fouled shooting a three?

I don't think so.

(Well I did briefly until I figured out that the years listed at the link appear to always refer to the year the season ended and therefore the 79-80 season was covered in the group listed at running from 1980. It looks to me like they always at least awarded 3 FTAs when fouled shooting a three.)

This is pretty interesting. Even if they were counting every missed free throw as a rebound as you eluded there should never be more rebounds than total missed shots as occurred in '58 and '63. Maybe they sometimes credited two players with a rebound when teammates battled for the ball. Out of curiosity do you know how the teams that Wilt and Russell played on compared to the league averages?

I rechecked the numbers, and yes, there were more recorded rebounds (66.725/g) in 1963 than total misses (66.45/g). Russell's Celts and Wilt's teams were almost always the league leaders in pace, which meant they almost always were the league leaders in tm rebounding. (I used TSA/g to approximate pace; it's usually very close to the official pace factor). Clearly stat-tracking was less reliable then.... Maybe team rebounds were being credited to each player involved as you suggested , in which case, Crow was on to something

I got interested in this because it occurred to me that we shouldn't be adjusting player rebound stats just by tm pace, or by tm RPG either. Pace gives you an idea of how many shots were attempted, but not how many were missed; lg RPG is close to lg missed FGA, but always lower (since 1970, at least). Every missed FGA is a rebound opportunity.

Another possible explanation for this oddity could be something to do with put-backs. Maybe official scorers of the time felt when a player tips a miss back towards the hoop without getting set he shouldn’t be penalized for a miss. Thus it would be possible for a single miss to result in both an offensive and defensive rebound. This is just a guess on my part though.

AYC wrote:I got interested in this because it occurred to me that we shouldn't be adjusting player rebound stats just by tm pace, or by tm RPG either. Pace gives you an idea of how many shots were attempted, but not how many were missed; lg RPG is close to lg missed FGA, but always lower (since 1970, at least). Every missed FGA is a rebound opportunity.

I adjust by compiled player rebounds to modernish norms - not team rebounds. I compile all the individual stats so that I can perfectly sum player ratings into the team ratings - having stats outside of individual stats (team rebounds) can skew that somewhat. My ratings adjustments are done team by team - to offset pace &/or weirdness of stat compilation (some teams being generous with home assists/blocks/ & I guess rebounds in the 60s).

If anyone is curious, I've done my player ratings back to 1995 - I have a project I'm working on to project players' future seasons (split into various skillsets and overall ratings) - and 1995 is when we have more of the modernish style (teams actually took threes) to build my dataset used for the projections. But - when I've finished this projection work - I plan on going all the way back with my ratings and see how some of those funky 60's seasons (especially Wilt) look compared to modern seasons. I'll obviously post everything here on the board when I'm done.

OK, I looked at the team by team numbers for 1964 and '65 and totaled them up:

1964: 47,423 total rebs, 47,464 total misses (FG and FT), difference of 41
1965: 48,433 total rebs, 48,440 total misses, difference of 7

Crazy as it sounds, I'm more convinced than ever that ALL missed free throws were being counted as rebounds during this time, even those in dead-ball situations. Big-men like Wilt and Russell probably have inflated numbers as a result

This is a real curiosity and surprise to me - especially as I pored over these historical data at length way back in the pre-modern early 90s, and I don't recall such anomalies. And the exercise at the time was to try to reconstruct rebounding rates on a per possession basis. (I suppose it's impossible that the old paper NBA Yearbooks had different numbers?.)

My failing memory aside, it should be noted that team-wise individual totals don't sum to listed team totals. So, it shouldn't necessarily be inferred that there is any stat padding.

Schtevie, you're a genius. I just took 10 minutes to sum the player rebound totals for 1964 (thank God they only had 9 teams then).

Tot Tm rebs: 47,423, tot Player rebs: 41,884, tot missed fga: 40,470
Tm rpg: 65.9, Player rpg: 58.2, tot mfga/g: 56.2

So, for whatever reason, it seems the league was counting missed FTs towards TEAM rebounding numbers, but not towards the players totals. Which means guys like Wilt and Russ are being underrated if we use team rebound numbers to adjust for era.

PS Statman, I misread your comment the first time; I guess you're a genius too

For instance, 1964 happens to be the best rebounding year of Russell's career; he had 1930 boards in 78 games, 24.7 rpg, and 20.0 r/36. If we adjust to 2011 avg numbers based on tm rebounds, he would have averaged 14.3 rpg and 11.5 r/36 this past season. But if we use the player total rebounds instead of tm rebounds, he would've averaged 16.0 rpg and 12.9 r/36

Team rebounds -- those not assigned to individuals -- are indeed included in total team rebounds in the '60s, and this practice seems to have been phased out over a few seasons. You should just add player rebounds to arrive at [what would now be considered] total team rebounds.

Player rebound rates are still speculative, since opponent rebounds were not kept before 1971 (except in the ABA).

1964 happens to be the best rebounding year of Russell's career; he had 1930 boards in 78 games, 24.7 rpg, and 20.0 r/36. If we adjust to 2011 avg numbers based on tm rebounds, he would have averaged 14.3 rpg and 11.5 r/36 this past season. But if we use the player total rebounds instead of tm rebounds, he would've averaged 16.0 rpg and 12.9 r/36

If we suppose the Celtics' opponents got the league avg rebounds/G in each year of Russell's career, then 1964 is in fact his best rebounding season.

"Team rebounds -- those not assigned to individuals -- are indeed included in total team rebounds in the '60s, and this practice seems to have been phased out over a few seasons."

I don't know, 9-10 team rebounds a game seems awfully high, even considering the pace then. And the correlation between "missing" rebounds and missed FTA is a hell of a coincidence