I recently re-read a 2018 article from The Coffin Corner by Kenn Tomasch, “The AAFC vs. the NFL: The Attendance Myth.” I had written an article for The CC on the subject in 2007 and was struck by discrepancies in NFL attendance figures from 1946-49 that Tomasch cited in his articles and those I found while researching my article. Here are the figures from the NFL Record and Fact Book I used. Since the point of my article was a comparison, I also include the AAFC figures:
1946: NFL total was 1,732,135 and average per game was 31,493; AAFC figures were 1,376,998 and 24,589
1947: NFL total was 1,837,437 and average per game was 30,624; AAFC figures were 1,828,480 and 32,651
1948: NFL total was 1,525,243 and average per game was 25,421; AAFC figures were 1,618,626 and 28,904
1949: NFL total was 1,391,735 and average per game was 23,196; AAFC figures were 1,122,811 and 26,734
1946-49: NFL total was 6,486,550 and average per game was 27,600; AAFC figures were 5,946,915 and 28,319
Based on what I assume are the NFL's official numbers, the AAFC outdrew the NFL in three out of four seasons (all but 1946) and cumulatively in the four-year period.
Tomasch cites team media guides and Total Football for his NFL figures:
1946: 1,873,229/34,059
1947: 1,882,713/31,379
1948: 1,629,093/27,152
1949: 1,621,843/27,031
1946-49: 7,006,878/29,817
If those figures are right, the NFL drew more twice (1946 and 1949), the AAFC drew more twice (1947 and 1948) and the NFL drew more for the four-year period. The AAFC comparison aside, why would there be discrepancies every year in which the NFL numbers from the Record and Fact Book are lower every year and there's a difference of almost 10% for the four-year period? Most importantly, which figures are right?
NFL Attendance Discrepancies, 1946-49
-
TanksAndSpartans
- Posts: 1206
- Joined: Wed Nov 18, 2015 1:05 am
Re: NFL Attendance Discrepancies, 1946-49
Andy, sorry I don't have an answer on the attendance figures, but I think the topic is interesting. Did you see Kenn's web site where he hints at a typo:
"If you notice, the discrepancy from the NFL book is exactly 10,000 per game, meaning it’s very likely just a sloppy mis-reading of the data on the part of someone at the NFL."
I didn't realize the AAFC's attendance was as good as it was, but it seems all the teams were down by '49 and with just 7 teams, I guess its not that surprising they folded/merged.
"If you notice, the discrepancy from the NFL book is exactly 10,000 per game, meaning it’s very likely just a sloppy mis-reading of the data on the part of someone at the NFL."
I didn't realize the AAFC's attendance was as good as it was, but it seems all the teams were down by '49 and with just 7 teams, I guess its not that surprising they folded/merged.
-
Andy Piascik
- Posts: 155
- Joined: Fri Oct 24, 2014 11:32 pm
Re: NFL Attendance Discrepancies, 1946-49
It certainly is surprising the NFL wouldn't have fixed such a mistake if indeed that is the case given how meticulous they are about such things. I have NFL Record Manuals almost 50 years old and every single one of them, right on down to the last print version of the Record and Fact Book, have the same figures.
My first thought was that some teams were overstating their attendance figures in their media guides and that the league office corrected them. If I'm not mistaken, there was a time when no-shows were not included in the official attendance numbers. If all those figures in the media guides are in fact accurate, it would be pretty simple to just correct the league's total attendance.
John, I only looked at the four years I mentioned. Does Tomasch say whether these discrepancies involve other seasons?
My first thought was that some teams were overstating their attendance figures in their media guides and that the league office corrected them. If I'm not mistaken, there was a time when no-shows were not included in the official attendance numbers. If all those figures in the media guides are in fact accurate, it would be pretty simple to just correct the league's total attendance.
John, I only looked at the four years I mentioned. Does Tomasch say whether these discrepancies involve other seasons?
-
Throwin_Samoan
- Posts: 168
- Joined: Wed Mar 02, 2016 5:17 pm
- Location: Phoenix, Arizona
Re: NFL Attendance Discrepancies, 1946-49
Annual yearly attendance doesn't make an appearance in the NFL Record Manuals until 1963. (It's listed in the table of contents for the 1962 edition, but page 77 - at least in my copy - is blank.) The totals printed then were no doubt carried through with each reprinting as they probably saw no need to go back and re-do the work, so if you look at all the figures in all the Record Manuals (later Record & Fact Books), you will see the same numbers for 1946-1949:
1946 -1,732,135
1947 -1,837,437
1948 -1,525,423
1949 -1,391,735
(Individual team totals are not listed in the 1947-1950 guides to check the math. When the NFL released its "official" counts, described as "turnstile counts," in late 1946, the overall regular season number was 1,950,154, which doesn't agree with anything else.)
And Record and Fact Book totals aren't even always consistent within the same edition (likely a paid vs. actual discrepancy, but I would have to check). Take the 1995 edition, just as an example. Each team has its total attendance listed in its section, and if you total up the 28 teams' listed figures, there's a discrepancy between those totals and the overall league total elsewhere in the book of 572,649 (or 2,556 per game). (The team totals add up to 13,457,786 - the league paid total is listed as 14,030,435 - where it remains to this day - or at least the 2018 edition, the last printed one I have.)
Team media guides don't always agree with the R&FB totals, either. For example, the Browns' 1994 Media Guide lists their AAFC totals as:
1946 - 339,962 (399,963 "Official" AAFC number from 1947 league guide, 399,962 Total Football sum)
1947 - 392,760 (390,939 AAFC number, 399,760 TF sum)
1948 - 318,619 (Same AAFC numbers, same TF sum)
1949 - 189,604 (188,947 AAFC number, 189,604 TF sum)
Total Football agrees on the Browns' media guide numbers, which are not always the same as the "official*" AAFC numbers, as you see above.
Total Football billed itself as "The Official Encyclopedia of the National Football League," so it's not some rogue publication, but even those numbers (where game scores are listed from 1920-1996) sometimes don't agree with the NFL numbers (the Bengals' 1994 TF numbers, for example, total 421,964, whereas their R&FB numbers total 415,368, matching their 1997 Media Guide numbers).
Point being: there are many different "official" sources (and I haven't even bothered to mount a newspaper box score search), and they do not always agree. In the absence of a table in the 1947, 1948, 1949 and 1950 NFL Record Manuals to check the math, you either have to piece together figures from media guides or Total Football or newspapers or PF-R, I suppose.
All the AAFC numbers in the original story are straight from the AAFC. The NFL numbers have to be pieced together, and they don't always agree. (The 1946 NFL numbers were off by 5, and the discrepancies between them and the first pass in years that followed were 12,853, 10 and 19,677. It's nice when they're spot-on, but they aren't always.)
As I'm sure many of you do, I have a handful of NFL Game Books from that period, but attendance is rarely if ever listed in them, that I can see. Back in those days, when there were more newspapers than there are now, you'd not only often see different figures (sometimes estimates) in different papers, you'd see different yardage totals and even player names.
At this stage of the game, these are discrepancies that will be difficult to reconcile. As there is no agreed-upon hierarchy of accuracy that I'm aware of, each researcher must make his or her own determination.
*"Official" just means "from the office," or "authorized or issued authoritatively," not necessarily "100% correct and consistent." Study this game long enough, you find myriad discrepancies and errors of math prior to spreadsheets and computers.
1946 -1,732,135
1947 -1,837,437
1948 -1,525,423
1949 -1,391,735
(Individual team totals are not listed in the 1947-1950 guides to check the math. When the NFL released its "official" counts, described as "turnstile counts," in late 1946, the overall regular season number was 1,950,154, which doesn't agree with anything else.)
And Record and Fact Book totals aren't even always consistent within the same edition (likely a paid vs. actual discrepancy, but I would have to check). Take the 1995 edition, just as an example. Each team has its total attendance listed in its section, and if you total up the 28 teams' listed figures, there's a discrepancy between those totals and the overall league total elsewhere in the book of 572,649 (or 2,556 per game). (The team totals add up to 13,457,786 - the league paid total is listed as 14,030,435 - where it remains to this day - or at least the 2018 edition, the last printed one I have.)
Team media guides don't always agree with the R&FB totals, either. For example, the Browns' 1994 Media Guide lists their AAFC totals as:
1946 - 339,962 (399,963 "Official" AAFC number from 1947 league guide, 399,962 Total Football sum)
1947 - 392,760 (390,939 AAFC number, 399,760 TF sum)
1948 - 318,619 (Same AAFC numbers, same TF sum)
1949 - 189,604 (188,947 AAFC number, 189,604 TF sum)
Total Football agrees on the Browns' media guide numbers, which are not always the same as the "official*" AAFC numbers, as you see above.
Total Football billed itself as "The Official Encyclopedia of the National Football League," so it's not some rogue publication, but even those numbers (where game scores are listed from 1920-1996) sometimes don't agree with the NFL numbers (the Bengals' 1994 TF numbers, for example, total 421,964, whereas their R&FB numbers total 415,368, matching their 1997 Media Guide numbers).
Point being: there are many different "official" sources (and I haven't even bothered to mount a newspaper box score search), and they do not always agree. In the absence of a table in the 1947, 1948, 1949 and 1950 NFL Record Manuals to check the math, you either have to piece together figures from media guides or Total Football or newspapers or PF-R, I suppose.
All the AAFC numbers in the original story are straight from the AAFC. The NFL numbers have to be pieced together, and they don't always agree. (The 1946 NFL numbers were off by 5, and the discrepancies between them and the first pass in years that followed were 12,853, 10 and 19,677. It's nice when they're spot-on, but they aren't always.)
As I'm sure many of you do, I have a handful of NFL Game Books from that period, but attendance is rarely if ever listed in them, that I can see. Back in those days, when there were more newspapers than there are now, you'd not only often see different figures (sometimes estimates) in different papers, you'd see different yardage totals and even player names.
At this stage of the game, these are discrepancies that will be difficult to reconcile. As there is no agreed-upon hierarchy of accuracy that I'm aware of, each researcher must make his or her own determination.
*"Official" just means "from the office," or "authorized or issued authoritatively," not necessarily "100% correct and consistent." Study this game long enough, you find myriad discrepancies and errors of math prior to spreadsheets and computers.
-
Andy Piascik
- Posts: 155
- Joined: Fri Oct 24, 2014 11:32 pm
Re: NFL Attendance Discrepancies, 1946-49
That's a tremendous post. What do you think about turning it into an article for The Coffin Corner?
-
TanksAndSpartans
- Posts: 1206
- Joined: Wed Nov 18, 2015 1:05 am
Re: NFL Attendance Discrepancies, 1946-49
That was a great post!
Andy, it seems Kenn focused just on the AAFC time period. I think he read in a book that the AAFC outdrew the NFL and checked into it himself. He says the blog post is pretty much the same as the CC article:
https://www.kenn.com/the_blog/?p=9307
Andy, it seems Kenn focused just on the AAFC time period. I think he read in a book that the AAFC outdrew the NFL and checked into it himself. He says the blog post is pretty much the same as the CC article:
https://www.kenn.com/the_blog/?p=9307