Thirteen years ago, Doug Drinen first looked at NFL player birth days (the month and the day, not the year) at the old PFR Blog. Six years ago, I took a second look at that study.  And over the last two days, I spent time looking not just at when NFL players were born, but when people are born in the United States. Today is a continuation of those studies.

There have been 17,027 players to enter the NFL and play in a game since 1970.  Given what we know about the birth rate in the United States, we can then estimate how many NFL players should have which birth date.  For example,  the most common birthday, according to one study, is September 9th.  Roughly 0.303% of all U.S. babies are born on this day, which means we would “expect” there to be 51.5 players in this time period with 9/9 birthdays. It turns out that there have been 48, which is pretty close. September 8th is the date with the most NFL players since 1970 — 72 — compared to an expected result of 50.2 players.

Those are just two dates, of course, so let’s look at the entire calendar. I performed this same calculation for every date.  As it turns out, January 1st is the biggest outlier: it is a very rare day for babies to be born, but it’s actually an above-average day for NFL players.  More on that in a minute.  To smooth out the chart, I looked at 31-day periods, so the month of July would be captured by looking at July 16th on the draft (each data point represents that day, along with the 15 days before and after).

Two other notes.  For the ends of the year — early January and late December — I pro-rated the numbers to make them work while still using just calendar year cut-offs. ^[1]Specifically, here’s what I did.   For days before January 7th, I excluded due to small sample size.  Beginning on January 7th, I looked at the 7 days from the 1st to the 7th, and then … Continue reading The blue line represents the actual number of NFL players born in each 31-day window.  The orange line? That represents the estimated number of NFL players based on the birth rates for the United States.

The results remain striking:  there are a ton of NFL players born in the first two months of the year, especially compared to expectations. Let’s look at the same data, but plot the difference between the two lines: i.e., actual NFL players minus expected:

As many people have pointed out to me on Twitter, yes, Malcolm Gladwell wrote “something like this” in his book Outliers.  There, Gladwell wrote that there were a disproportionately high number of Canadian hockey players born in January and February. However,  as I wrote yesterday, that reasoning is not applicable here: the thesis behind Gladwell’s story was that in Canada, there is a national cut-off for school and travel hockey teams is based on a calendar year, with January 1 being that date.

It’s a beautiful example of a self-fulfilling prophecy. In Canada, the eligibility cutoff for age-class hockey programs is Jan. 1. Canada also takes hockey really seriously, so coaches start streaming the best hockey players into elite programs, where they practice more and play more games and get better coaching, as early as 8 or 9. But who tends to be the “best” player at age 8 or 8? The oldest, of course — the kids born nearest the cut-off date, who can be as much as almost a year older than kids born at the other end of the cut-off date. When you are 8 years old, 10 or 11 extra months of maturity means a lot.

So those kids get special attention. That’s why there are more players in the NHL born in January and February and March than any other months. You see the same pattern, to an even more extreme degree, in soccer in Europe and baseball here in the U.S. It’s one of those bizarre, little-remarked-upon facts of professional sports. They’re biased against kids with the wrong birthday.

There are four reasons why this explanation does not suffice for the NFL.

• The majority of states have September 1st, not January 1st, as the cut-off, with most having some cut-off of between September 1 and October 31st, and yes, some having January 1st as a cut-off. In Texas, kindergarten eligibility is based on a child being 5 years old by September 1st.  In Florida, 9/1 is also used. And the other big produces of NFL players — California — also uses September 1.  Even if you believe 100% in Gladwell’s work being applicable to this study, then we should be seeing September and October as the two big months, not January and February.
• There is no national cut-off the way there is in Canada: as you can see, each state is different. So even if this effect was true, it wouldn’t be uniform the way it is in Canada.
• Finally, I don’t think Gladwell’s work is applicable here because football is pretty different from hockey, and you don’t have the steep learning curve of skills.  In other words, you aren’t skating. Many of the best NFL players are remarkable athletes who didn’t start playing until high school. These are not people who have been taught the fine art of backpedaling since they were five.
• Even if you disagree with everything I wrote, Gladwell’s theory requires December babies to have really bad chances of making the NFL. We see that is not the case in these graphs.

That said, there’s no denying that January and February do produce a lot of NFL players.  I thought it would be interesting to start the graph from September 1st — the most common cut-off date for schoolchildren — and see how it looks.  This is the graph from above, showing how many more NFL players come from each 31-day window of birth days relative to expectation (based on the U.S. birth rate):

There is also a really sharp decline that is hard to miss.  The 31 days around April 6th have 1,351 NFL players and would be expected to have 1,410 NFL players (-59). Meanwhile, the 31 days around March 3rd have 1,578 NFL players compared to an expected result of 1,421 (+157).   I can’t think of a reason why late February/early March would produce a ton of NFL players, but late March/early April would not.  It is possible that this is just random variation, of course, but over sample sizes this large, that usually isn’t the case.

What do you think?