How Important is Age?

Anyone over the age of nineteen knows from personal experience that age is a relevant factor in one's ability to perform physical (and mental) tasks. In this article, I'll examine the aging process and its effects on the physical (and mental) task of piling up fantasy points.

In order to do this, I cracked open my season-by-season database, which contains complete career data for every player who was active in the NFL in 1998 or later. Then I proceeded with the following plan:

  1. Find all seasons in which the player in question was age N, and note his performance. The metric I'll use is Fantasy Points, defined as follows For example, in 1996, Terrell Davis was 24 years old, and he tallied 275 fantasy points.
  2. Note how the same player did the next season. Continuing with the same example, Davis, at age 25, racked up 294 fantasy points in 1997. Technical note: I only considered players who scored at least 50 fantasy points in one of the two seasons.
  3. For each N, count up how many players were better at age N than at age N+1. In the above example, Terrell Davis improved from age 24 to 25, so he serves as evidence that running backs perform better at age 25 than at age 24.

Let's just get to the results for running backs before we talk about what it all means.

Running Backs

           TOTAL      BETTER AT    WORSE AT      PCT
N         NUMBER       AGE N+1      AGE N+1    IMPROVED
-------------------------------------------------------
21          11           5             6         45
22          43          23            20         54
23          69          40            29         58
24          77          44            33         57
25          62          32            30         52
26          57          31            26         54
27          39          20            19         51
28          34          13            21         38
29          26          10            16         38
30          19           7.5          11.5       39
31+         12           4             8         33
Just to be sure we're all on the same page here, look at the 24 row. It says that I found 77 24-year-old running backs in my database. 44 of them did better at 25 than 24, and 33 did worse. Thus 570f all the backs I considered improved from age 24 to age 25.

Sure, it's true that many backs in the study improved or decline for reasons having nothing to do with their age -- injuries, changing teams, a change in surrounding personnel or coaching staff, etc. These are all things that this study does not account for. The idea is to collect enough data that these external factors even out (e.g. the same percentage of 24-year-olds changed teams as 28-year-olds). I'd be more confident in the results if I had more data, but for now this is the best I can do.

Let's organize this into bigger groups in order to extract the general trend:

             TOTAL      BETTER AT    WORSE AT      PCT
N           NUMBER       AGE N+1      AGE N+1    IMPROVED
-------------------------------------------------------
under 27     319          175           144        55
27            39          20            19         51
over 27       91          34.5          56.5       38 
This tends to suggest that the peak period for running backs is age 27-28. As a group, running backs under 27 tend to improve, and running backs over 27 tend to decline.

Note that these results describe a general phenomenon, and are not to be taken as a prediction about any individual running back. Saying that 60% of all RBs Emmitt Smith's age decline the next year is not the same as saying that Emmitt Smith has a 60hance of declining. To see the difference, consider: about 400f all people will become pregnant at some time in their life. Does that mean you have a 40hance of becoming pregnant? Not necessarily. There are some factors inherent to you that might explain why your particular chance is greater than or less than that of the rest of the group in general.

Back to Emmitt Smith. 600f all backs in his age group tend to decline, but there may be factors inherent to Emmitt that will cause him to buck the trend. If you think you can identify some factors like that, by all means don't be afraid to project improvement for Emmitt.

The bottom line is this: this study tells me that there is a general force pulling Emmitt down. Based on that, I refuse to project improvement for Emmitt unless I can think of specific reasons why Emmitt's situation is different from that of a typical 30-year-old RB (turning 31) -- maybe a couple of new offensive lineman, maybe a new system that will get him more touches, maybe he's healthy this year after fighting through injuries last year -- things like that. With respect to Emmitt in particular for this year in particular, I don't see any real reason why he'd buck the trend. Thus, I expect Emmitt to decline. You may come to a different conclusion, but just make sure you have a good reason for doing so. Also note that Emmitt may decline and still be a very valuable back. By no means am I telling you to cross him off your list.

One more note before we leave running backs. It's not at all clear to me how much of the "aging" process comes in the form of the physical pounding running backs take, and how much is just general age-based decay that we all experience. For example, Stephen Davis is 26 this year, but it's possible that his body thinks he's more like 24 because he hasn't taken nearly as many hits in his NFL career as most backs have by the time they're 26. That's not an unreasonable assertion. On the other hand, studies have shown that baseball players (hitters, anyway) peak right around the same time -- age 27 or 28 -- and they don't take any physical pounding. It's a tough call, but my opinion (i.e. guess) is that the pattern we see here for RBs is due to regular old aging and is not being unduly accelerated by the particular physical requirements of the job. I'm inclined to treat Davis like any other 26-year-old, but I won't try to change your mind if you think otherwise.

Receivers (including TEs)

           TOTAL      BETTER AT    WORSE AT      PCT
N         NUMBER       AGE N+1      AGE N+1    IMPROVED
-------------------------------------------------------
22          48           35           13         73
23         103           71.5         31.5       69
24         117           63           54         54
25         113           60           53         53
26         109           53           56         49
27          82           35           47         43
28          61           30           31         49
29          52           26           26         50
30          34            8           26         24
31          24            7           17         29
32+         33           10           23         30
The trend here is not quite as clear in terms of finding the peak, but note the cliff at age 30. Let's try slicing it this way.
              TOTAL      BETTER AT    WORSE AT      PCT
N            NUMBER       AGE N+1      AGE N+1    IMPROVED
----------------------------------------------------------
under 24       151         106.5        44.5        71
24-29          534         267         267          50
over 29         91          25          66          27
This is a much different picture than what we saw for running backs. Young receivers are more likely to improve than young RBs. My guess is that this occurs because young RBs are more likely to already be good than young receivers are. Old receivers are more likely to decline than old RBs. I can't think of a good explanation for this, but the data is compelling.

Did you know Marvin Harrison will be 28 this season and Randy Moss only 23? Something to think about for you dynasty league types.

Quarterbacks

I expected to find no clear trends at all for QBs. Let's see.
           TOTAL      BETTER AT    WORSE AT      PCT
N         NUMBER       AGE N+1      AGE N+1    IMPROVED
-------------------------------------------------------
22           7            6            1         86
23          33           24            9         73
24          49           28           21         57
25          51           24           27         47
26          46           25           21         54
27          41           18           23         44
28          35           21           14         60
29          32           12           20         38
30          30           14           16         47
31          29           14           15         48
32          24           17            7         71
33          19            8           11         42
34          16            5           11         31
35          21            7           14         33
36+         26           11           15         42
It's clearer than I anticipated.
              TOTAL      BETTER AT    WORSE AT      PCT
N            NUMBER       AGE N+1      AGE N+1    IMPROVED
----------------------------------------------------------
under 25       89           58           31         65
25-32         288          145          143         50
over 32        82           31           51         38 
This looks fairly similar to the receivers, except that the peak appears to be a little longer and a little later. Again, it's no mystery that QBs under 25 tend to improve. They typically start off so bad that there's nowhere to go but up. Then, there's about an 8-year stretch where age doesn't seem to be much of a factor. Finally, while there always seems to a few oldsters out there putting up solid numbers, the decline phase does kick in around 32, so remember that those guys are the exception rather than the rule.

Short Summary:

According to this data, the word "old" means 28 for a running back, 30 for a receiver, and 32 for a quarterback. This is probably not a surprise, as it squares fairly well with conventional wisdom.

Old players, as a group, will decline, although the phenomenon is far from universal. In fact, the age factor is much weaker than I would have guessed, and should probably rank relatively low on your list of considerations when projecting players. Still, it is there, and while it's not uncommon for old players to improve, it's probably not wise to count on it, unless you have a good reason for doing so.