A thought experiment on aging…
Suppose I play in a corporate basketball league with 4 teams. Let’s say all the teams are about equally talented. The league has been going, more or less unchanged, for a long time. Then one season, one of the teams decides to quit the league, but the three other teams all return and with equally talented rosters. Importantly, assume the average age of the returning teams remains unchanged (e.g. they swap in some younger players and everyone else ages by a year). Now imagine that the lost team is replaced by one made up entirely of recently retired NBA players. What will the league look like the following season?
Well, quite obviously, all the non-NBA players will look worse overall! Instead of playing all their games against equally matched teams, each returning team will play one third of their games against a much more talented squad. In those games they’ll all more points, score less, and record fewer rebounds, steals, blocks, etc.
Now let’s say that we wanted to look at this league to come up with aging curves. This season I’m 38 years old, and I played last season too (my age 37 season). To figure out how my performance changed from age 37 to 38 the standard thing to do is to just look at my performance over the two years and take the difference. Let’s say I was about average in my age 37 season, and that I’m still about average among the non-NBA teams at age 38. My stats will show a huge decline for me from age 37 to 38! But since I was average at 37, and still average among returning players at 38, my skills probably haven’t declined much at all – recall the assumption that the average age of returning teams and their overall talent level is unchanged. No, my skill level remains the same – what has changed is the league context.
Okay, so what? When we’re trying to forecast player performance, which is what aging curves are about, we have to deal with the context we have, not some hypothetical alternate context: if talent is increasing, then a formerly average player becomes below average – blaming everyone else rather than the player himself doesn’t change the reality. However, I think there are good reasons for trying to separate the physical effects of aging from competitive changes.
Of course saying it is one thing… doing it is another. But I have some thoughts on that too. To be continued…