Monday, January 04, 2016

Statistical Error

I got annoyed by a tweet this morning. Here it is.


You can see I replied a couple times, but he only engaged once. Which is fine. If I was in his shoes, I wouldn't have answered me at all, so it was good of him to bother.

The chart is basically a tease for his article on the Hall of Fame and most of it is considering which active players might get in. It's worth reading.

But that chart... is very misleading.


There are a couple reasons for this. One is that there are a lot of players whose careers go back through most of the 90s and who aren't yet eligible for HoF consideration. The other is The Veteran's Committee.

It's mostly the latter. The Veteran's Committee has added a lot of players over the years who weren't able to get in on the standard voting process. Their voting process is complicated, but they essentially aren't considering anyone who started playing later than 1986. So that's my cutoff year.

So let's look first at the same chart, 1900-2009 but without the Veteran's Committee additions.

Well, the first thing to notice is that the early part of the century suffers. No surprise, the HoF wasn't around then, so they're mostly committee selections. But more interesting is that the latter decades exceed the 20s and 30s. Let's cut it off in the mid-80s to account for recent players not yet eligible.

How about that? Turns out that the recent "decline" is a product of the lag between playing time and election or selection. The original chart is technically correct, but the conclusion is erroneous since no account was taken for why the later years differed so dramatically from the earlier. Odds are, we're watching (and have been watching) just as many HoFers as dad and grandad did.