Quick review: league-wide wRC+ distribution

When we asked the Talking Chop commentariat what they wanted to see in 2021, one item that stood out to me was for a refresher of sorts regarding what’s considered “good” or “bad” with respect to some of the commonly-cited baseball metrics. One obvious place to look for answers for this query is the Fangraphs glossary, which includes neat tables like this for most stats:

https://library.fangraphs.com/offense/wrc/

So, one way to provide this refresher is just to say, “Go look at the Fangraphs glossary.” But, I think we can do a little better, in a way that a glossary page can’t. That’s what this post is about, brief as it is.

First: what the hell is wRC+ and why are we talking about it? Well, you could just go to the glossary. But if you won’t, it’s essentially the aggregate offense measure on Fangraphs. It answers the very basic question of, “After adjusting for home ballpark, how much offense does the player produce per plate appearance, based on his outcomes?” Each integer above/below 100 indicates a one percent increase/decrease in productivity: a 110 wRC+ means a player was 10 percent more productive, per PA, than a league-average hitter; a 75 wRC+ means a player was 25 percent less productive, per PA, than a league-average hitter. This is, honestly, way better than most other offensive metrics: average, OBP, and slugging all tell only part of the story; OPS is a mathematical frankenstein. wOBA attempts to combine things and put them on an OBP scale so it at least looks like the same three-numbers-after-a-decimal number people are familiar with, but it A) has no adjustment for park, and B) honestly, saying someone has a .330 wOBA is meaningless unless you know what league-average wOBA is. wRC+ solves these issues; you never have to look up league-average anything to understand wRC+, since it’s scaled to 100 by definition.

Note that wRC+ is based on outputs; “luck” a player experiences on balls in play will show up in wRC+. Further, wRC+ is not the only attempt to make a holistic, scaled-to-100 offensive metric — but it is the one that is most clearly tied specifically to outputs, and its derivation is pretty not black box-esque compared to others. (You can always use Baseball Prospectus’ DRC+ as well.)

In 2019 (we’re ignoring 2020 because, well, it was 2020), 635 position players had at least one plate appearance. Their wRC+s ranged from -100 (the worst possible value) to 216. Remember that by definition, league-average wRC+ is 100, and the weighted average wRC+ for these 635 is 100... because it has to be. Yet the straight average is 79, and the median is 88. What gives? This might be obvious and not worth saying, but essentially, better hitters play more. They’re overrepresented in the total share of plate appearances.

Why does this matter? Well, one main reason. A lot of times, I find myself asking, “where does this guy’s [insert stat here] rank among all MLB hitters?” The thing is, as we can see from the numbers above, comparing [stat] to all MLB hitters is probably not a good idea, since what you don’t want to compare anything against a bunch of guys getting only a few PAs, most of whom aren’t very good and not really deserving of more. Anyway, the answer to this question is embedded in the actual visuals that show the league-wide wRC+ percentiles, so just take a gander:

(Qualified means “qualified for the batting title, the cutoff for which was 503 PAs in 2019.)

Long story short: 200 PAs is the cutoff you want to use. This has been consistent for a while now; I’ve been using the 200 PA cutoff for about a decade now due to doing this exercise a while ago, and every time I do it, it’s still 200 PAs. It’s also 200 PAs in 2019. In short: if you want to compare a player’s percentile rank against other players, use a 200 PA cutoff, since that one ensures that the median player is an average hitter. So, there you go.

One more parting shot, just because I thought it was funny. One key thing the above implies is that, well, many hitters that get at least one PA during a season are not very good. What if we wanted to look at the breakdown of how not-good, in terms of buckets? The buckets are purposefully uneven, because I think this is an amusing way to look at it:

This is the same data as above, just in terms of “what proportion of players have a wRC+ in one of these buckets, under the given PA cutoff.” You can see that under this very specific definition which is not an appropriate definition of buckets and I chose it because it was amusing, it doesn’t actually look very normal. If anything, players get kind of bunched towards “worse than average” until you get nearly full-timers at 400+ PAs, where “better than average” takes over. Again, the point is just that most hitters aren’t actually particularly good or even average, and the class of MLB hitters that is average-or-above is generally a minority.

Anyway, now you know. Pitchers and FIP-/xFIP- are next, but those are much thornier problems, because a hitter is a hitter who gets PAs, but pitcher roles and innings are pretty finicky.