Wednesday, August 5, 2009

The All-Average All-Stars

Early last month, I wrote this post called "The Most Average Player of Any Given Season", where I tried to find the player whose season statistics across all categories were closest to the league-wide average. This is what I said in the original post:
Using the seasonal statistics found in the Lahman Database, I decided to look for the average offensive line put up by all regular players in any given season. By that, I mean, if you were to take every full-time player in a season and find the average number of at-bats, doubles, home runs, strikeouts, walks, etc. that that group put up, what would that statistical line look like? Furthermore, once you have the average statistical line for that season, is it possible to find the one player from that season who is closest to the average across all stats - to find, in effect, the Most Average Player for that season?
The answer to that question turned out to be Tony Pena in 1984, whose only real deviation from the season average line was in walks (52 walks for the average player versus 36 walks for Pena). The biggest complaint about that post - and it's one that I recognized and that I tried to address in the post itself - was that it seemed unfair to call Pena's season "average" since, as a catcher, his stats were decidedly above-average. It wasn't a criticism I could much deny, except to say that the exercise was from a purely offensive standpoint and it wasn't really trying to account for defense. It's a weak argument, I know.

So I decided to do something about it and take everyone's advice. Using the same data as the initial post, I broke every player into their primary position (ie, the position that they played the most that season) and found the average across each position. Ron, in the comments, called it "an average all-star team."

There are a few caveats about the data. As in the initial study, I limited the pool of players to those who qualified for the batting title. This means that for a couple of positions - catcher and DH, mostly - there are years where very few players meet the criteria. When this happened, I just excluded it from the final results. It made little sense to go back and try to find a suitable cut-off point for those years and, besides, it was consistent with the rest of the study. The other thing to note is that, for years prior to 1996, I don't have any position-specific outfield data. Instead, all outfielders are lumped together. In the output, then, I include three generic "outfield" spots for years between 1961 and 1995. I also include a LF, CF, and RF player for the 13 years since then.

With all that said, here is the list of most average players, by position. Or, if you prefer, the "All-Average All-Stars". For an example of how the points were calculated, see the example in the original piece. Otherwise, just remember that the closer to zero, the more average the player's season was. You can find the spreadsheet here.

Catcher: Ray Fosse, 1971 - 23 points

First Base: Tino Martinez, 1999 & Kevin Millar, 2003 - 23 points

Second Base: Tony Bernazard, 1985 - 30 points

Shortstop: Jose Pagan, 1962 - 30 points

Third Base: Ken Caminiti, 1991 - 24 points

Outfield (1961 - 1995): Bob Skinner, 1962 - 28 points; Dave May, 1971 - 31 points; Dave Winfield, 1976 - 33 points

LF/CF/RF (1996 - 2008): Moises Alou, 1997 (LF) - 33 points; Carlos Beltran, 2005 (CF) - 36 points; Moises Alou, 1996 (RF), 26 points

Designated Hitter: Edgar Martinez, 1999 - 23 points

There are a couple of things worth noting here.
  • First, the 1999 season actually gives us two of our All-Average All-Stars, with Tino Martinez and Edgar Martinez both making the list. I don't know what was in the water that year, but it is pretty remarkable.
  • There's also the fact that Moises Alou makes the list for two consecutive years at two different positions. That's quite the feat.
  • Finally, as was the case in the initial study, it's pretty interesting to see the big names on the list: all-star Ray Fosse, Tino Martinez, Kevin Millar, Ken Caminiti, Dave Winfield, Moises Alou, Carlos Beltran, Edgar Martinez... each of these guys carries some kind of clout and might surprise some people to be found on such a list. My guess is that these guys show up on the list when they're young, old, injured, or otherwise not at their best.
There's always going to be problems with a study like this, whether it's the issues detailed above or something else that I've been lucky enough to avoid so far. All we can ask is that the question we were asking is fair and sensible and that the answer makes sense. I think that's all very true for this post, no matter how surprising it is to see the likes of Winfield, Beltran, and Edgar on the list. If you have any other thoughts on how best to break down this data or another way to approach it, let me know in the comments.

No comments: