Thursday, December 4, 2008

Career length of players by position played

Well, I'll say it: if you aren't reading Joe Posnanski's blog every day, or buying his book, then you are missing the best sports writing there is. I don't think I'm more excited to see a post show up in my RSS reader than I am when I see a new post from Joe.

Yesterday, Joe posted 4,400 words exploring the make-up of the Hall of Fame, trying to walk us (and himself) through its 75 year history, to see if we can get a handle on what it truly is, and not just what it claims to be. The post is informative, well-written, and thought provoking. I highly recommend it.

As good as the post is, though, the comments are just as good, with people chiming in on what makes the Hall special and why certain people were inducted when others weren't. Reading through the comments, I came across this one from Brent, who is apparently a big fan of second basemen in general. He asks this question:
I think 2B is a hard position for the Hall to judge. The only two recent (if the early 80s are recent) are Joe Morgan and Ryne Sandberg (and the dubious inclusion of Rod Carew as a 2Bman). Why is that? Let me propose two possible reasons.
The two reasons he proposes are sensible, and boil down to this: second basemen, due to the unique nature of the base on the field and due to the physical toll one takes turning double plays and the like, play much fewer games than players at other positions. So, when the Hall comes along, their counting stats are inferior to those of other positions, leaving them in the cold.

It seemed like a good theory to me, so I went looking at the data to prove it. For each defensive position on the field, I found everybody who played 1000 games at the given position (it seemed like a decent threshold for a “long-time” player), and then I found the average total games that those players played (for example, Cal Ripken would be on the "SS with 1000+ games" list, but his games at SS and 3B would be added into the average total games, since it shows how long he was able to play the game). I also ran the list for players who played 1200 games at a given position, to try to limit it to longer tenured players. This is the data that I came up with:

…………MIN: 1000 G ……………………MIN: 1200 G
……………
#Avg Games……………#……Avg Games
C…………
109…1406.83………………67…1576.25
1B…………125…1709.25………………92…1823.05
2B…………107…1623.43………………72…1748
SS…………120…1716.86………………90…1862
3B…………100…1667.15………………67…1825.64
OF…………424…1583.91………………288…1754.27
DH…………8………2181.38………………5……2237.6

Taking a quick look at the counts, it seems that the 1200-G group might be the better dataset. The count of players at the infield positions stay pretty steady between the two datasets, but the outfield count seems a little high in the 1000-G group. We would expect the ratio of outfielders to, say, first basemen to be about 3:1, since there are three outfield spots. In the 1000-G group, however, the ratio is a little more extreme. The ratio in the 1200-G group is much closer to our expectations.

Now to look at the games played by position player: Catchers play the fewest games in their career, by far, with more than a season's worth of fewer games played than the next lowest position in each dataset. Besides catchers, the next lowest position is inconsistent between the two, with outfielders taking that ranking in the 1000-G group and second-basemen taking that ranking in the 1200-G group. But we've already established that the 1200-G group is likely the better dataset, so we'll use that. And it does show us that second-basemen play the fewest career games outside of catchers. Outfielders are very close behind, though.

The count of players at each position is also an interesting datapoint. In both groups, the third-basemen give us the least number of players with the qualifying number of games. Does this tell us something about the demands of the position, or is it more a comment on the types of players who play that position (eg, are they more injury prone, etc)? In both groups, however, the count of second basemen is at the bottom of the list.

With the shortest average career besides catcher, and with the fewest number of players even able to play that long, I think it's pretty clear that there's something about the second base position that keeps players from having long careers. Whether its for the reasons Brent brings up - not able to switch to another position, the physicality of turning the double play - or some other reason, I can't really say, but I believe that there's something there.

One final note about the numbers: in both groups, shortstops have both the longest careers and the (second-) most players making the list. I haven't investigated this further to prove this suspicion, but I suspect that this is because both high-offense and high-defense shortstops can hang around for a long time, either as a bopper at another position (Banks, Cal) or as a defensive specialist whose offense can be overlooked (Ozzie, Vizquel). It's good to be the shortstop.

6 comments:

JojoBebop said...

Excellent post, very thought-provoking. I actually followed this link from your comment on Posnanski's blog. Like Brent I have a special affinity for second baseman (played the position in high school).

As far as 3rd base is concerned, one interesting study was done by Hardball Times. Basically, third base is a transient position, so except for the all-time greats who came up on the position (schmidt, brett, chipper, etc.), teams are willing to use it as a weigh station of sorts for middle infielders who've lost a step or young players on their way to the outfield or first base (Miguel Cabrera is a prime example).
So overall you'll see less career third baseman because of the way the position is utilized from a roster management perspective. But because third baseman hit more than second baseman they have more potential to be moved to a offensive first position once they've outlived their useful life (defensively speaking).

Enjoy the blog by the way!

jmeasor said...

Superb investigation ... even for those of us that don't have peculiar interests in middle infielders in particular :-)

Any way to filter out those who shift from SS to 2B? That would limit the listing to those who 'come up' as 2B (Biggio aside) and probably shorten the pool considerably.

All my best,

John

lar said...

@Jojo: Thanks for pointing that out. It makes a lot of sense, and pretty easily explains the numbers that we see in the post.

I was having an email discussion about this with my buddy earlier today, and he brought up a point about second-basemen along that same vein. He said: "when you are young and you are a great athlete or hitter, you don’t play 2B. You play CF, 1B, or you pitch. These players who are at 2B are undersized (generally) and have had to work at their trade rather than be recognized early on for their talent and placed elsewhere."

I doubt it's 100% true, but I believe there's probably a lot of truth in that. And when you couple that with Brent's original point about the bat disappearing and the physical toll of the position, it's easier and easier to see why second basemen just don't play as long and, thus, don't accrue those shiny stats that make a Hall call so much more likely.

@John: Good question. There's probably a couple of ways that could be done, though it might be a little tougher than you think it would. the big question is determining what we mean by "shifted from SS to 2B". Are we talking about someone who spent some time at another position in his rookie year and then moved over sometime fairly early on - think Cal Ripken moving from 3B to SS in 1982? Or are we talking about someone like Pokey Reese, who played his entire first year at SS before being moved around and finally ending up at 2B? Basically, we'd have to decide on a threshold for games played at SS early in his career that would cause us to drop him from the list...

I'll noodle it around. Let me know if you have any thoughts.

Anonymous said...

Also saw Brent's comment on Posnanski's blog. I thought that it had some merit, but your data (1200 games) seems to suggest that there is little difference between 2nd and the outfield in longevity.

lar said...

That's a good point, and without delving further into the numbers (maybe later... I'm at work) I can't really say.

I'm wondering if maybe it's easier to play a lot of games in the outfield, and so even the 1200-G threshold isn't high enough (ie, there are too many people who barely made the cut for other reasons - like not being good enough to have a longer career to begin with - bringing down the average).

I'll try to look into it.

lar said...

I ran the data again, limiting the groups to players with 1300 and 1400 games at the specific positions. Here are the numbers:

………………MIN: 1300 G ……………………MIN: 1400 G
………………#……………TOT G…………………………#……………TOT G
C……………48…………1696.13……………………34…………1816.24
1B…………82…………1850.49……………………66……………1941.76
2B…………59…………1829.68……………………45…………1934.13
SS…………78…………1918.54……………………66…………1977.86
3B…………60…………1854.6………………………46…………1958.65
OF…………229………1848.46……………………180………1948.31
DH…………4……………2233.25……………………3……………2217

As you can see, as we limit the group to players with longer careers, our pattern of catchers and second basemen having the shortest careers remains. The outfielders' average career length also separates itself some from the second basemen, though not by a whole lot.

I can't say for sure if we've cleared up the issues with the outfielders' numbers (if there was an issue), but I think restricting the results helps give us a clearer picture. For the 1300-G and 1400-G results, the ratio of outfielders to other positions stays at a healthy 2.7:1. The prior datasets are a little less steady, with a 3.4:1 ratio for 1000-G and a 3.1:1 ratio for the 1200-G dataset.

Regardless, second-base keeps coming up at the bottom of these lists (besides catchers). It seems pretty clear to me that there is something about the second-base position that keeps players from having long careers.