Oakland roster turnover
Here's a neat post by danmerqury at Athletics Nation (one of my favorite AN writers, by the way -- and there are many days when I'd say the only AN writer I make it a point to read) about whether the A's have more roster turnover than most teams. Answer: not at all. By the measure used, the A's are dead in the middle of the league.
Now, that measure (average tenure of Opening Day 25-man-roster members) isn't the only one that could be used, but it's a pretty good one. A team with a bunch of four- and five-year team-veterans on it (like the Yankees or Angels) will be a team that feels like it's been together forever. A team with a bunch of AAAA players and recent trade acquisitions like Pittsburgh will feel like whoever happens to be on he roster that day, with no real connection to the team itself.
There are some tweaks possible, of course -- Opening Day is an arbitrary point, and in fact something like June 1, where you've gotten back the guys from offseason injuries, the position battles have shaken out into the group of players that'll be with the team, and so forth, might be better. Still, it likely wouldn't change the results all that much.
Another point might be to weight the question by playing time. Every team has a certain amount of churn at the edges of the roster, as utility players and sixth men in the bullpen are all interchangeable and aren't who you remember when you think back to the teams of your youth. I remember Matt Stairs and Ben Grieve and Scott Spiezio, not Kurt Abbott and Jay Witasick. The A's would probably still do better than you'd think after accounting for playing time -- even though the high-profile players who leave (Giambi, Tejada, Hudson, Mulder, Zito, Ramon Hernandez) are the high-usage players, Oakland tends to keep these guys for a number of years before sending them on their way. It's churning through Ryan Christenson and Terrence Long that kills your numbers, not losing Jason Giambi after six years by the bay.
Beaneball by Jason Wojciechowski is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.