FanPost

Market Size, Divisional Familiarity, and All-Star Selection

After watching Joe Buck and his minions forget to mention any Brewers during Tuesday's All-Star Game, lest he miss an opportunity to talk about Numbah Two, I remembered that I'd started doing some All-Star roster analysis for a fanpost and then forgot to finish it up. So here it is!

I initially set out to look at two factors for who becomes an All-Star: home media market size and the manager's division. Unfortunately, I couldn't find solid info on who was voted in by the players and who was a manager's selection, so that part will be a little inconclusive. I do have a nice trend for market size, however. Here, MLB media markets (excluding Toronto and Montreal) have been ranked by size, 1 (New York) to 25 (Milwaukee). The data points seen here are average ranks for fan-voted players (including the "final vote" winner) and the other players on the roster, for 2003 to 2012 (the first year the game "mattered" through the last year of the previous divisional alignment).

Markets_zps717c3641_medium

A couple clear trends pop out. First, the players' vote/manager selection group is pretty stable. That group is from a slightly bigger than median market on average, but in no year is the average higher than 10 or lower than 12. The fan vote is much more volatile and on average in favor of players from bigger markets. Only twice in 10 years did the fans' choices come from smaller markets than the rest of the rosters, on average.

Some of this certainly has to do with which teams are good in which years. 2006 featured the biggest market votes by the fans, and that year both the Mets and Yankees had the best records in their leagues; four Mets and two Yankees were voted in that year (and in one of my favorite bits of ASG trivia, Carlos Beltran played the entire game).

Meanwhile, when it comes to those that aren't voted in by the fans, I've long wondered whether familiarity with the players in one's own division makes managers more likely to select those guys. As I mentioned, I can't separate out the players' votes from the manager's selections, so this is basically offered without much comment. This is the average number of non-fan voted players per team in the manager's division compared with the other two divisions.

Divisions_zps09b8a086_medium

At best, there's significant noise here that can't be parsed out. However, nearly all instances of noticeable differences are in favor of the manager's division getting more players on the roster. In both these examinations, the N is too small to worry about statistical significance, so these charts are for entertainment purposes only.