Fun with home/road splits: Dunedin Blue Jays

Something notable happened in Dunedin last night. The D-Jays blew a 8-4 lead after seven innings, falling 9-8 to the Palm Beach Cardinals, which though a painful way to lose is hardly unprecedented. But it was just the fourth home loss for Dunedin in 25 second half games.

The Florida State League splits its season into 70 game halves, with the first half and second half champions in each division qualifying for the playoffs (if the same team wins both halves, it's the team with the next best full season record). The D-Jays were leading the first half standings until a nine game slide starting May 30th sent them spiraling, and they ended the first half a mediocre 34-35.

The second half didn't start much better, with a 2-5 road trip. But then they returned home and swept Daytona, and this set the tone for the second half. At home, the D-Jays have been almost unbeatable, with a 21-4 record (0.840). On the road, they've struggled to avoid being swept, posting a record of 8-19 (.296). The difference in winning percentage is an incredible 0.544!

Of course, we know that home teams have an advantage in baseball, with MLB home teams generally winning about 53-54% of the time. So we should naturally expect a small differential. However, that effect should be more muted given (1) sparse attendance across the FSL, with Dunedin by far the lowest ; (2) lack of geographic differences influencing play; and (3) lack of ability to tailor a team to the home park, given that players are constantly moving through.

It's also a pretty small sample size at less than 30 games of each, so I was curious as to the statistical significance. It turns out that the difference is statistically significant at just shy of four standard deviations (z=3.94 to be precise). That would put the chance of this occurring randomly at something like 1-in-25,000 assuming there was an equal chance of winning at home on the road. Even adjusting for something like a conventional home/road split, it's still still highly significant at over 3 standard deviations, making it around a 1-in-20,000 event.

What makes in even more fun is comparing to the first half. At home, Dunedin was 17-18. On the road, 17-17. In other words, about as neutral as it gets. MiLB doesn't seem to have the 2015-16 standings readily available, but looking at the five years before that (2010-14), Dunedin had a .573 winning percentage at home, and .510 on the road. Which is right in line with normal.

Is there anything to this? Probably not, it's just one those huge outliers. I do like it for one reason. Dunedin only has their own broadcasts for home games, and when the D-Jays win they play the full "OK Blue Jays" at the end, which I find enjoyable. It's been played an awful lot the last 50 days.