In my effort to avoid doing my actual research, I'm still doing a bit of a baseball research on homefield advantage. I have a link in my signature to my articles from a few weeks ago at TGP about it. Today I decided to check out whether home teams won certain number games in a series more often than other games in a series. As it turns out, I got a very interesting result:

From 2004-2007 (9,719 games):

First game of series: 53.1%
Last game of series: 53.0%
Middle games of series: 55.7%

This is VERY statistical significant. What on earth could be the reason that home teams win more games in the middle of series than at the end of beginning? In case anyone is wondering, there is virtually no difference in between the home team winning the first game of a series when they are coming back from being on the road compared to when they were home for their previous series.

I'm really baffled, but the odds of getting a result this extreme by random chance are about 1/70. I have to go out for the evening, but I'd love if anyone could give me any thoughts on this to look at when I get back. Thanks.

Were these only 3 games series with 1, 2, 4 and 5 game series not included in the study?

lethal wrote:Were these only 3 games series with 1, 2, 4 and 5 game series not included in the study?

it's all series that took place between 2004-2007 regular seasons.

Dunno. Maybe there were enough "just coming off the road" (mad hot luvin with the wifey) 1st games of series and "hittin' the road tomorrow" (OMG, did I remember to pack my teddy bear?) last games of series to account for a 2% diff?

Just saying that maybe, if the series is either preceeded or proceeded by travel, the players are a little more "settled in" for the sandwitch games.

My thought is that the opposing team's players are much more likely to hit the town on a Friday night than on a Thursday night--when they are travelling--or on a Saturday night--when they have to be at the stadium early Sunday for a day game. Similar logic could apply to many midweek series, as well.

i wouldn't even have known that a 2% difference was statistically significant unless our resident statistician had pointed it out. good stuff. more analysis, plz.

td11 wrote:i wouldn't even have known that a 2% difference was statistically significant unless our resident statistician had pointed it out. good stuff. more analysis, plz.

2.7% difference is significant it is if you gather 4 years worth of data. think about it-- if the dataset were just 2 games, going 0-1 in the first game of the series and 1-0 in the second (a 100% difference) isn't significant. but if it happened every series for a decade it would be.