Earlier this year I read on a college football blog (or was it a message board?) that home field mattered a lot, "especially in conference games." There's not a lot of unambiguous evidence that it really does if the teams are evenly matched, as I conjectured about this time last year.
It is true that over time the home team wins about 60% of the time in almost all division 1 sports, but before you draw any conclusions from that you have to take into account that the correlation may be to be something other than the game location.
In Division 1 sports, games between teams that have a large difference in "strength rating" (whatever that is) are predominately played at the stronger team's location. Most such games would be won by the stronger team no matter where they're played.
It is difficult to sort out the real correlations in football or any other sport where teams only play at one team's home in a given year because of the large number of changes in personnel from year to another and the paucity of home-and-home contests in any given two-year period.
One way to look at a single year is to compare scores from only games that involve teams from the same "level." In the table below we consider just games where both teams are from the Bowl Championship subdivision. The column labeled HFA* is just one half the difference in home team score and visitor's score. That's not really a meaningful number, but it illustrates the point (and is surprisingly close to Sagarin's more sophisticated definition.)
|
The overall home winning percentage is pretty close to the 60 percent usually cited, and the overall "home field advantage" pretty close to the 3 points many use. But if we limit our comparison to conference games, the split is closer 55-45, and the HFA* only about a point and a half.
It would appear from the data that the "home field advantage" is not as large in general as we might think, and probably only comes into play when teams are very evenly matched.