Posted by Doug on June 6, 2007
A couple of weeks ago I threw out a crazy theory that possibly some teams might be at a slight disadvantage late in the season because of the peculiar characteristics of the cities they call home.
I was thinking primarily of heat and possibly altitude, but Commenter Extraordinaire JKL suggested a different theory:
I might suggest a different possibility, rather than altitude. Denver has 2 major factors that make it unique in the NFL. The obvious one is the altitude. The other one is that Denver is the only team within the Mountain Time Zone, so every time Denver plays a road game, it is in a different time zone. (And conversely, every time a team comes to Denver, it is playing in a different time zone).
So I decided to check and see if road teams playing in different time zones tend to fare worse than road teams playing in the same time zone. I took every regular season game since 1978 and ran a regression of road team's point differential versus the following input variables:
1. the difference between the road team's win total for that year and the home team's win total for that year. Technically, to account for strikes and ties, I used win percentage times 16 instead of wins. And I threw out the game in question.
2. the absolute time zone difference between the road team's city and the home team's city. E.g. when Denver travels to Miami, that's a time zone difference of two. When Jacksonville travels to Miami, that's a time zone difference of zero.
The time zone difference variable had a coefficient of -.14, which would mean that a trip from one coast to the other would cost you about .42 points. However, the coefficient was not statistically significant (and wasn't close to being statistically significant), which means that, given the 30 years of data we looked at, we have no real reason to suspect that the true coefficient isn't zero.
OK, so it looks like there's probably not much there. Doesn't hurt to check.
But then I forgot about the time zone thing and just started playing around with the data a little, and I noticed something:
Week 1: road teams lost by an average of 2.78 points.
Week 17: road teams lost by an average of 4.22 points.
Now, that split suggests, but doesn't necessarily prove, that there is a general sort of "road wearying" affect on teams. The later it is in the season, the harder it is to win on the road. To be a little more rigorous, I ran a regression of the road team's point differential versus the road team's win difference (as above) and the week number.
The week number turned out to be highly significant (p=.006, if you're into that kind of thing). Here is the equation:
Pt. differential =~ -2.19 + 1.22*(win diff) - .0924*(week)
So if two evenly-matched teams (windiff=0) are meeting in week 1, the road team should be about a -2.28 point underdog. If those same two teams meet in week 17, the road team should be a 3.76 point underdog. Interesting.
Just because it's June, and because I enjoy a good conspiracy theory, let's calculate which franchises have been most helped and hurt by their schedules since 1978. As you'll see, the NFL seems to have a slight preference for playing late-season games in domes and in milder climates. This means that teams like Buffalo, Green Bay, and Chicago tend to play slightly more early-season home games and thus slightly more late-season road games, whereas the California and Florida teams have the reverse.
Assuming that every passing week does indeed cost the road team about a tenth of a point, here is a list of how many points the schedule has helped or hurt each team from 1978--2006. A positive number means the team has been helped.
San Diego Chargers +19.4 Hou/Ten Oilers/Titans +15.7 San Francisco 49ers +15.5 Oakland/LA Raiders +15.2 Miami Dolphins +14.6 AZ/STL Cardinals +11.4 Seattle Seahawks +11.3 Dallas Cowboys +9.8 Detroit Lions +7.0 Baltimore Ravens +6.3 Tampa Bay Buccaneers +6.1 St. Louis/LA Rams +4.3 Carolina Panthers +2.4 Minnesota Vikings +2.0 Philadelphia Eagles +1.9 Cincinnati Bengals +1.5 Houston Texans +1.1 Indy/Balt Colts +0.0 New York Jets -1.2 Jacksonville Jaguars -1.3 Atlanta Falcons -1.8 New Orleans Saints -2.5 Pittsburgh Steelers -3.9 Washington Redskins -5.5 New York Giants -7.0 Kansas City Chiefs -7.1 New England Patriots -9.1 Denver Broncos -13.4 Cleveland Browns -13.5 Chicago Bears -19.0 Green Bay Packers -26.8 Buffalo Bills -33.4
Check out the Bills' schedule by quarters of the season:
Games 1--4: 60.2% home games
Games 5--8: 52.8% home games
Games 9--12: 50.0% home games
Games 13--16: 37.0% home games
This year, the Bills play four of their last six on the road.