Posted by Doug on October 2, 2006
Longtime readers of this blog know that a guy with the handle JKL always has interesting things to say in the comments. In particular, he did a couple of very interesting home field advantage studies awhile back. I asked him if he'd be willing to write those up and let me post them here. He agreed, and so I'll be posting them today and tomorrow. Thanks JKL!
Kansas City and Denver are generally considered to have two of the best home field advantages in the NFL. More often than not, this is attributed to, in Denver's case, the altitude, and in Kansas City's case, the environment ("crowd noise") at Arrowhead Stadium.
But is there another explanation? Denver and Kansas City are more geographically isolated from their conference opponents than any other franchises in the NFL. In fact, they are each other's closest geographic neighbors, separated by approximately 567 miles and 4,000 feet in elevation change. Denver's next closest AFC opponent is division rival San Diego. Kansas City's next closest is non-divisional Indianapolis. Compare this to Pittsburgh, who is within 300 miles of all 3 divisional opponents, and within 500 miles of several conference opponents (Buffalo, Tennessee, New England, New York Jets, Indianapolis).
To examine this more closely, I looked at all divisional series played between 1986-2005. With a few exceptions from the 1987 strike season, all divisional series have been played on a home and away basis over that span. This method was chosen because it allows us to compare 2 games, at different sites, each year, to see which series feature a high percentage of home team splits, compared to those that feature more sweeps or series splits where the road team wins both times. Obviously, if a Super Bowl contender is playing a first overall pick contender (like Indianapolis and Houston last year) it is going to make little difference where the game is played. The hope is that over the course of 20 seasons, things will even out, and the series will show roughly even distribution of years where the teams were relatively close, and those were one team was significantly better than the other. I was not concerned about which specific team in the series was winning more, but rather, how often the home team in the series won.
I then looked up the geographic straight-line difference (in miles) between each city in a series (as opposed to the driving distance) because I wanted to get a good approximation of travel time by air on a charter plane. With the exception of Green Bay playing down the road in Milwaukee, I excluded those years when one team played its games in two or more cities, such as New Orleans of 2005. Below are the winning percentages of the home team in divisional series, sorted by geographic mileage between the two cities.
Distance W L T PCT within 400 363 293 0 0.553 401-800 248 205 0 0.547 801-1200 332 235 1 0.585 1201-1600 172 102 0 0.628 1601-2000 56 42 0 0.571 2001 or more 74 55 1 0.573
Conventional wisdom is that distance matters at extreme distances such as cross-country trips. However, the series beyond 1600 miles, which consist of pre-2002 series involving Arizona in the NFC East, and San Fransisco and Los Angeles versus the other old NFC West opponents, show no greater home field advantage than those at, say, 1100 miles difference.
For some concrete examples of the effect of short distances, here are the divisional series that have been played between 1986-2005 where the teams are within 300 miles, with the miles separating the teams, followed by the home team's won-loss record:
team 1 team 2 distance (mi) Home W Home L nyg phi 90 22 18 rai (la) sdg 108 7 11 cle pit 125 19 15 phi was 133 23 17 chi gnb 167 17 23 rav (bal) pit 198 11 9 cin cle 210 19 15 nyg was 219 21 18 atl car 228 15 7 cin oti (ten) 245 5 5 chi det 247 26 13 cin pit 253 19 21 clt (ind) oti (ten) 260 4 4 gnb min 266 26 14 det gnb 283 26 14
Notice the winning percentages in the series involving one dome team against an outdoor team, versus those involving two outdoor teams. In Outdoor only series, the home team was 181-174, a paltry 51.0%. In series involving a dome team, the home team was 97-52 (65.1%). Basically, in outdoor series where the teams were relatively close geographically, the home team has had virtually no home field advantage. This has been borne out again this season, as New England won at New York Jets, Cincinnati won at Pittsburgh, Chicago won at Green Bay, and the New York Giants won at Philadelphia. Does this mean the road victor is going to cruise to victory at home later in the year in these series, because they already won on the road? If history is any indicator, the answer is no.
It also raises the interesting question of climate, and how much of what is showing in this distance data is actually due to climate differences, rather than the effect of travel time. Most of the dome series results are from the current NFC North teams, where the difference in playing temperature between playing outdoors and playing in a dome can be very significant.
Tomorrow, I will look at climate differences, and also put the old Black and Blue Division under the microscope. The NFC North teams provide a good case study because not only are two teams outdoor cold weather teams, and two teams indoor controlled weather teams, but because all four teams used to play outdoors. Thus, we can see the effect that changing to domes has had on home field advantage in those series.