Pro Football Reference Blog

In one of the greatest Super Bowl upsets of all time, the upstart AFL Champion New York Jets beat the NFL Champion Baltimore Colts, 16-7, in Super Bowl III, to give the AFL its first Super Bowl victory and instant credibility. Let me give you an explanation you've never heard to make sense of such a game. The location was the Orange Bowl, where the Jets had played just four weeks earlier in the regular season finale against Miami. The Jets team played there in '67 and '66, too. Baltimore? They had never played a game at the Orange Bowl--never even played a regular season game in Florida.

Fast forward to last weekend, when the Arizona Cardinals traveled to Carolina, where they had lost by 4 points in the regular season. The Cardinals had struggled over the final month of the regular season, were 0-5 in road games played on the East Coast in 2008, and would be without WR Anquan Boldin. They weren't given much of a chance on the road, and entered the game as a sizeable double digit underdog. Two years ago, in my very first post as a member of the "staff", I predicted that the Arizona Cardinals would enjoy an increased home field advantage in 2007, with the rationale being that teams in new stadiums enjoy a familiarity advantage over road teams who have never visited the stadium before, particularly after the home team has had some time to develop its own comfort with playing in the new stadium. Over the last two seasons, the Cardinals have gone 12-4 at home (and one of those losses in 2008 was to a team who had previously played in AZ) and only 5-11 on the road, for the largest home/road differential in the league over the last two years.

In the first round of the playoffs, thanks to the rule that gives division winners with worse records the higher seed, the underdog Cardinals drew another first time visitor to University of Phoenix stadium, the Atlanta Falcons, and jumped on them early en route to a 30-24 victory. But the divisional round game against Carolina was on the road. Yet, despite that, familiarity may have still given the Cardinals an assist in reversing the regular season fortunes. Now, they return home to face another team that has never played a game at the new stadium, the Philadelphia Eagles. The winner will become the second team to advance to the Super Bowl with 9 wins, since the league went to a 16-game schedule. Oh, and by the way, the last team to do so, the 1979 Los Angeles Rams, advanced by avenging regular season road losses at Dallas and Tampa Bay. The Rams then played the heavily favored Steelers in the Super Bowl, not in their home stadium, but a few miles away at the Rose Bowl, and had the lead entering the final quarter before the Steelers pulled away late.

Now I won't be so naive as to tell you that familiarity with a stadium is outcome determinative. But I think it's a small, and more importantly, underappreciated part of determining who wins games (or who stays competitive in them). Why is it a factor? Well, it could be related to many different things, all which may come into play. Field surface and conditions could be just one of the factors. Wind patterns (created by unique features of a stadium) could be another. Sound conditions created by each stadium design could be another. Weather patterns and climate is certainly another strong one, where teams from similar playing climates are comfortable and used to playing under the conditions, while those from disparate climates are not. Sun conditions and lighting patterns in the stadium. Heck, the color of the locker room, types of showers. Anything that might be different could come into play, almost imperceptibly, to cause sub-optimal performance.

I've cited three examples where it may have helped some underdogs make history, but let's look at all the info, not just isolated cases. Since 1970, and including last weekend's games, there have been 129 occasions where two teams have played at the same location in both the regular season and post season. I think it's important to further sub-group those games. 56 of those were divisional rematches, and 73 were games between conference opponents. Last January, I also wrote about the playoff bye week and whether it increased home field advantage in the divisional round (concluding it does not). That post also contains data on climate effect in the post season. I think its also important to account for the powerful influence of climate differences on home field advantage in the post season, so I divided those 129 games by whether they were divisional or conference rematches, and by whether the road team had a cold weather climate disadvantage or not.

Type			Season	PF	PA		Playoff PF	PA
============================================================================
DIV-Weather		3-12	20.5	25.3		2-13	14.3	25.6
DIV-No Weather		17-24	19.4	22.2		19-22	18.8	21.4
CONF-Weather		4-15-1	18.2	26.6		6-14	16.9	24.1
CONF-No Weather		9-44	17.0	27.2		15-38	18.6	25.0
============================================================================

And to make it a little clearer, here is a continuation chart showing the average point differentials in both the regular season and playoff games, for each category.

Type			Season PF-PA	Playoff PF-PA	Point Diff. Change
============================================================================
DIV-Weather		-4.8		-11.3		-6.5
DIV-No Weather		-2.8		-2.6		+0.2
CONF-Weather		-8.4		-7.2		+1.2
CONF-No Weather		-10.2		-6.4		+3.8
============================================================================

So let's talk about what is here. When a division opponent with a climate disadvantage gets a second shot, it doesn't really matter. In fact, the climate effect is strong and the drop off is severe. When division rivals meet again, and the teams are either from similar climates or the colder weather team is on the road, there is no change. This makes sense. For these teams, one extra game isn't going to make much difference. They have already played in the stadium continually for several years.

However, we do see improvement in the second game when the opponents are not from the same division. Here, the teams are not accustomed to playing on the same field every season, so the additional chance to return to the same stadium a few months later improves the performance. For the games where the road team is not at a climate disadvantage, the improvement is by close to four points on average. Only three of these conference road rematch teams had a twenty point or more dropoff from the previous result (Denver at Indianapolis, 2004; Minnesota at San Fransisco, 1988; Pittsburgh at Oakland, 1973). In contrast, twelve of the road teams had a twenty point or more improvement over the regular season result.

road	home	year		Reg Season		Playoffs
===========================================================
PIT	IND	2005		L 7-26		W 21-18
LA RAM	TAM	1979		L 6-21		W 9-0
NYG	LA RAM	1984		L 12-33		W 16-13
ARI	CAR	2008		L 23-27		W 33-13
LA RAM	DAL	1979		L 6-30		W 21-19
PIT	DEN	1989		L 7-34		L 23-24
SEA	CHI	2006		L 6-37		L 24-27
NWE  	PIT	2004		L 20-34		W 41-27
DAL	SFO	1981		L 14-45		L 27-28
KAN	HOU	1993		L 0-30		W 28-20
MIN	LA RAM	1977		L 3-35		W 14-7
DAL	LA RAM	1978		L 14-27		W 28-0
===========================================================

I don't think the +3.8 point improvement for these conference road teams is entirely due to the familiarity effect of getting to return to the same stadium. I think there are two other factors at play as well. First, most of these teams lost and performed worse than expected in the first matchup, so there is some natural regression. Dallas, for example, lost by 31 to the San Fransisco 49ers during the regular season, but that result was the anomaly. Second, there is probably also some overconfidence/let down by the home team. After all, they've already beat the team once before at home (often impressively), so they should be able to do it again. Of the nine teams that won the regular season matchup on the road, only three won the playoff rematch (including Baltimore at Miami this year). Thus, twelve of the forty-four losers were able to reverse the results, and ten others lost by a touchdown or less. That said, I do think the familiarity concept is playing a role.

How might all this apply to this weekend's championship games? Well, in the AFC Championship Game, Baltimore and Pittsburgh are similar climate division rivals. As I first discovered in this post, these similar climate rivals (all AFC North games, New England/NY Jets/Buffalo, Chicago/Green Bay, Philadelphia/NY Giants/Washington, San Diego/Oakland, and recently, San Fran/Seattle) show virtually no home field advantage. We've already seen this last week, as the Eagles won at the Giants for the second time in one season. Thus, it should matter very little that this game is being played in Pittsburgh. Pittsburgh is trying to beat Baltimore for a third time, but both games were extremely close and Baltimore had a lead in both late.

In the NFC Championship game, on the other hand, the Eagles will be playing at University of Phoenix Stadium for the first time. Most objective measures would rate the Eagles as the stronger team, and they have already defeated the Cardinals handily. However, the home field advantage that the Cardinals are currently enjoying is an equalizer. I think the Eagles should still win because they are the better team, but IF the Cardinals advance to their first ever Super Bowl, the concept of stadium familiarity will have provided a silent assist.