In honor of the start of football season, I thought it would be a fun topic for this week’s dip into self-delusion to cover the undeniable faith of football fans. And lo and behold, a new study just out in PLoS One covers this very subject.
Fans, like professionals assigned to cover a team, were overly optimistic about their team’s prospects. The opposite pattern was found for teams that fans disliked. Because success within the NFL is zero sum, these results make clear that bias exists and that collective decision making is inconsistent.
A team of researchers from the UK looked at fans pre-season predictions of how their most favorite and most hated football teams will do in the coming season. Yes, they are in the UK, but they looked at NFL football.
They used the NFL because it is a very distinct and clearly defined zero sum domain. For each team that wins, another one loses. The average wins for the season has to come out to exactly eight and the average losses for the season has to come out to exactly eight.
Then, they used statistical regression to model what one should objectively expect for a team based on its record from the past year. It turns out that teams’ records are both autocorrelated and revert to the mean. In simple terms, any team that was below .500 one year will be a little closer to .500 the next year (a little higher). Any team above .500 will also be a little closer to .500 the next year (a little lower). Of course there are variations, but this is the average. So if you look at thousands of fans across all 32 teams, this is what you would find if people are completely objective in their predictions.
But what they found was quite different. The average predicted by fans was 9.6 wins for their favorite team rather than 8. Note that this includes teams with terrible records the previous year (i.e. less than 8 wins) and teams with great records (i.e. more than 8 wins). They also calculated regression models to look at individual teams and found this overconfidence across the board. No matter what a fan’s favorite team’s record was the previous year, they overestimated its expected record in the following year. As a comparison – just estimating that a team would do exactly the same as they did the previous year was more accurate than what fans really predict.
A team with 12 wins the previous season should regress towards the mean the next year. Some of that outperformance was due to fortune – fewer injuries than other teams, a good trade or two, and so on. The bad teams should be better competition – fewer injuries, a good trade or two of their own. But fans don’t consider this.
Most Hated Teams
As a good control, they also looked at fans’ predictions for their most hated team. These were underestimated. The average for their most hated team was 6.1 wins. Again, this covers both teams that were good and teams that were bad the previous year. So the average should be 8 wins. And the regression showed that it was true for all fans, regardless of which team the fan hated. So there is overconfidence there too.
They also report the results of a test of ESPN’s local commenters. As you may know, ESPN has specialists assigned to each team. They should be objective, right? Nope. They are a little less overconfident than the fans are, but not by much. The average predicted by experts was 9 wins for the team they covered.
There are a few good human factors reasons for the overestimations and underestimations besides just wishful thinking. A fan is likely to know more about their favorite team and pay more attention to their favorite team. They are more likely to pay attention to good news about their favorite team and to bad news about their most hated team. This is natural. After all, sports fandom is not intended to be a fully objective process. What fun is that?
To check this hypothesis, they looked separately at the predictions of more knowledgeable fans. They asked respondents how many yards an offsides penalty is. 86% answered correctly. These fans predicted an average of 10.3 wins – even more overconfident than the average. Not less.
I am sure there is also a good bit of pure wishful thinking too. Even conscious wishful thinking – “Maybe they are losers, but they are MY losers.”
Who are the Favorite and Hated Teams?
In case you were wondering, the same teams that were favorites were also the most hated. That is what I expected – it is all about jealousy. There is no sense hating a bad team, they are no threat. So the New England Patriots showed up as the Most Liked (by 7.7% of fans) and the Most Hated (by 17.3%). And keep in mind that the study was done before DeflateGate, so that isn’t it. The Packers and Cowboys also show up on both lists. I have a sense that Seattle will show up if they keep up their overperformance.
They also looked at the specific teams with the most overconfident and least overconfident fans. The most overconfident were Cincinatti and Arizona (overpredicting by 6.6 wins!!). The least was Denver (only overpredicting by 0.7 wins). Since this was just one year of data, these could be anomalies – perhaps Peyton Manning’s injury kept overconfidence at bay.
If you are a fan of any sport, you have probably experienced this effect personally. A player’s footsteps just at the sideline. You are positive that the foot is in-bounds. The fan sitting next to you, with the same view but cheering for the other team, is just as positive that the foot is out-of-bounds. “Ref, are you blind?!?!?!”
Please share your stories, interpretations and human factors insights into this domain. There are many more factors to consider, but not very much research on the topic. So opinions will have to do. And anyway, when it comes to sports, opinions are much more fun.
Image Credit: Giovanni Gallucci