Does travel distance impact performance in World Cup Qualifying?

By Jackson Weaver

Qualifying for the 2018 World Cup has come to a close, and for many fans, including myself, anger and outrage fill our hearts over our respective country’s inability to qualify. Beyond poor performance, there are a number of factors that disgruntled fans tend to blame, among them injuries, region, and faulty coaching. One of the most common excuses is travel distance and harsh away conditions.

I decided to examine the the impact of travel distance on away results in World Cup Qualifying after reading an older HSAC post on the effect of travel in MLS by current HSAC co-president Brendan Kent, one of the more geographically diverse leagues where home field advantage is considered incredibly important. I came across this article days before the US lost away to Trinidad and Tobago, which led me to a state of massive denial. I wanted to blame CONCACAF for the US’s failure, so I decided to research this topic.

To test whether longer travel distances negatively impact resulting in World Cup Qualifying, I fit two models: a logistic regression model where the response variable indicated whether the away team achieved a result (a win or draw), and a linear regression model, where the response variable was goal difference from the perspective of the away team. The dataset includes World Cup Qualifying matches in Africa, South America, Asia, and Oceania. The proxy for travel distance is the distance from away country’s home stadium to the opponent’s stadium and the results and locations are those reported by FIFA. In the case of countries like the United States who use many different home stadiums, the stadium in which the team most recently played a home match was used. Additionally, In both models, team strength was controlled for using the difference between the ELO ratings of the two countries.

In both models, travel distance proved an insignificant predictor of away team performance when controlling for relative team strength. In the logistic model, the p-value of the travel distance covariate was 0.27, while in the linear model, the p-value of the travel distance covariate was 0.72. The plot of goal difference vs. travel distance (below) illustrates this lack of a relationship between travel distance and goal difference:

While these models demonstrate that travel distance has not impacted away results in World Cup Qualifying, it does not necessarily prove that travel has no impact on player performance. It could just as well be that teams plan travel to minimize it’s impact – this could mean, for example, flying to games multiple days in advance. Additionally, this study focuses only on team travel, and it is possible that travel-related performance issues could be more related to the distances players fly from their club to join up with their respective national team. For example, Christian Pulisic must travel across the Atlantic to most US camps.

Unfortunately, this result debunks my excuses and denial of the US soccer program’s failures. Ultimately, the US cannot use CONCACAF travel distances as an excuse. Although the model does not include data from CONCACAF qualifying, it is inexcusable that the US lost this game and could not place top four out of a group with ELO rankings of 17th (Mexico), 30th (Costa Rica), 47th (Panama), 61st (Honduras), and 102nd (Trinidad and Tobago). It was a sobering loss that needs to lead to an in depth revision of the US Soccer System, from the President down to the youth systems.

About the author

harvardsports

View all posts

Leave a Reply

Your email address will not be published. Required fields are marked *