USMNT Mythbusters: American MLS starters doesn’t correlate to national team success

Taylor Twellman made headlines at the beginning of the MLS season for highlighting the decreasing number of American-born starters from 2014-17. “That is a concerning trend when you look at opening day starters. Cause expansion is not going anywhere. So for the next 5-7 years if that trend continues, we are gonna look at 2022 [World Cup], the American starters being less than 33%,” said Twellman. He was very concerned about the effect on the USMNT, and U.S. Soccer’s response. He cited England as an example of a strong domestic league, but a struggling national team. However, in my research, I found that there is no statistical backing to his claims.

Statistical Background

In order to figure out his question, I decided to see if there was any correlation between the highest USMNT Elo ranking of the year and percentage of American MLS starters on opening day. I used highest USMNT ranking of the year because it provides a summary of the USMNT’s year. In a study of soccer rankings methods by professors at the University of Amsterdam, Elo was the most accurate at predicting soccer matches. The number of American starters came from two sources: Elias Sports Bureau (2014-2017) and WhoScored (2008-2013). The sample size is large enough to make comparisons, but with a little bit of caution.

Statistical Test and Analysis

I used a common statistical measure of the relationship between a cause (independent variable) and an effect (dependent variable), called “R”. In this case, the R value of .3202 demonstrates little to no correlation between the number of American MLS starters and the highest USMNT Elo ranking. A second calculation, called “R2,” is a statistical measure of how often one variable changes when another variable varies. In this case, the R2 value was .0914, leaving almost 90% of the change in the variable accounted for by other factors. In short, the percentage of American starters in MLS and USMNT Elo ranking are statistically unrelated, making Twellman’s statements off-base.

Another approach for evaluating Twellman’s claim is a linear regression t-test. This test gives us a line of best fit, and will tell us if Twellman's hypothesis is supported by this evidence. Conducting this test leads to the conclusion that there is no significant statistical linear correlation between the highest USMNT Elo ranking of the year and percentage of American MLS starters on opening day.

Conclusion

While Twellman is a knowledgeable tactical soccer analyst, his statement is not backed up statistically. Only time will tell how this affects the USMNT. The non-American talent hopefully will push Americans to fight harder for their places in the starting 11.