But on average that thin-faced man is going to be at a genetic disadvantage, which I believe is the point of the article. Of course there will be outliers, and of course you can overcome some natural deficiencies with training, but on average they will hit limits that wider-faced men will naturally surpass. The article doesn't go much into the strength of the correlation, so I'm not sure how big a difference it really makes, but it clearly makes a difference. The original paper probably has more on this.

So there are several metrics that are not included here. First, what's the standard deviation on that average. How easy is it to be an outlier in this scenario? Second, how strong is the correlation?

More importantly, having a wider face is not a direct advantage. It's an indication of a disposition. This is dissimilar to height of players in basketball where you can in fact be too short to be effective. Because of this I'd argue that while looking at it as an average the findings could be very much valid, looking at an individual fighter will yield very different results. You could find that a thin faced fighter might actually be superior to a wide faced fighter precisely because he is thin faced.