The figure is also in the origional paper
. The strength of Yann le Meur's info graphics tend to be that they stick very closely to the papers they are trying to convey.
The figure does show some people improving and others getting worse. A back of the napkin calculation puts the top line at +~6% (320/300). As I said there is no way of empirically knowing the direct causes of any change.
What you say is not wrong, and perhaps I should have been clearer - but there are differing goals between the scientific and the high performance community which should be considered. The athlete or coach does not care about statistical significance. They care about performance.
Science is based on the methodical accumulation of evidence. That study doesn't provide strong evidence that there is a performance benefit to a gluten-free diet in non-celiacs. However, the evidence it provides against a performance benefit is also not strong. There is a large amount of inter-individual variation. One study on 13 varied mixed sex participants will only ever graze the surface.
Thus for an athlete seeking a performance benefit, especially if you have had problems with GI distress - what have you got to lose? Try it in the off season, see if there are improvements.
GG - Sadly not; ba-ba - No comment