You are weighing the tentative results of one study equally against many others saying helmets reduce injury.

The study is also weird in that in the course of being passed by 2500 vehicles, he was struck twice. That is an astronomical accident rate that leads me to think his test was abnormal in some way.