So yeah, the risk thing. I've posted similar to this before but.
-Cycling is already pretty low risk of any sort of serious injury (IIRC, something like 500 per billion km)
-Ergo, the risk of a head injury of any sort is even lower
-Of those, the subset where a helmet could prevent a death or serious injury is smaller still
-That's why we cannot detect any protective effect in the data - the specific circumstances where a helmet could help is so vanishingly remote that it is undetectable.
-And once again, there are plenty of activities with a broadly similar risk of head injury that we do not bother with a helmet for - because, like cycling, the risk is so low as to not be worth the bother
I do get somewhat peeved with the "But if it saves just one life brigade", it's such a moronic argument. Luckily there aren't very many of those on here.