Mark Thompson wrote:
3. Helmets protect better against impacts of a smaller force than of a
larger force.
4. So if x% of head injuries are prevented by wearing a helmet, then that
will generally be the least serious x%
5. So to start preventing significant amounts of serious head injuries
cycle helmets would have to reduce head injuries by more than 95% - and
that's just the ones serious enough to make people pop into the hospital!

I have two questions.

A) Is there a flaw in my logic?

Helmets could plausibly prevent (i.e. make non-injuries) only 5% of
injuries and still reduce some serious injuries to not-serious.
You only have to prevent over 95% of injuries to make some of the
serious injuries into non-injuries.

Now other figures suggest that helmets don't actually provide that
benefit, or only to a very small extent, but I don't think the argument
above is enough to show that, at least without some elaboration of 4.
(And if that elaboration turns into "helmets are, on the whole, useless
in serious impacts", you can omit the other steps.)