this is one of those impossible-to-prove-a-negative situations. It's very easy to point to instances where a helmet saved a life in an accident (and not just isolated anecdotes, definitely enough to make meaningful statistical statements).

And it's easy enough to look at more extreme cyclist deaths (helmeted or no) and to say "helmet couldn't do anything about that".

But it's near impossible to look at helmeted cyclist deaths and point to which ones might have been avoided if the cyclist was not wearing a helmet and thus had more vigilant, cautious riding behavior.

And yet the statistics (increasing helmet use is correlated with increased death rates, not decreased death rates) imply that there are a lot of those cases -- more than there are lives saved by helmets.

So do you want to wear a helmet as protection for your inability to ride completely vigilant and cautious all the time?

Or do you want to skip a helmet to help force yourself to ride as vigilant and cautious as possible all the time?

Of course everybody wants to answer "I will wear a helmet 100% of the time AND be 100% vigilant and 100% cautious 100% of the time". But to believe that's actually 100% possible is at least 57.3% foolish.

(And of course, there must also be an effect of, going with out a helmet increases vigilance and caution at first, but then you get used to it and grow complacent. It's only human)