The NY Times article whose headline states that helmets do not reduce head injuries does not in fact say that; only the headline writer says that. The article does claim that the incidence of head injuries has gone up but offers a number of convincing reasons why that should be--primarily an increased incidence of risk taking behavior as "extreme" skiing becomes more promoted and more popular, especially park skiing. I would add that continuing improvement in gear has made higher speeds far more common. Without knowing the denominator--the total number of participants, and more particularly, the number of participants skiing and riding in a particular fashion, the overall number of head injuries is meaningless. If you truly wanted to find out if helmets do any good you could prospectively monitor a number of terrain parks around the country, count the number of skiers and riders and the number of runs they take and the number of jumps per run, note whether the participants were wearing helmets or not, and then see whether or not the number and severity of head injuries differed between helmeted and helmetless participants. Obviously such a study would be extremely difficult to conduct, but if someone really wants to know if helmets help a study of that sort would be necessary. Obviously expecting helmets, especially the minimal helmets most of us who wear helmets wear, to prevent all head injuries is unrealistic, but that doesn't mean they don't do any good. And on the other side of the issue--a story about someone who survives or is uninjured after they hit their head and their helmet breaks doesn't prove that helmets save lives either.
The absurdity of the attitude towards helmets was particularly well expressed by an article I read recently about an avalanche victim who was found dead after a several hour complete burial--the article noted he was wearing a helmet.
I personally believe that helmets do mitigate, but do not prevent, brain injury, but I have yet to see data that proves or disproves my belief or gives me any idea of the magnitude of the benefit. I also believe that helmets could be designed to be much more effective, that the current testing criteria are not adequate, and that data such as the number of G's a headform is subjected to during a standard test should be available to the public for all helmets as well as for the unprotected headform, so that buyers can make a rational choice.