That is an excellent video, thanks for sharing it.
Now,
@Jeff S ...
They are (usually) not tested to failure - which would be required to say one helmet is "safer" than another. So: certification sets a floor on performance, not a ceiling.
If I am interpreting what you are saying correctly, then I disagree. First of all, I don't agree that it would be required or even worthwhile to test any helmet "to failure". Certainly there is room for disagreement about the particular standard that's in place and whether it should be changed, but I don't think anyone would truly advocate for a minimum standard sufficient to cause all helmets to fail eventually. I'd also disagree that certification sets a floor, not a ceiling. I mean, technically, that's correct. Because the ceiling is not relevant. A gradient may be relevant in some cases, like the SHARP ratings for helmets, in which case you might be able to determine which helmet may perform better in specific instances, and indeed you might be able to rank one as safer than another if it consistently outperforms on all of the SHARP tests. But it should not be necessary in most cases to test any of this to failure, but instead simply test to the degree of intended protection. To make an extreme example, there may be some helmet out there that will repel a 5.56 round and another that will repel a .22 round but not a 5.56, but that's not relevant to a side-impact test since there is no reason either of them should have to repel gunfire. So to this end, if the standard is sufficient and all helmets pass, then in fact they are all equally safe. Otherwise, the standard is insufficient to evaluate safety, which can be true, but doesn't necessitate testing to failure to become sufficient.
I do appreciate how the guy on the video clarifies some of the assumptions of causation vs. correlation. Safer motorcyclists choose to wear helmets, and also make other safer choices, so they have a lower death rate than those who choose not to wear helmets and likewise also choose other riskier behaviors. It doesn't mean that if you put a helmet on an idiot hooligan it will double his chance of surviving the inevitable crash, or that when I test-ride my scooter around my cul-de-sac without a helmet on, I am twice as likely to die from riding my scooter.