If you think about it, these results add credibility to Zebralight's own explanation. They said they had a bad batch of H501s in early 2009, and that's where you see the majority of failures. Once you take out the H501s, it's hard to make the case that Zebralight is, across the board, a manufacturer with horrid reliability, as has been implied countless times here at CPF.
I'm not considering the H600 at the moment because it's new, and the 'n' is low. But I sure as heck hope that isn't history repeating itself...rushing a new light out the door before it's been completely tested? That would border on unforgivable, given that the lesson should have been learned with the early H501s.
Honestly, the results as they stand (caveats aside about the methodology, as has been noted) make me wonder if I want to be first in line for the H502, or do I want to wait a couple months? If Zebralight is using the marketplace as their quality control testing laboratory, then that tells me to wait, and let someone else be the experimental subject...!