Statistics are not a substitute for the long-tail effect.

Lack of reporting from non-technical people who aren't aware they can opt-in cannot be corrected statistically, as the two categories of people (technical, non-technical) use the browser very differently.

For made up example, if you type "Yahoo" into the search bar and then type "Search" into the field and then type your search into the third page, you'll be acting as many normal world users do, and you may uncover crashes on page #2 at Yahoo that a technical user would never encounter, simply because they wouldn't type the word "Search" into the search field at Yahoo and trigger a JS bug where "Search" or "Yahoo" gets used one too many places and ends up crashing the CSS parser because it race conditions with repaint.

If that problem affects 0.01% of the Firefox population, that's a lot of people who don't think technically, and do feel regret when we crash and can't help them because we can't see where it crashed.

(Yes, employed. No, I didn't talk to anyone else before I posted here. My own thoughts, I am not a number^Wcorporation, etc.)