I'm comparing two OSes and pointing out when one OS has issues there's a single source to point a finger at. When the other OS has just as many or similar issues there is no one to point a finger at. So if you total up the finger pointing, MS will appear to be doing worse than they are. Where as if you total up the issues instead of the blame, they're doing probably no worse than average, maybe better.
If we're talking services, it's similar. 100 companies, 50 using MS, 50 using random non-MS software. 10 breakins in each category. MS gets the finger pointed at them 10 times. 50 random non-MS companies each get the finger pointed at them just 1 out of 5. But both MS and Non-MS have the same amount of issues in this hypothetical example, but one looks worse, even if they're not.
In fact there could be 5 breakins with MS and 15 with non. But MS would have a finger pointed at them 5 times and 15 of the 50 random companies would each have a finger pointed at them only once. Yet, if you added up the numbers you'd be safer with MS (5 failures out of 50) instead of random non-MS (15 failures out of 50).
I'm not saying that's how it is. I'm saying it's plausible.