I can only hope they kept logs of all takedown requests. Now they only need to find a reasonable sample. Honestly, I believe that even if they tried hard not to make mistakes, at this volume they would get some false-positives - and that's as good explanation as I can imagine. Now the question is - can Mega* guys prove enough of cases like that. Or maybe they have other explanation for the quotas - I don't know - I'm eager to hear more about it.
Assuming the limit was close to the average number of take down requests it's reasonable protection to prevent a buggy third party script from harming their site.
