I'm pretty sure that doesn't mean: affinity should reach the goal of 0 errors. But some of those 98 errors might be easy to fix, other might be very important. So it's still valuable feedback.
I don't agree with "any application will fail that test" … You simply didn't get the purpose of it.
Affinity should take that suite and try their best to make as less errors as possible. and: some of those problems might fall out of the 80/20 rule … I'm sure affinity is able to identify the important issues AND the low hanging fruits as well!