By strict logical standards you would be right. If they can prove it won't damage then it fails the test. But if that is the case then also "if they can't prove it will damage the network" also fails the test and they could block those.

I would think if we looked at the specific ruling it would be:

can't prove it won't damage -> allowed

can prove it will damage -> can block

can prove it won't damage -> allowed

can't prove it will damage -> allowed

So it would have been clearer if they said if they can prove it will damage the network.