I honestly don't think those tests are accurate at all, I think it's due to test factors not being normalized correctly - tricky things like lever actuation force input point, and more obvious things like just using the same pads. It's surprisingly difficult to conduct a completely objective and accurate test. I'm glad they try though, perhaps in future methods will improve.
For reference, I own the one brake that obliterates all these tests.
For reference, I own the one brake that obliterates all these tests.