His response actually sounded a bit defensive. Trying to deflect the discrepancy as AMD might be pulling shens. Not liking what I'm reading @ TPU lately???
That sound rather trollish, i quote this guy sentences :
It might also be possible that the variation between Hawaii GPUs is very large and I got an unlucky sample.
Why unlucky.?.
Generaly reviewers complain when they have a sample that ock badly or consume too much, here the guy say that he would had prefered something not as good, that would had been luck...
Let's just not hope that reviewers get low-power-picked cards and the high-power cards end up with customers.
He imply that this lack of luck could be AMD s fault, that is, they deprived him, and others, from cards that would have showed some "expected" results rather than those "unexpected" results, but consumers, be warned , thoses results are possibly wrong and you ll end with cards that have the results we "expected"...
To summarize they are incapable to do a review and take their own numbers for what they are, they are saying that they are skeptical of their own measurements due to some AMD possible manipulation..