I did, you said he used a China only model (that's an AMD reference card from what I can tell) instead of using a global release version (with altered power targets I assume?).
What is the "global release" reference card and what are it's power limit? Or has it not been released yet? If AMD choose not to release a reference card in other markets, that's on them. If AMD however have a different reference card for other markets, even if it's not out yet and only annoucned, that's on HUB.
I'd also point out, if AMD have a china reference model gre, and are planning on releasing another, differently spec'd, global spec gre that's also on AMD because that'd be stupid, they should have named it differently.
I see no issue testing a reference model as a base line, testing an aftermarket one with altered specs as a baseline is a no IMHO. It's fine if they're reviewing specific card, i.e. comparing a Nitro+ to an XFX Merc, but not for trying to establish baseline, otherwise which do you put into charts? The slowest one? The fastest one? All of them? Then do you need to do that to all the cards to make it fair?