I agree, this test is not very scientific. I also can’t read Chinese, so while Google auto-correct seems to do a decent job some of the details may be lost in translation. It’s possible that some of the perceived improvement in position (e.g., when standing next to the building) could be at least partially attributed to dual frequency GNSS, but given the design it’s impossible to tell. When using closed-source 3rd party apps it’s not clear what the exact position source is. It’s also not clear if dual-frequency use of GNSS is active during these tests. I’d like to see a test of two Mi 8s side-by-side, one configured to have dual frequency turned off and the other to have it turned on.