That's one way to not measure power usage. Normally reviews test it with media playback, games and office utilities tests. Ie I want notebookcheck to give a review before we go blowing horns over how good it is.
Laptopmag are quite well known,and are part of the same group as Toms Hardware and Anandtech.
They have done the most important thing that even NBC and other sites don't bother to do - test both the Intel and AMD versions of the SAME laptop.
Everything in the two laptops are the same,outside the CPU,motherboard and case colour.
It also means both chassis have equivalent cooling capability and the same screens. Exactly the same battery in both. They also weigh the same and have the same dimensions.
So if the same test is run,its accurate.
Remember,they are running the same benchmarks and same testing on both systems,so ultimately this is an accurate relative comparison.
All previous reviews have tested the Ryzen 5 2500U based X360 against different laptops. Plenty of us benchmark ourselves and we know very well its important to keep common components,ie,if that CPU or GPU is better we are actually looking at those parts in isolation.
If you don't equate all the other parts,you are not testing CPU performance,but CPU performance and the implementation of said laptop.
If you don't believe me - look at every site that tests a graphics card or CPU/Motherboard/RAM,every other part is identical to stop any bias in the results even down to the case used.