[Review] Indigo Xtreme vs. AS5, MX-2, IC Diamond, Shin-Etsu X23-7783D

[Grinchy]

and if top 20 differ by 0.7C, and possible error sum might be more then 5C, how someone can trust those results at all?

that's why these tests are of a dubious reliability and that's why in that review they warn not to watch at the 0.1 difference but they divide tims in categories.
These methodologies are not good at all. Standard tests should be made taking results in W/°mC, applying same layer of thermal material,constant heating....and blablabla
Saying that with a cpu a tim heats 0.1 or 1°C less than another doesn't necessairily mean that the former is better than the latter