Pretty sure they used the staircase method. You know, write the name of each engine on a seperate piece of paper, stand at the top of a staircase and throw the stack of papers in the air. Note their respective positions upon landing and there you have it: a ranking system.
They probably just take the developer's word for the rating. Doubt there is independent testing, but then I'm wrong a lot.
Does anyone know how the ELO of Lucas Chess engines were determined? For exemple - I'm a beginner I know - I can win against the "Gaia 1" 1000-something engine but the "Fixed 600 ELO" Rodent plays me in a corner??