Ugh, could they have introduced any more variables into that to muddy the results? Why didn't they test Force 1 against Force 22? Not that I would expect the outcome to be substantially different, but in a test that's trying to quantify small differences in efficiency between 1x and 2x, you'd think they'd want to isolate the changes to only those required for 1x. Run the same chain. Change the cassette ratio, but get it from the same family. And choose two RD's from the same manufacturer, at least.
"They're made of latex, not nitroglycerin"