I haven't seen anyone doing full runtime test, but krikzz has measured the power draw compared to a retail game, which was pretty muc identical to the EDGBA.
The problem with those sorts of tests is that it's possible that 90% of the time it draws low power and 10% high power or some other mix. It's clear to me from lots of experience/observation that the best benchmark is a test that's consistent and as real to real life as possible*. So, yea, I'd still prefer a full runtime test. :/
* There's also knowing the testing setup, as demonstrated repeatedly with Microsoft "proving" Edge lasts longer in web browsing because they defined web browsing as watching video and then optimized video playback for low power consumption. That's a whole other part of gaming the test. :/