So all the skepticism of the Magnic Light's efficiency has me wanting to devise some reasonable and repeatable test. This is what I have come up with so far. If you find issue with it, please let me know. Otherwise I'll see if I can find time to do it this weekend:

Premise: I live at one end of a block with gentle hills on both ends and a dip in the middle.

Assumptions: same bike and rider, minimal ground contact on launch and stop. The total rolling distance is hard to judge w/o actually performing the test but perhaps 1/4 mile.

Test:
#1 : start strava
#2 : roll down hill with no brakes or pedaling
#3 : come to a dead stop
#4 : stop strava
#5 : measure distance for the run

I'd repeat this test twice (3x?) for the bike with Magnic Lights, w/o Magnic lights and dyno lights off, w/o Magnic lights but dyno lights on.

The only issue is whether or not to use a separate front wheel or use the dynamo front wheel for the Magnic light test. I can get a pretty close setup (32c top touring on an T520 vs 32c Marathon Supreme on an A719) but I don't have access to two identical front wheels (where the hub is the only difference).

So, would this be a worthwhile test to perform?