I don't see how you can say that "X is far more efficient than Y" if you haven't 1) measured the electricity produced by X and Y and 2) measured the drag produced by X and Y. Since we're talking about lights, we probably should also 3) calculate the lumens/watt for their emitters as well.

Ultimately, you could compare watts to drag or lumens to drag -- take your pick, but you're going to have to get quantitative data on at least two variables here.

You've given some quantitative data for the two lights in your other thread (i.e. you can feel the drag from one but not the other), but you haven't actually measured it that I can tell. And while you've talked about how bright they are -- 30 lux vs 60 lux -- note that lux doesn't tell how much light the light emits, only how much light reaches one particular point (the brightest area) or the average over an area that they consider to be "the" important area -- "lux" is quite vague and not suited to comparisons regarding efficiency.

Comparing average lumens (average, because many dynamo lights have a visible flicker) would be ideal, and if you want to compare beam patterns you could count only the lumens that go where you want them, but then it becomes subjective.

As for typical values, I've heard that hub dynamos are around 50% efficient (so a hub generating 3 watts of power would generate about 6 watts of drag if it was 50% efficient), and modern LEDs tend to put out about 100 lumens/watt. I don't really have any idea how efficient the magnic light's electrical generation would be, but I wouldn't expect their LEDs to be particularly unusual.