Only if the scene you are metering is 12% reflective, otherwise your cameras meter is trying to turn that scene the equivalent of middle gray. The incident meter will be more accurate because it doesn't rely on reflectivity. I have no problem with using your method because you can chimp to get an overall impression of the brightness in the review screen and go from there, but the inference that its more accurate is mistaken.
I never meant to state it was more accurate - just stating how I go about doing it. We all have our different ways of going about metering ambient vs. flash.