No. No it doesn't. You are talking about averages over many thousands of miles. One poorly calibrated pump isn't really going to affect the data pool in any significant way. Statistically speaking of course (and that is what you are dealing with here, statistics).
On the other hand, a poorly calibrated OBC will severely affect the data pool, and thus makes the whole exercise pretty pointless. Sure, it will at least be consistent, for that one vehicle, but it does mean that trying to draw any reasonable comparison BETWEEN vehicles is pointless, as there is no regulated standard for calibration between manufacturers OBC's, whereas there is a regulated standard for petrol pumps.
Basically, your method promotes whole vehicles as potential outliers when doing comparisons, compared to taking pump data where there may be an occasional outlier pump that doesn't really significantly affect the accuracy of comparison between vehicles, unless someone exclusively uses a dodgy pump. Moreover, you are severely restricting the gathering of data with such inaccurate rules. Since the very people you really want to be posting in this thread, are the people who will only gather accurate data over many thousands of miles (I have over 100k of tracked miles between different vehicles in fuelly for example). But these people are unlikely to use inaccurate ways of gathering their data. Look at the amount of posts on here that use some sort of fuel tracking app, or website to see what the general trend is for gathering fuel economy data.
Maybe someone should start a sensible fuel economy tracking thread. It's a topic I have some degree of interest in, as I find the average fuel economy of a real life vehicle is usually the better metric to use when deciding if a vehicle will fit my needs for a particular task. I often browse Fuelly to see real world numbers for cars I have a passing interest in.
