I just wanted to issue a reminder to everyone to keep seasonal variations in mind when reporting results. As weather changes, particularly temperature, FE results will get skewed. Just to prove the point, I recently made this plot from my daily log data from the last 2 years, showing average fuel economy for every month:
Jun-Aug are typically good months because it's so warm, Jan & Feb not so much. Keep this in mind as your looking at and reporting the effects of mods. A-B-A testing is of course the best way to do things if you can. The issue with only A-B testing is that weather conditions (and other factors) can artificially skew things. So, as things cool down, don't be discouraged if you see little or no increase in FE numbers. And, when things warm up, be careful to not inflate the beneficial effect of a certain mod.