[MetroMPG]

Tank to tank testing is not a viable way to test modifications.

It is simply too noisy (too many uncontrolled variables affecting fuel economy).

If you are convinced that the ECU is affected by these magnets (requiring significant time/distance to "settle" between changes), then you are correct in saying that only laboratory test conditions are acceptable.

At this point you should have the intellectual honesty to agree that this isn't testable, tank-to-tank, and that conclusions can't be drawn from attempts to do so.