Plug readers

IMHO...

Run tests in the same cyl. Under as close as the same driving and outside air temperatures as you can.

Remove as many variables as you can.