Run a type A analysis on each run parameter... 10 runs each parameter, keeping all the contributing factors within a margin for each parameter, and Ithink your repeatability will fail what you are attempting to preach.
I could run 50 and there would still be excuses. I'm not preaching anything, just reporting my results. I don't sell these things or even use them. To say there is no increase in power is just 100% wrong and the facts don't back it up. Maybe someone else can do 500 pulls and show there is no increase as that is now the opinion with no data.