Personally i believe (like a Zealot you would say) that there is way more going on with load tuning than can be classified as statistically relevant or proven.
If you lined 100 Statisticians up, ask them to statistically analyse some data they would never come to a consensus on the "results". The best bit of lining those 100 Statisticians up is you can identify 100 people you probably wont want to have a conversation with...

