Now I will show some data plots for AMD gains...
The photo shows all pairs..
Note that 4 points (see arrows) are away from others that's why I BANNED them from data modeling!
But I kept total samples = good samples + 4.
Here is
the data modeling function (red color).
Green dashing curves are the upper and lower limits of each region.
As you can see 4 points are outside of areas (show arrows again).
Would you like 100% possibility?
OK it is easy!
But if program output "Estimated area is 2700 - 2800 MHz and possibility is 100%" is this useful?
That's why I keep estimated areas 2 - 4 * multiplier MHz for all estimators.