Interesting. They even use the Gaussian scale for mean probability in the interpretation of charts in futures commodities, like wheat, corn, soybeans, and light sweet crude oil. Even though a surface(s) may have the same Ra, they are effectually different, at least in the 3 examples shown.
Without wading through a lot of the feldercarb, it give a scientific approach of what accuracy means in determining surfaces and how their (base surface, ground, planed or machined) will turn out within a reasonable value. Major peaks and valley, at least the mean 3 values of peaks and the mean 3 values of valleys give an engineer some basis to evaluate the probability of the wear, given a known hardness, and so forth.
It's like knowing within certain values, of a projectile, like a .223 coming out of a rifled barrel of a Colt, and knowing standard pressures and the rifling of this particular barrel, with certain accuracy, if you know the velocity and the pull of gravity you will know the amount of drop at 100 yards, 200 yards, 400 yards and 600 yards of this particular bullet. You also have to know the coefficient of friction and drag through the air of this particular bullet, and you can get very good at dropping them in the right ring at whatever distance you need.
What does all this have to do with oil? Just a figure of how physics and knowing what properties you are dealing with can make all the difference in predicting the near term outcome of a test object. To the majority of oil users (consumers) knowing the microscopic peaks and valleys are not as important is what kind of beer tonight.. Bud or Bud light. But to the people working in the labs of oils and filters, metallurgy takes on a whole new meaning and what can they do given the "known" to make a better product.