I debated whether to add the Cd to that chart but since I did and you "caught" me I'll elaborate a bit.
One of the issues [not really a problem] with computer models is they are perfect!
None of the lumps, bumps, rivets, waviness, seams, warts, wrinkles or variations of real world parts. So that model says if those parts were exactly those shapes and they were placed exactly the same in a perfect wind tunnel this could be the potential Cd. Cd is a dimensionless number! WTF does that mean? We can measure density, frontal area and velocity but there is no number to measure shape - hence a dimensionless number is derived that enables us to compare different shapes. If the same methods and measurements are used to compare shapes then the "value" of that number has no meaning other than one shape is better than another shape. As has been stated - don't get too hung up on that dimensionless number!
Think of a pressure gage with no increments - if the needle goes CCW that's good but if it goes CW that's bad or vercey-visey! It still gets me the improvements I want just not quantified to any standard.
So the new chart show what happens
if when we don't have perfect parts. So if we fudge some of the drag values to guesstimate what might happen if the tank is not as slippery or the tires are bigger or a different shape or you have a roll cage hanging out in the breeze you can see that the Cd goes south in a hurry. In all cases the tires and axles still have the most drag but the tank is getting pretty bad at the end. Wonder what some F1 tires would do? Double or triple the tire values and you will have an idea. If you can't change the shape you have to make it smaller.
Me thinks that just confirms many of the real world comments already stated in this and other threads.
As my byline says: All models are wrong but SOME are useful! That's my story and I'm sticking to it!