first paragraph is the "ideal and why" for the sensors placement -- cleanest signal at CG

Second paragraph is "ideal and why" for the airframe's CG -- centroid of the motors, since the CB will have to level out the motors anyways, it's best to make it easy on the CB and put it there naturally.

the second answer completes the first.

One caveat: the centroid of the motors is *NOT* the center of a circle all the motors rest on (not fig 2), but instead a weighted averge in X and Y of the motor's positions -- it's really the point where the force vectors actign at that point balance out any torque.

We assume all motors being equal, the weighting is all 1, so the math becomes easy. Also the roll-symetry between the motors makes the pitch position averge trivial -- it's the centerline of the craft.

Still with me?

So if you pick the rear boom line as your 0 along your roll axis the rear motors both contribute weight to the averge, but not distance, so if the length to the front booms is L:

CG position forward of the rear motors = (0+0+L+L)/4 = 2L/4 or 1/2 L.

Figure 1 is the most correct, but it depends on equality in motor balance -- if your tail motors would be half the strength but co-located (like in a y4) The true balance point would be farther forward of the rear motor, becaues the total weight contribution on the bottom woudn't be 4 -- it would be 3 (.5 +.5+1+1), and CG would be 2/3 L, just like a tricopter.