logicators, thanks for the update. I have a few quick points of note:
First, your 2nd graph, the one you say is no-eq, no-smoothing, has a different label (and avg dB rating) listed than the original one you posted, so it looks like a different measurement. It may be nitpicking, but since you're going back to these saved measurements to generate the graphs, go back to the original measurement from the first 1/8th smoothed graph and just change it to no smoothing.
Second, instead of screenshotting the entire program, you know that the "capture" button near the top left corner there will create JPG or PNG files automatically that show your labels and how much smoothing is applied. That'll make it much easier on you by saving some time.
Third, your last no-eq, no-smoothing graph shows a response from "P2," which I assume is position 2. Is this different from your main listening position (MLP) as labeled on that same graph?
Lastly, I see that the first graph you posted is a full-range measurement, with all speakers on, but you singled out the 5-80Hz range. If that's indeed the case, I would definitely be worried about your major suckout centered at 65Hz. When measuring full-range in REW, a suckout there means a major room mode (about 20dB down from a large portion of the rest of the bass region in your graph), and at a place in the spectrum where there is a fair amount of content that you don't want to miss. I'd reconsider your XO and see if raising it higher and/or adjusting phase doesn't eliminate or at least ameliorate it.
Good luck, and let us know how it goes.
I get what you're asking about WRT how finely grained we hear, and I don't know. I can't find any definitive answers to the question, although I may not know what to search for.
As regards your other comments about the comparison of systems with EQ performed from different smoothed starting perspectives, I would definitely think you can hear the difference, but it depends on your definition of "ruler flat." Do you mean +/- 3dB, or +/- 5dB? I think that makes a huge difference, especially since smoothing can hide a wider & deeper dip if your definition of flat is relaxed.
It also matters where the peaks/nulls are located. The range of 30 - 50Hz, for instance, has a TON of very high amplitude content in almost all bass movies. 60-80Hz? Not so much. I'm guessing that if you have a deep, 1/6th octave null somewhere in 30-50Hz range, you're gonna notice it much more than a similar null in 60-80Hz.
By the same token, a major peak in the higher range will definitely be noticed as a pronounced, boomy sound (which I've experienced, for sure). Both of those might look like 5dB swings in a heavily smoothed graph, when in reality they might be 10 or 12 decibels. (Hope my math isn't wrong here, I didn't do the log calculations!)