It seems as though the two of you only use incomplete and inconclusive (as all measurements of loudspeakers are) measurements on which to base your opinions. I have a serious problem with that, especially when it comes to telling others how to spend their money.
Actually, the measurements here are very informative. And they are enough to recommend it as a linear device for the purposes I suggested it for. The lower than normal cabinet resonance is also a major factor here, at this price range.
Over the last 3 years I've probably taken 50,000 measurements of individual drivers and complete systems. But more importantly I've spend thousands of hours listening, and what I learned is that NO loudspeakers measurements I know of are enough to draw conclusions.
It is no fault of mine if you have not made yourself familiar enough with the founding perceptual research in order to correlate the magnitude, of specific types of measurements, and the complex factors that determine audibility on human subjects. However, I have specialized in this field of study, and design accordingly with expected results in almost every case. I add in randomized double blinded testing, also, as this is the only true way to evaluate things. Proper understanding of the measurements in relation to human perception is critical to understand the actual effect measurement variables play.
If you're going to come to any sort of conclusions based on measurements alone, you need a LOT more data than is given by Stereophile. CSD and on-axis FR are the same data, off-axis FR is good data, and accelerometer readings are good data too, but the sum of these are NOT enough to make any conclusions. There simply are too-many unknowns.
The only unknown here is non-linear distortion. It can be presumed pretty safely that B&W did not design this system and leave significant distortion as to be audible in music program.
I have heard 805s' a fair amount, a good friend is a manager at a local B&W dealer called The Sound Room, and he's a B&W fanboy ... and thus makes me listen to a pair whenever I visit

They're a fine speaker, I'd say they'd be a great option at around $1000-$1500 ... but at $2800 there are MANY better options.
It really does not matter what you(or I) think of them in their inherent response so long as their is no gross distortion or severe FR anomalies, or severe resonance(
which in fact many people do not identify properly though they hear it and it negatively affects their impressions). The frequency response is really undesirable in the treble here for using for most commercial music without tonal shaping. In addition, it likely needs more BSC compensation. But the weakness of it sounding ideal used by itself is exactly what makes it so good for manipulation: it's neutrality. This is a very linear device given it's price range - and it's trivial to adjust these things to ideal preference in a DSP unit such as I suggested. There is no magic or mysterious factor(s) at play. I have studied this field of perceptual research and conducted my extensive double blinded tests of many factors - most things people like to claim(
and this specifically includes man experienced DIYers and professional designers alike) are nothing more than speculations based upon incomplete understanding and relying upon one's biased sighted listening tests. The measured performance here depicts a speaker that is excellent for the price range for the specific purposes that I outlined in my first post.
-Chris