What i suspect might be happening here is that Bernard was initially undecided whether to get the D800 or D800E. He chose the D800, presumably on the basis that the D800E had only a marginal increase in resolution but also an increase in aliasing artifacts which he thought was more significant than the increase in resolution.
I think Bart has tested that the D800E has only a 1% increase in resolution, over the D800, after appropriate sharpening has been applied to the images from both cameras.
The disturbing issue about the D800E is that it has unnecessary AA filters, which cancel each other. The light signal from the lens is therefore passing through an unnecessary process which might degrade the signal to some extent, depending on quality-control variations.
Since the D810 has no AA filters at all, it's quite likely that on average, the D810 will have a greater than 1% increase in resolution, compared with the D800, in relation to Bart's methodology. In fact, if Bart had tested a few other copies of the D800E, he might have found that some of those copies had a greater-than-1% increase in resolution compared with the D800.