Couple of question I have from that review... which overall seems very well written and fair...
1) Wouldn't interpolating re-sizing the canon image from 22MP to 36MP potentially destroy some of the original data and increase the SNR in an unfavorable direction when trying to recover data from the shadows? Could any additional color noise or banding been introduced as a result of the resize?
2) Would shooting with a Canon Mount Zeiss have made any difference over using the Nikon Mount version with the EOS adapter? Could this have introduced any unwanted characteristics?
While I don't believe he should have reduced the D800 file, as I don't think that would be a fair comparison either, I would like to have seen a recovery performed on the original image, not interpolated. Just for comparison.
It appears normalization is the bane of any comparison between these two bodies because the MP gap is just so large.
I'm asking these questions, because I honestly don't know the answers. I'm making an assumption based only on my experience with reworking images.