Ah -- that version three mix is better!
Glad I wasn't going crazy about NX being switched in. I've not used that specific system, but I've heard similar ones, and thought I recognised the artefacts. Anyway, now that's sorted, I'd agree with you that it sounds like you've been fighting to make sense of what you're hearing through NX and the headphones, because the overall tonality of the mix is very middly, with the region from 100Hz to about 1kHz dominating over the rest for the most part. That said, there also seems to be quite a lot of energy in the top octave -- on the vocals sibilants and snare in particular -- and the 2kHz zone seems a bit recessed. No panics, though. I've had this kind of thing when recording on headphones (or when dealing unfamiliar speakers and acoustics) on location, so I know how easy it for this to happen, and I know that it's possible to tackle. The main thing is (as you've already surmised) to give enough time to the mix-referencing process, as that'll make any spectral biases of the monitoring more obvious to you as you work.
As for specific suggestions for references, this contest mix was tricky in that respect because Diesel left the field wide open for new interpretations, rather than dictating the sound he was after, so it puts the responsibility back onto each mix engineer to decide what sound they're aiming for, at which point they need to find specific references to support that. In acknowledgement of that I personally threw my net fairly wide when selecting things for comparison while doing these critiquess. For instance, I referred to a couple of dozen things out of my own personal reference collection (though most frequently to Paolo Nutini's 'New Shoes' and James Morrison's 'One Last Chance') and then worked through about 20 likely-looking tracks I'd recently come across in my work for The Mix Review column, ending up with a shortlist of James Bay's 'Hold Back The River', Chris Stapleton's 'Traveller', Alabama Shakes 'Don't Wanna Fight', and The Steeldrivers' 'Hanging Around' so that I had plenty of bases covered. Of course, if I were mixing this myself, I'd probably shortlist a little more stringently, in order to decide for myself what the most relevant tracks were to my own choice of vision for the mix. It's a bit of a pain in the neck, but what I can definitely say without reservation is that it's worth the effort!
It's a bit tricky to comment about more subtle balance issues in you mix in the light of the tonality concerns, but I suspect you've got your heart in the right place, as many of the levels do nonetheless seem instinctively pretty sensible. I suspect you could, like many others, benefit from more low-end on the bass guitar, and some de-essing on the backing vocals would make sense, even with less HF fizz, as the consonants are pulling those singers too far forward for me. A bit of extra stereo width in the riff and verse sections would probably help give more of a sense of openness to the mix too, once the overall spectrum's more balanced.
Hope some of that's useful -- and thanks for sorting out the updated upload so quickly, too!