ForgReaver, would you mind linking each build to the post where it was introduced? Compacting the justification for damage into a single line is insufficient and hard to read.
Personally I would have allowed the players 2 rounds and to take the average of both rounds. This would have allowed for a bit of buffing, but both are biased in some way. Personally I found sustained damage more helpful than nova as sustained damage pushed towards having resources later in the day. This is going to be difficult to model as some tables frequently have 1 encounter per day, where nova would be king, while others are closer to dungeon crawl where a nova build will have sacrificed endurance for the nova itself. A way to get a mix would be that the build needs to explain what they do for 2 encounters for 2 rounds and they take the average or median of those 4 rounds. Average would bias towards nova, median would bias towards sustained damage.
You made reference to Damage Kings. As author of the DPR King Candidates thread in 4e I've had some experience doing this sort of thing. You said you were not wanting players to simply one up eachother with a tweak that ends up being unplayable, yet the top entries in your list are one trick ponies nonetheless. I found that you allow these tweaks and rank them all the same, then append the user's name on the entry. The competition ended up being healthy for the optimization and community as new exploits were found.
Another problem I found with my DPR King Candidates was differing interpretations and cheese. Like you I realized there was no way to have the same rules apply to all builds. I found that labeling the cheese or side of an interpretation was helpful as many builds ended up relying on them. I then added these labels as tags to the build so one could quickly filter out builds that wouldn't work at their table. I often had to make DM-like calls on some of these interpretations. Don't be afraid to do so here.
I noticed you didn't account for area damage. According to your rules I could just make the assumption that I'm dealing with 21 HP mooks all in a giant ball and wreck the rankings. You would counter saying that is not an assumption that would be universal, yet I've cited my assumption. The way I tried to account for this was to separate out a build's single target damage from their area damage and annotate the per-target area damage with the area it applies to, then cite the additional single target damage. I was unable to find a satisfactory conversion from area to targets. For ranking it I had to guesstimate.
Lastly have you considered having a section dedicated to guiding people through the DPR analysis. Advantage and Critting was one thing that many didn't find intuitive to calculate. I would go so far as normalizing for level by dividing the expected damage by some HP(level). I called this KPR (kills per round) and found it helpful putting builds on the same platform.