I think you misunderstood. It's a question of how you represent your data. If the top dps job (lets say SAM) does 5800 on a fight, then one point on the normalised scale is worth about 60 dps. If the top dps job at a different percentile (let's say MNK) does 1800, then one point on the normalised scale is worth about 18 dps. So even if you look at the same fight, on the same tier, the normalisation scale varies based off of percentile. In fact, the normalisation of each job varies with percentile, so it's actually meaningless to say "job X has a score of Y". Does a 3 point difference represent 54 dps, or does it represent 180 dps? It depends on the percentile. If you can't even use normalised values to assess a single job, on a single fight (because it varies with percentile and the top dps done by any job at that percentile), then imagine trying to compare two different jobs, in two different tiers, in two different expansions. Good luck with that.

As far as speedkills go, some players just have their preferences, either because they really like a given job, or because they have invested a lot of time training on it. Even in HW, you could outplay a more powerful job if you wanted to with tighter uptime and overall better play. I'm not asking who the twenty best players are. But that doesn't mean that the right player on the right job couldn't push those numbers even further. The problem that I'm seeing is that there's enough of a difference at the moment that more good players are starting to switch off of DRK because of the incentives than the ones sticking to their guns (or greatswords, as it were.)