I'm not "failing" to understand anything since the point is hardly some revolutionary complex theory. What I am saying is that, when Federer dominated in 04-07, nobody else was even close. It was dominance unlike anything we've ever seen (which I'm assuming is what you were alluding to when you said his dominance was more "dominant" than Novak).
As such, it is totally flawed to compare it to an era in which multiple players dominated. In other words, if you're going to say Novak dominated from 2011 and 2016 because even though other players dominated as well, you can still be dominant while having someone more dominant than you (which for my money is extremely debatable), then I'll just as easily claim that Roger's dominance was from 2004 to 2010. Yes, Nadal was more dominant between 2008-2010, but Roger dominated as well (using your logic). Roger won a slam in 2008, reached two other finals, won 2 slams in 2009 and reached two other finals (and finished as world number 1), and won a slam in 2010.
So why are we ending Roger's dominance at the 2007 mark? I'll give you 2010 since he didn't do much after winning the AO, but surely you'd have to include 2008 and 2009 in there.
1. Yes, Fed's dominance was way too much and we have never seen anything like that before. That is
exactly what I alluded.
2. Technically as per the blogger's definition of dominant stretch, Fed's dominant stretch is from 2015 Wimbledon
to 2010 AO, a span of 19 Slams. But, Novak's dominance is from 2011 AO to 2016 AO, a span of 24 slams.
That is why I said, Novak's dominant stretch was longer, but Fed's dominance during his stretch was definitely
greater than Novak's dominance during his stretch.
3. It is very clear that you are fixated on the notion that there can only be one dominant player in an year.
On the contrary the blogger's metric of dominant stretch is formulated in such a way that it can even
possibly three dominant players at the same time, while not requiring even one dominant player at
any time. However, the formulation prevents four or more players being dominant at the same time,
as it is crazy.
We could combine the two metrics of length of dominant stretch and the level of dominance into one metric by multiplying
them together. But, I don't have time right now to do so. But, I will attempt to do so soon. For example, if some one's stretch
is say 20 slams and he won 60% of max possible points in GSs during that stretch, then his figure will be 12. On the other hand
if someone else had a dominant stretch of only 16 slams, but he won 90% of max possible points in GSs during that stretch,
then his figure would 14.4 and this is better than the previous guy's performance.