Keep in mind that for many of those people like Finkel, PVDDR, Nassif, and so on, their 65% win rates (or Finkels which is I think 73% or something absurd) was against their fellow best Magic players on Earth, mainly PT players where this was tracked. Magic is not purely a game of skill, but if one person does significantly out skill their competition their win rate will be a lot higher and that can easily skew results.
This is easy to track too. The average win rate in a PT is lower than a GP day 2 which is lower than a GP day 1 for stronger players. Most online testing that better players do is against the average skill level of whatever platform they test on. But then others will pick up a deck those players popularized and have lesser results. Not because the deck isn't strong, but because the skill gap isn't there.
Case in point. Jund is fringe competitive right now. Reid Duke is knocking out 4-1's and 5-0's with it like it's trivial, which to be fair it probably is for him. The guy could probably find a way to win a game playing just 60 basic forests in his deck.If people adjusted their win-rate expectations in Magic as a whole, let alone Modern and our conversation about Modern top decks specifically, players would enjoy the game a lot more. They would definitely enjoy their options within different formats more too.