I agree with MustaKotka, but I also think the problem extends further than that.
The biggest issue with determining a power level is that EDH as a format has such a variety of players that want different things.
The clearest distinction for what I mean is EDH vs cEDH.
In EDH, people are usually looking to play a good game of interactive magic, playing powerful cards and doing cool stuff. It is not uncommon, nor unwanted, for games to go 10+ turns even without hard stax effects. When people play like this, it's not really a big deal if you play your commander out a bit later because a) they have a high cmc or b) it better suits your
game plan. But this kind of play style simply will not cut the mustard in cEDH.
Personally I feel that comparing cEDH and non-cEDH decks is like comparing lemons and limes. They're similar sure, but ultimately different. I think it would probably be worth having separate systems for ranking cEDH and "regular" EDH decks. I think the criteria for what makes a deck effective and powerful are different enough that there could be two systems. There could even be a pre-cursor check list for the cEDH list. A list of criteria (to be defined) that a deck must fulfil to meet the requirements before it's even assessed to stop people overestimating their deck's power level, assuming it's a cEDH deck or close to one and scoring it incorrectly.
I touched on this earlier in the thread when I suggested the idea of weighting categories based on priority, but I think that remains one of the biggest issues that these kinds of scoring systems need to get around. Different decks are going to care about the same categories different amounts.
As MustaKotka points out with his HoG example, there will always be times when a given deck will answer a certain category in a way that belies it's power level. HoG will likely never be a cEDH deck, but has the potential to be
reasonably powerful. However, it is always going to boss the "When can you get your commander out by" question.
TL;DR
1) There is an argument that cEDH and EDH decks can and should be assessed, ranked and scored with different systems.
2) Being able to make scores contextually relevant to the deck so that it's score is based on areas that the deck wants to prioritise and focus on is still a fairly major hurdle as there are too many edge cases that
undermine regular questions.