Regardless of whether I agree about the statements being made regarding the CR system (I don't agree obviously), I think that if you are going to try and come up with a more accurate system (which you could try and do)... then your first step should not be to ignore powerful magical items, and assume 4 players. Weren't those the things that everyone mentions as changing the difficulty?
If you want a system that accurately tells you if your encounter is challenging enough, then you need to factor in all the variables, and not neglect them. So include all levels of magical equipment that the party could have, and all numbers of players. You may even want to add modifiers based on whether certain classes are represented in the party. Having a paladin or cleric in your party, for example, will no doubt make encounters with undead a lot easier. And what about difficult terrain? Do the opponents have a height advantage? Do they have easy access to call reinforcements? Are they in their natural terrain?
If you want a more accurate system, you can't ignore those details. Go all the way, or don't bother.
sigh...Our entire society, including this forum, your computer, and virtually everything you have, was invented and constructed using the scientific method. Its basis is the controlled experiment, where you set values for all variables but one, then experiment to determine the value of the remaining variable. It is especially useful in complex situations like this. So, at the outset I set several baselines. Thereby allowing us to have a discussion and possibly get some results. It does not preclude you from adjusting them based on your relative skill levels, the amount of magic in your campaign, or anything else that you have different in your campaign. But it gives us a starting point.
Its like saying you cant make a household budget because you dont know for sure what expenses you might have, and that unknown expenses might come up, etc. (OMG!!! There's too many variables!!! I cant do it!!!) Of course you can.