I agree with you on 3E. CR sucked. (Or should I say, CR still sucks).
But I do strongly disagree with you on 4E.
Due to recent responses on other topics I'll start with the point that I fully respect other people enjoying different aspects and styles. And I endorse the expectation that WotC strive to provide the best game for everyone.
I see 3E CR as a terrible system, that could easily be ignored. I see the 4E system as a great system for making the math work, but a terrible system for modeling diverse, dynamic population of things to encounter. In great stories the authors are never worried about "balance" or if the math works. "You must be this tall" is a filter that excludes things. I don't want things excluded. I get that excluding things and putting brackets around the limits of things provides reliability and reduced prep time. I respect that. But there is another side and there is merit to not having "you must" or brackets or math as prerequisites for your key building blocks of narrative.
The good thing is that I can ignore CR in 3E and can do the same in any other version. But because the "math works" was so built into 4E it is reflected in the monsters with unsatisfactory results. (again, "for me")
If 5E takes a mechanical "must be this tall" approach to monster design, that will turn off a lot of people.
Except you're conflating two different elements of the issue.
4e's issue, as some people took it, was the constantly scaling monsters, and the math that created a "viability window."
But that's something apart from 4e's encounter building guidelines which took all the numbers and crunched them down into something clear and useful. It was really easy to guess the consequences of various monster compositions, and after building a few encounters of their own "by the numbers" most any DM could be confident when turning the various dials. From a DM's perspective this predictability of outcome is invaluable, regardless of edition.
The concern here is
1) if CR 2 means "this is an appropriate monster to use against a level 2 party" then the jump from level 1 to level 2 is crazy steep. Using this ogre against a level 1 party, even alone, runs a solid risk of killing one or more party members. That's an insane jump, and a huge step backwards from 4e's gradual, predictable upwards curve in difficulty as monster levels went above party level.
1b) 1st and 2nd level characters are kinda stupidly fragile, which is counter intuitive game design, and spitting in the eye of one of 4e's best received changes.
2) All the bits and pieces we've seen of the CR system don't make any cohesive or intuitive sense, so even if they do have a solid underlying system it's already a step back from 4e's instantly intuitive level/role/job monster classification.
As far as "must be this tall", the game needs to have
some measure of that, otherwise the DM's job is stacked against them. Dracoliches aren't appropriate foes for a 1st level party. Now maybe the DM just kinda knows that because, you know, Dracoliches. But what level party
is a Dracolich appropriate for and how are you supposed to know? And what about, say, gnolls? The intent behind the 5e system is that the DM's toolkit of monsters only expands: as the party levels up monsters are added to the table, but they never drop off the bottom. But, and this is a huge but, that execution relies on the game accurately telling the DM when stuff gets added to the table, and what the boundaries (and consequences) of playing near the edges are. If it doesn't, if, like 3e, the numbers are ultimately meaningless and misleading, then the system can actually be worse than saying "**** it, just guess."
Jeez. All I wanted to find out was if Black Dougal has half a chance of surviving this, and instead I'm having my nits picked. 
