It is possible to get it close to correct. I will echo some of the things other people have said before hand. You can guess that a particular encounter LIKELY won't kill people, but there is always a chance.
I find that CR worked...roughly. There were certain monsters who were just way too powerful or underpowered for their CR. They were labeled CR 5 but should have been CR 8 or 9...or vice versa. And because of the way the math worked in 3e, a particular monster could be as tough as a CR 5 against a party with a Wizard and as tough as a CR 10 without a Wizard in the party. Because one number(like say, AC) could be much, much higher than the average monster of its CR while another number(like say, Will Save) could be way lower than normal. But, the system was better than guessing randomly. Like what you had to do in 1e/2e.
4e XP system works pretty well. Although, it can strain if you push it really far. Like deciding to use the entire XP budget on 1 really high level standard monster as an entire encounter rather than using a lower level solo monster. But if you recognize that all monsters really should be +/- 4 levels, it works well. Magic items make a fairly big difference as well. If most of the party has +2 items at second level, they'll probably be able to take another level higher monster.
As for the spreadsheets. It is possible to go that far. I'm naturally mathematically inclined, so I look at these things myself. I have a number of friends who do as well. We're not as bad as the people you've listed. We're accepting of any class. On the other hand, if someone showed up with a Rogue with 14 Dex, we'd likely recommend over and over again that they change characters to one that was better. I certainly don't agree that Paladins are below the curve. DPR is not the only thing in the game.