Most of the challenges I listed are logically obviated on successive playthroughs. If those elements were really as important to your enjoyment as you say they are, and so many of them manifest due to a lack of foreknowledge/familiarity, that doesn't add up. If literally "half the fun" of playing a bad build is discovering it's bad halfway through the game and you've played the game over and over and over, how does that circumstance (the discovery) even arise for you?
As for attributes and bad builds, let me put it this way: in D&D 3rd Ed., let's say Str, Dex, and Con are the only three stats in the game. You make a fighter and you can have an 18 in one, 16 in another, and a 10 in the last. The things that fighter is good at will shift significantly based on how you place those stats -- but the character is still clearly a fighter. You may be able to make a convincing case that one build is markedly better than another, but they'll all give you different strengths and weaknesses. Now figure out a way to do the same thing for Int, Wis, and Cha in the core rules and that's essentially what we're trying to do for PoE.
"No bad builds" is not the same as "all builds are equal" and "all builds are functionally the same". It means that if you distribute your points in different ways across our attributes for a character of any class, the strengths and weaknesses of the character will shift in interesting ways and still be viable. If you dump Resolve for your fighter, you will suffer. If you boost Resolve for your fighter, you will benefit. Some classes in D&D already do this better than others. Monks and paladins have a more difficult allocation of stats to consider than fighters. If you want to shift the difficulty of combat, we have a level of difficulty slider. The attribute system is not meant to be a covert way of haphazardly achieving difficulty.