they want balance as it relates to DPR. That desire is the direct cause of their dissatisfaction. They have a problem because DPR is their main concern.
If the players didn't worry about parity in the DPR area, then when the multiclassed warlock/sorcerer used his Exploitation Blast to do 846 points of damage in one round, the other players would simply say "wow, impressive" and there would be no issue.
They want DPR builds to balance with DPR builds, that's not the same thing as DPR being the main concern, or they'd be all over trying to boost DPR outside the weapon users who have little else to contribute. Instead, the excess DPR of the feat-optimized fighter is viewed as merely rendering the class viable.
Tony Vargas's post seems accurate to me.
But even suppose that the OP and the OP's group did have DPR as their main concern.
What would be wrong with that? It's obvious that balance in damage dealt is an important design consideration for 5e (there's no other reason why the spells no longer have the canonical damage ranges, but rather have all these weird damage expressesion, like a fireball cast by a 5th level character doing 8d6 rather than 5d6). And probably no single rules element gets as much attention (in PC design, in resolution mechanics, in spells and items, in monster design) as the rules for action economy, attacking, saving throws and inflicting damage.
I read the Mythological Figures column every time it is posted. And every time it's just a combat stat block. King Arthur is a set of combat stats for a 19th level battlemaster/paladin. Lancelot is a combat stat block for a 13th level champion/paladin. The only person in those threads who offers up Ideals, Bonds and Flaws for these personalities (which at least might start to flesh them out as the mythological figures that they are) is me.
If so many people are playing D&D essentially as a type of wargame - for which there is a very long pedigree, going back to the classic game - then what is wrong with focusing on DPR?
the -5/+10 aspects of those feats...and the ones that grant extra attacks, as well...don't really add anything to a character concept. I can make a Robin Hood type character without them. No one looks at a Robin Hood movie and thinks "wow he must have done +10 damage on that shot!"
The -5 actually fits neatly with BA: without BA, it'd be too great a penalty at low level, and trivial at high.
Here's my take on some aspects of the design challenge around -5/+10.
First - and contra the claim that "no one looks at Robin Hood and thinks that he must have done +10 damage" - there is an expectation that weapons can kill. That Robin Hood can take down a guard with a single arrow. That Conan can cleave an enemy in two with a blow from his axe. How to account for this in the context of a hp-based game? Allow those weapons the chance to get a damage add, which lifts their max damage to 20-odd, which is enough to kill an orc (hp 15) or a guard (hp 11) or a scout (hp 16).
(Other builds get different stuff: shield masters get to do protective stuff; fencers maybe get to be rogues?)
Second, the significance of the -5 to hit is incredibly variable. The OP regards it as just a speed bump on the way to optimisation. But somewhere out there a relatively new player is just now working out that, with a well-time buff from a friendly cleric or bard and manoeuvring for advantage, s/he can offset the -5 and get a real lift in damage output! For that new player, that's an enjoyable experience of learning to play the game better.
Is the game being designed for the OP, or for that new player? It's not easy to design for both, because building in scope for the new player to discover new tricks is at one and the same time building in break points for the player who can see and routinise those tricks on the basis of his/her experience.
This second half of this post ends up returing to the first half. It's possible to have a RPG in which mechanical tricks - playing with the dice and numbers - is not an important part of play. RPGs which aren't, and can't be, wargames. Those RPGs exist, and I probably have more familiarity with them than many ENworld posters, though not as much as some others.
But D&D is not one of those RPGs. It's chock full of opportunities for mechanical tricks - exploiting buffs, optimising action economy, finding clever ways to minimise damage taken, etc. The design doesn't just allow for that, it actively encourages it. So it's no mystery that some tables take this seriously. But it's hard to design a system of that sort that will be robust in the hands of both amateurs, who only stumble onto occasional tricks, get a buzz out of them, but don't systematically exploit them; and in the hands of "professionals" who systematically seek out those tricks and want to routinise them.
I was never a terribly good MtG player, but had friends who were national champions. Playing with them, and with their decks; compared to playing with random members of the University RPG club; was like apples and oranges. It wasn't the same game. D&D, on the other hand, is the same game - the OP is building PCs from the same build elements as the newcomer - and it's remarkable that it's as robust as it is across those differing sorts of play environments.
Ultimately what 4e got wrong and 5e got right was being genuine to D&D, and that means being 'bad' in a lot of ways that D&Ders are just fine with... 
