I was about to say, the actual solution to this problem is to introduce degrees of success (and its less-talked about no less-important counterpart, degrees of failure) to skill checks. It makes every point, every bonus actually matter. Also, helpful: success-with-consequences, but that's for solving a slightly different issue (bottlenecking) springing from the same problematic mechanic (binary success/consequence).
That makes sense. It was composed late at night from bed and is a lot more disorganized than I meant it to be . my point involved a comparison between the previously discussed self-obsoleting skill system of 5e, which has been discussed in previous posts, and one version of what existed before 5e. It somewhat depended on the linked text of the subsystem being known
A favorable circumstance gives a character a +2 bonus on a skill
check (or a –2 modifier to the DC) and an unfavorable one gives a
–2 penalty on the skill check (or a +2 modifier to the DC). Take
special note of this rule, for it may be the only one you’ll need.
Mialee runs down a dungeon corridor, running from a
beholder. Around the corner ahead wait two ogres. Does Mialee
hear the ogres getting ready to make their ambush? The DM calls
for a Listen check and rules that her running from the beholder
makes it less likely that she’s listening carefully: –2 penalty on the
check. But one of the ogres is readying a portcullis trap, and the
cranking winch of the device makes a lot of noise: –2 modifier to
the DC. Also, Mialee has heard from another adventurer that the
ogres in this dungeon like to ambush adventurers: +2 bonus on
the check. Her ears are still ringing from the shout spell that she
cast at the beholder: –2 penalty on the check. The dungeon is
already noisy because of the sound of the roaring dragon on the
level below: +2 modifier to the DC.
You can add modifiers endlessly (doing so is not really a good
thing, since it slows down play), but the point is, other than the
PC’s Listen check modifier, the only numbers that the DM and the
player need to remember when calculating all the situational
modifiers are +2 and –2. Multiple conditions add up to give the
check a total modifier and the DC a final value.
Going beyond the Rule: It’s certainly acceptable to modify
this rule. For extremely favorable or unfavorable circumstances,
you can use modifiers greater than +2 and less than –2. For
example, you can decide that a task is practically impossible and
modify the roll or the DC by 20. Feel free to modify these numbers
as you see fit, using modifiers from 2 to 20.
The problem is that presentation matters. Once the rule books discuss degrees of success in concrete terms, players expect degrees of success to exist. The very fact that it would be described in the book will alter player behavior towards mechanical over narrative thinking.
In pure mechanical terms or game terms, the existence of degrees of success is perfectly fine. Mechanically, it's not broken at all. The trouble is caused by describing degrees of success in a player-facing book.
This is why it was important to rename Thief to Rogue. It's also why when you sit down at a D&D 2014 table, you assume that Drow Assassin is an available option, while at the same time you do not assume that a Aasimar Death Cleric is even though they're both in the core books.
A game is about the things it chooses to spend time making rules on, but TTRPGs are weird because they're also not. TTRPGs are actually about role-play, so they have to be incredibly careful about how they present mechanical rules. If you spend too much time on combat rules to the exclusion of everything else, then the game looks like it's entirely about combat. 4e D&D made this exact mistake in presentation, which is why it earned a reputation that it wasn't about roleplaying at all. The game spent about 310 of it's 330 pages in the PHB on abilities that were all but exclusively relegated to combat. I guess the game is 95% combat then?
Going to extend on @EzekielRaiden's point about other systems by showing that 3.5 had degrees of success and degrees of failure before I get into this
I think that you might be working under the idea that degrees of success require anything & everything to have a roll, and that only rolled checks hold significance later on, but that's just not accurate... 3.5PHB pg65 has a section for "checks without rolls," and I wouldn't be surprised if some 4e book mentioned something similar.
In both of those past versions, we had a rules subsystems that enabled almost anything to matter a little (or a great deal). That included elements that were initially introduced as seemingly inconsequential fluff at the time & past actions that were too trivial to require a check at the time. However, in 5e, we really only have (dis)advantage, which comes with a massive all-or-nothing bonus or penalty too big for casual details & trivial checks that needed no roll to matter. On top of all that, the skill system tends to self-obsolete after a few levels & once that happens we wind up with any given possible check effectively being a matter of if the GM will allow the player to accept their guaranteed success or if the player needs to find an acceptable loophole in the block by fiat.
Degrees of success and Degrees of failure are highly adaptable without necessitating a roll for every little thing, but the system itself provide a solid framework for it to work effectively. By dropping the ball on that framework in so many ways , 5e itself forces rolls that on things that previously would not required anything in the past.
. my point involved a comparison between the previously discussed self-obsoleting skill system of 5e, which has been discussed in previous posts, and one version of 
