Depends on the game. If your "optimizer" has charisma 8 and no persuasiom and tries to sweet talk an NPC with "good roleplaying" , then still let them roll for success.
Some good roleplaying powergamers can afford to just optimize combat because they are good at talking and their 8 charisma is often forgotten. And if you as a DM don't enforce rolls, their is no downside to optimizing for combat.
But, again, the critical thing is that failing a social roll almost never results in losing your ability to keep playing the character. Even failing a complex social scenario almost completely doesn't. Because the main way to take away the character after a bad social encounter....is to have them get into a fight. Which they're already optimized for.
Keep in mind, I
do care about this sort of thing, and as far as I can tell I have done a fairly good job of getting my players to care too. I'm talking about players who
don't care. Who only care about survival value. How do you give those flopped social rolls bite? How do you make it so they have consequences the player will care about? You must threaten their ability to keep playing their character. How do you do that? By threatening bodily harm or the loss of material goods. How do you cause bodily harm or take away material goods?
Violence.
5e, like most editions of D&D,* simply doesn't offer grounds under which serious threats to the character's wellbeing can be
mechanically articulated outside of combat. And because they can't be mechanically articulated, simply
inflicted as a punishment, they're going to come across as arbitrary and punitive in many, perhaps most, situations. Optimizing for combat means optimizing for dealing with the most significant punitive thing the DM can do besides rug-pulling and "rocks fall, everyone dies" type stuff, which would be dealt with via the social contract regardless.
That's why you don't see a more mixed presentation of this stuff--why it almost always centers on combat, even though there are lots of things to optimize for. 5e, like most editions of D&D,* doesn't actually support using it for situations where the greatest threats the player characters might face have nothing to do with physical violence.
*There's a reason I use this phrasing.
This is why I still maintain that a lot of posters on this site (or at least several prolific ones) are big 4e fans.
Regardless of your perspective, the rules are still considered more important than the story, such that the fiction should be molded so that the rules effect makes sense. I dont recall any advice suggesting otherwise, and back in the day when I asked about this stuff, people would encourage me to literally make something up so that the rules effect remained unchanged. I found that extremely irritating, and I still do. Its one of the reasons that I left 4e after playing it for about 18 months.
Anyway, bashing 4e is not my intent. I just didn't care for it personally.
It took me some time to ruminate on this, but no, you are not correct here. The rules are
not considered more important than the story--or at least that wasn't the intent. Nor was the intent to have the fiction "molded" around the rules. Because that, again, puts the rules absolutely, exclusively first, and that's incorrect.
The idea is that the two are
simultaneous--and that you should, ideally, be able to think about things in narrative terms and derive correct mechanical conclusions, or in mechanical terms and derive correct narrative conclusions. Hence why I love the LoH structure--because, while I have no actual documentation to back it up, I am completely confident that they sat down and asked, "How would we make a power that has the Paladin sacrificing some of their own health to heal someone else?"
If you have the fundamental underlying math down, so that it really,
truly works, in straightforward and well-explained ways, then the difference between thinking about things in mechanical terms and thinking about them in narrative terms blurs, and hopefully disappears entirely. There ceases to be any meaning to the idea that "rules come first" or "fiction comes first." To think about the rules is, by definition and without separation, to be thinking about the fiction; and to think about the fiction is, by definition and without separation, to be thinking about the rules.
This is why I am such a fan of both Dungeon World and 4e and its relatives (like 13A.) They reach the same point from opposite directions. DW abides in the fiction and very, very carefully constructed and tested its rules until following them was, in all but highly unusual cases, identical to whatever the fiction actually called for. 4e built a robust, scalable rules framework, and then carefully constructed rules elements so that using them is (or should be) the same as telling a story about them. Rules only "came first" for 4e in the sense that the mathematics were nailed down by the designers and made consistent
in order that storytelling would be as close to 1:1 equivalent with invoking the rules. Once you, the end user, are invoking them? Telling a story should be using the rules and using the rules should be telling a story. Consistently.