To play devil's advocate, what's the difference between a player showing good tactics in a combat (especially theater of the mind combat), which can have mechanical effects (more targets in an area of effect, positioning) and a player show "good tactic" in a social situation (correct levels of politeness, good cop/bad cop, flattery, proper accent, etc.).
A player showing "good tactics" in combat that have an in-game mechanics impact is showing skill at "the game system".
Any player can learn the game and learn how combat works and make good choices or bad choices for that and get the same benefits.
A player does not get "mechanical advantages in combat" for describing well how his character is making the attack, how his character's sword strike is curving around the shield, etc.
Exact same thing for a social encounter - especially as seen in play many times in my game -
Getting "help" by say getting someone else the NPC likes as a backer - gain advantage.
Arming oneself with info from prior "research" - no different from doing research to show you that trolls need to be burned or that Gruesomes fear silver. Likely also results in advantage but might even move to auto-success depending on the extremity. These are the social equivalents of using terrain.
But, neither of those rely on how well the player decides to "flatter or intimidate" or describe how he does it or chooses *his* words as a player.
The players are going to benefit from knowing how to play the game and how their characters work and play to those strengths - even when - especially when - that means handing "go get me an edge" to an ally by research (or by standing within 5' on an enemy so the rogue's sneak hits.)
So, your knowledge of playing the game plays at all times - but your own personal player-side IRL knowledge of "how medicine works" or "how allista work" or "how actual real magic spells work" do not get into the resolution process.
There is a ton of abstraction within the game's resolution system and it hinges resolution around the character and the challenge and the environment/scenery - not the player knowing the "skill".
In my games - the "describe how the swing goes" is part of the result not part of the attempt.
Then again, for non-instant tasks i always use three-way tasks resolution (akin to death saves) and that encourages multi-faceted approaches and even multi-person efforts.
As an aside: have a highly charismatic character in one of my games. Character is an unabashed rake/flirt. Player often expressed things as fairly direct double-entendre - sometimes single entendre.) The resolution determines what the character says and how it is received - not the player's phrasing and word choice and ability to keep a straight face. That may mean others observing from the side may wonder "how did that line work" when the gentleman takes it in a humorous or intriguingly direct way. In fact, that same line trying to be delivered by their non-proficient and non-charismatc 16 character might well fail, badly.
thats because it is the character who performs the task and delivers the pitch - not the player.
just like i do not make the player sing when their character performs and use that for resolution.
Edit to add - noticed did not specifically address the examples of AOE blasts and positioning... it should be obvious from the above but those are examples of a player knowing how the game works - what the radius of his spells, how movement thru enemies or AO work, are etc and how to apply them. So, no "player IRL knowledge of the task" applies.
Oddly tho, while ToM was mentioned, both of those seem to be more applicable to grid-based play *or* character-based skills since ToM doesn have players choosing exact positions on any actual sense, just seeking results in character "she tries to catch as many enemies as possible" or "she tries to block the approach" etc which would come back to "character traits" not what/how well the player says things.