In Gygax's PHB, a ranger PC must have at least 13 STR and INT, and at least 14 WIS and CON.
In his DMG (p 100), a ranger NPC is built quite differently:
Fighter: strength +2, constitution + 1
Ranger: as fighter, 12 minimum wisdom
So it turns out that players are only allowed to play the tougher, cleverer rangers. That is a feature of game play. It is irrelevant to whether or not the setting is verismilitudinous.
All versions of D&D have manticores in the fiction. No version of D&D I'm aware of sets out rules for a player to play a manticore. So basically all the manticores in the fiction are not protagonists in the play of a game. That is also irrelevant to whether or not the setting is verisimilitudinous.
In my 4e D&D game, the player of the chaos sorcerer had abilities on his sheet that let him manipulate dice rolls. This was a gameplay technique that reflected the character's control, in the fiction, over the forces of chaos. Similarly, the player of the Deva Sage of Ages had dice-manipulation abilities that reflected the character's knowledge, in the fiction, of both the past and the future.
In my Torchbearer 2e game the players have all sorts of ways, consistent with the game's rules, to manipulate dice pools and dice results. Some of these reflect in-fiction capabilities, like knowledge and personality traits. Some are purely meta. In the fiction, the upshot is simply that
things happen of the sort that might happen to lucky people with that knowledge and those traits.
The attempt to infer from
stuff that happens at the table, in the process of authorship to
the verisimilitudinous nature of the fiction, is hopeless. And the reason is easily explained: whether fiction is verisimilitudinous depends upon its content. But the content of a fiction is independent of how it is authored. QED.