I'd add one think to the topic of believability.
There are two kinds of believability, two ways in which it can be achieved. They come from two different ways in which we may perceive and analyze a fictional world: as a world, ruled by its internal laws, and as a work of fiction, ruled by laws of art. They are not opposites, and they are, in some cases, very similar - but they can also be extremely different.
An setting may be believable in a "scientific" sense. That does not mean it follows real world's science, but that it consistently follows its own laws, laws that can be discovered by someone who lives there. The laws may be strange, but they are there and they work. Books by Brandon Sanderson and Jacek Dukaj are great examples here. It's hard, if possible, to "overthink" such setting. A lot of fun comes from figuring out how it works and (in RPG) exploiting it. In this kind of logic, when one asks "If a portable hole placed in a bag of holding explodes, may I use it to make a bomb?" is "Definitely yes.".
A setting may also be ruled by laws of its genre; laws that treat it as a work of art. Indiana Jones, Star Wars or Pirates of Caribbean have very little "scientific" consistency, but are fun anyway. We don't ask how is something possible, or why a character acts as he does. We ask what is appropriate, interesting and fun in this kind of story. Trying to apply scientific analysis to this kind of setting is an exercise in futility. Applying literary analysis, on other hand, works. In this kind of logic, an answer to "Why can't I trip an opponent more than once an encounter?" is "Because it would be boring if you did it all the time.".
It is possible to "internalize" a genre, by explaining and rationalizing the tropes in-setting. This gives a game the "scientific" believability without losing genre coherence, thus allowing for more kinds of exploration. Exalted and Earthdawn are both examples of such approach.
Many problems we encounter in RPGs come from confusing this kinds of settings and kinds of believability, by game designers or by players. If a game is written with genre consistency, but presented as if it had scientific consistency, it leads to frustration and abuses. That's a problem that plagues most editions of D&D. A game that has scientific consistency and is treated as if it had genre consistency won't create interesting story, as it won't guide the play as the players expect it to.
There is no direct relation between the type of consistency and the game being rules-light or rules-heavy. Rules-heavy "scientific" games have rules that may be treated as (an approximation of) their "laws of physics". Rules-heavy "genre" games have strong metagame rules that enforce appropriate tropes. Rules-light games make sure that players know how the world works (in "scientific" ones) or how the genre works (in genre games) and use this knowledge in making their decisions and rulings.