Realism and rules light games

So what is realism in an RPG? Is it second by second simulationist perfection? Or a very good description of the setting, character and events? What is realism in a game that considers magic real? What is realism in a fictional sci-fi setting?

We could go into a lengthy conversation as to what realism is. Is it representing a characters ability to jump correctly? The precision of the bullet? The technobable involved in the game? The fact that a long sword really does deliver 1d8 HP of damage? What is a hit point?

I think the issue with realism comes when a player is happy when his magic user's fireball goes off and kills the orcs (as if fireballs and orcs were very real), but complains of suffering an attack of opportunity that could have been prevented. The player then goes on complaining how that's not possible, it should be evaded, that's not true, in real life his character could have dodged the axe. In real life his character wouldn't have fireballs. So what is real? What is realism in an RPG?

How do we define real in a game who's setting isn't real in the first place? Should we use a better term? A term like consistent? If a game is consistent the players and GMs know what to expect even if the setting itself isn't realistic.  What is the game rule's role in maintaining this consistency and what is the player's mind role in maintaining this consistency?

Consistency can be understood as the game rules contradicting themselves or not. If no contradiction exists the game is said to be consistent with itself. Clearly the bigger the rule set the more prone it is to have a contradiction. It has two or more rules that contradict each other. It is clear then that rule light games are more prone to be consistent than rule heavy games. A poorly designed rule light game can still have contradictions, it's simply easier to have them in rule heavy games.

We can also talk about consistency between player expectations and outcome. By this I mean is the outcome of a certain action what the player expected or is the player constantly taken by surprise by apparently incoherent outcomes to the character's actions. Is the game consistent with the player's expectations? Although it is a valid usage of the term, I prefer to keep consistency applied to the rules and use the  term plausible when talking about player expectations.

Instead of using the term realistic or consistent let's think for a moment about "plausible". Is the outcome of an action expected? Magic isn't real, but the effect of a spell must seem plausible for the game to be accepted by the players. In combat an attack or outcome must be plausible if it is to be accepted. A monster that swings at everything that passes by, while possible by the rules, is hardly plausible and being against the player's interests, a highly disputed by the players. The more plausible the game is, the less game flow will stop to challenge the GM, look up rules, check up tables, etc. Plausibility benefits play.

So to conclude, I'd like to lay out the following points I'm using to better describe and design games:

• "Realism" is a term that should not be used to describe a game, instead "plausible" gives a much better term for what is commonly meant by "realism". 
• Consistency is another term that should be used to describe a game, since an inconsistent game will surely be less plausible than one without internal contradictions. 
• Consistent and plausible games run smoother than less consistent and less plausible games. They're more fun!
• Rule light games are easier to keep consistent than rule heavy games.
• "Plausibility" is not only enabled by the rules, but also by the GM. Being the GM a human being he or she is clearly benefited by less rules than more rules, as there is less to keep track of, less to stop and check and it is quicker to learn. Leaving more time to concentrate on the game and creating a plausible adventure.
• It is of uttermost importance to select the essential rules to write. Any small set of consistent rules will not automatically ensure a plausible game. The rules must be cleverly selected to represent the essence of the game. They are to be the base from which all other dynamics in the game emerge. 

This final point is what makes creating a good rule light games hard. A designer must distill the essence of the game and write it in a way that allows it to grow beyond the bound of the written rules, but still remain playable as the characters gain power. A shortcut to this is to bound the game's "power range" for example characters can go up to level 6 or have so many power points, etc. This binds the game to a range within which it remains stable. To let it move beyond it requires a set of rules that self regulate the power curve of the game by means of some negative feedback.

Thoughts?