"Let's say that you're playing a character who the rest of us really like a lot. We like him a whole lot. We think he's a nice guy who's had a rough time of it. The problem is, there's something you're trying to get at with him, and if he stops having a rough time, you won't get to say what you're trying to say.
Our hearts want to give him a break. For the game to mean something, we have to make things worse for him instead.
I'm the GM. What I want more than anything in that circumstance - we're friends, my heart breaks for your poor character, you're counting on me to give him more and more grief - what I want is rules that won't let me compromise.
I don't want to hurt your character and then point to the rules and say "they, they made me hurt your character!" That's not what I'm getting at.
I want, if I don't hurt your character, I want you to point to the rules and say, "hey, why didn't you follow the rules? Why did you cheat and let my guy off the hook? That sucked." I want the rules to create a powerful expectation between us - part of our unity of interest - that I will hurt your character. Often and hard."
So what I think Vincent's saying here is that the point of game rules is to make things happen at the table that no one wants to see happen, but players are invested enough in the rules of the game that they go with it anyway.
It's that bit in the game where a character is on the brink, hanging off the edge of the cliff, and everyone's cheering for him to drag himself back up, but they know that if that die comes up a 1, he's going over.
It's that bit where the couple is fighting and hurting each other, and everyone's hoping that they'll stop it and get back together, but the dice go the other way, and everyone's heart breaks but they go with it.
That (Vincent argues, and I agree), is what conflict resolution is for.
"Radiance", (link is to pdf) a free game by M.J.Graham, is designed with a different philosophy, I think.
What the game's rules do basically is apportion out narration rights to various players. One player describes some stuff happening, other players suggest complications or "redirections" that may occur, and you draw beads from a bag to see if the redirection is relevant or not. If it is, the player who introduced it gets to narrate what happens next. There's no direction as to how the "redirection" affects the story, just a change in narrator.
So the game doesn't introduce anything into the fiction that the players don't want to be there.
But that's good, right? It means the only stuff that happens is stuff that the players want to happen!
I don't agree. Vincent again (from the same essay):
"The only worthwhile use for rules I know of is to sustain in-game conflict of interest, in the face of the overwhelming unity of interest of the players. "
In other words, if your game rules aren't providing some tension in play, if they're not "pushing back" against the unity of interest of the players, they don't achieve anything. I think what "Radiance" boils down to is M.J.Graham giving you permission to tell a story with your friends.
Let me expand on that. What the rules of "Radiance" do is apportion out narration to various players. They tell you who gets to speak at what time. But there's no rationale for why this particular configuration is better than any other. What does distributing narration around the table add to a story? Why do you need a bag of beads and some candles to do that? If you're going to tell a story with your friends, why would you bother with this, rather than any other way of telling a story?
The author reenforces this impression at the end of the document in a "Q&A" section:
"Why don't you give more tips or instructions on how to make a story?
Because I don't think I know more than you do about creating good stories. I certainly don't know more than you do about creating the kind of stories that you enjoy. Besides, there's nothing I can tell you that you won't pick up faster and more completely than by creating your own stories. "
So what's the game for?