Traditionally, RPGs follow a very binary success/failure ratio. When you roll high—or low, in a few crazy games like GURPS and Alternity—you succeed; perform the opposite, and you fail. In D&D, success means killing the dragon, while failure means it killed you. While you can find a third route out—flee the dragon, barter with the dragon, subvert the binary pass/fail by co-opting a league of dragons to fight against the specific dragon you’re trying to kill—most often, the game pushes the pass/fail goals as the primary route.
Granted, you might have to take many recurring steps to get there, but in the end, enough successes equal a pass. Meanwhile, all it takes is one failure more often than not, and bam, you’re gone: one failed Climb check, one failed Save Against Death Magic, one failed Reflex save against a ray of disintegration. That’s the entire reason for the rebellion against save-or-suck spells that dominated 3rd Ed D&D (and ICONS for some strange reason), but I think the results (4e and its anti-save-or-suck balancing, for example) are just patching the problem instead of finding a solution.
An example from my recent dead Pathfinder game. When the group was “ambushed” by pterodactyls while crossing a rope bridge looming hundreds of feet over crocagator-infested rapids, one of the players tried to go all cinematic and grapple one of the dinosaurs into submission. He jumped up and grabbed a pterodactyl with a decent success (27 is, for most things in d20, a damn decent success). When he tried to get a better grip, to control this thing to go after the other ‘dactyls, he rolled a nat one. And when that happens, there’s no real way to prevent extreme, gripping failure, even with the “That’s Fucking Cool” bonuses I’d factored for him: he fell off and went plummeting into the rapids, barely surviving the crocagator attack.
Yeah, I could have swung something to keep the character on the dinosaur. Given a few seconds I probably would have come up with something. But before I even knew what he rolled, he’d decided his character had plummeted into the rapids—the pass/fail mentality is hard-coded into D&D and its mindset, since it’s a big part of the Rules As Written. It’s something I’ve seen come up time and time again.
In, say, Exalted, the player would have received bonuses to make the attempt, and wouldn’t have gone for the binary pass/fail but a degree of success—rolling a pool of dice, where 7s and up count as successes, compared to the basic D&D difficulty class, which you must beat in order to survive. And, granted, he still could have botched and fell off in Exalted—which would require him to roll zero successes and at least one 1 on the dice, which is harder than you’d think in the recent White Wolf systems.
More modern games have introduced degrees of success, such as the Exalted example above—there’s a “bare minimum” success threshold, and everything over that increases the attempt’s effects (e.g., hitting with a melee weapon and passing the required threshold = more damage). Rather than pass/fail, it’s more of a question of “how well did you succeed?” Making failure all the more interesting.
In a cinematic game like Exalted or 7th Sea, it’s also easy for the GM to justify lowballing a pass to keep a favorite character alive—you can still succeed when you rolled under the target number, but the success might not be pretty, or go horribly awry. In the above attempt, maybe the character got their foot caught in a rope tied to the pterodactyl, so while he’s not falling into the rapids, he’s being dragged through the air twenty feet behind an angry ‘dactyl. (Okay, damn, that would make a fantastic cinematic sequence, crawling up the rope to regain control of this impromptu mount.)
Lately, there’s been a rash of games which rethinks the traditional pass/fail mentality. FATE, for example. One of the big elements that turns up in different FATE games is altering the scenery or situation. A success doesn’t always have to be killing the orc; it might involve spending some player currency to alter the situation—set the room on fire, find a lockpick hidden in your boot when you need to get out of a burning room. Even moreso: instead of “succeeding” to find a hidden door, you can spend character currency, and bam, you’ve just found a hidden door. Was it there earlier? Who knows. As long as the GM allows it—and unless they’re bad at thinking on their feet, or the players are dicking with them—there’s no big reason to deny it.
Similarly, player-derived failure. In FATE, players have the option to “fail”—rather, be compelled to take immediate minor setbacks (or major complications) related to their character in return for ingame currency, and the hope/option to succeed, or succeed better, at a later point. In a sense, it’s picking and choosing your battles—losing something now for the options to excel later, when you want/need to.
It’s another form of thinking that I’ve noticed takes some getting used to. Heck, coming from the D&D mentality, I’ve noticed the majority of new Exalted players don’t want to “stunt”—perform cinematic high-risk, high-reward actions for free bonuses—for fear of failure. Or, when they do, they don’t know what makes things cinematic. Altering the game via narrative control in FATE is even more extreme, an entire new way of thinking about handling situations: you don’t have to just find the secret door, you can create one on a metagame level. And the failure thing is another leap of logic—nobody wants a complicated situation or ongoing failures, yet those are what make sessions, games, campaigns memorable: the successes that come after overcoming obstacles, the humorous Rube-Goldberg-Meets-Benny-Hill situations you find yourself in.
That narrative control was one of the turn-offs for me at first, but after playing for a while, I realized it fit my ad-hoc style of generalized planning better, since the players would feed me plots, routes, and situations to put them in, free of charge. It’s not so much a good thing or a bad thing as much it is a different way of approaching problems. And looking at how different game systems approach problems, complexities, conflict, and the like gives the GM more ideas and tools to use in their own game of choice.