Ah, dissent. Saelorn, this is why you're a valued member of the Modos RPG community.
That seems really vague, though. For most tasks, success and failure are clearly defined. With a system like "favorable" or "unfavorable", it seems like a lot is left up to GM interpretation, which can create an issue where the player thinks a particular goal has been achieved and the GM ends up narrating it in a completely unexpected manner.
I have to disagree with this first premise, and suggest that only tasks found on tables and meta-rules have clear lines of success and failure. If a table lays out your outcomes:
[TABLE="width: 500"] [TR] [TD]Roll[/TD] [TD]Outcome[/TD] [/TR] [TR] [TD]1-5[/TD] [TD]Spell fizzles[/TD] [/TR] [TR] [TD]6-10[/TD] [TD]Spell hits caster[/TD] [/TR] [TR] [TD]11-15[/TD] [TD]Spell hits target[/TD] [/TR] [TR] [TD]16-20[/TD] [TD]Spell doubles damage[/TD] [/TR] [/TABLE] well, each outcome is subject to interpretation, but yes, there's a clear definition that you will have one of four outcomes. Or if your rule is based on rules (a meta-rule), then that could also have a clear definition. For example, when a player's attack roll exceeds his opponent's - sorry - equals or exceeds his opponent's armor class, then the player is allowed to roll damage and apply that damage to his opponent's hit point total. That's pretty clear. What, exactly, that represents in-game is not always clearly defined.
So what are the other clearly defined tasks? A survival check to see if you salvage enough food for the day? What if you collect more than enough food, but half of it goes rotten? Is that a success or failure? A perception test with four stunt-points to read a photo's inscription? Success could be simply noticing a marking, understanding half of the words, or being able to read the secret code written in fine characters between the words.
That's a problem with any game where "complications" are part of the success table. Note that with D&D, in my above example, the specific checks would involve riding and pushing an animal for an extended period of time, and the PCs can be reasonably assured of the likely results before any check is made - if it took you six hours to deliver a message, but the opponent got there in four hours, then it's because you didn't push the mount hard enough (or pushed too hard and exhausted it). One of the major benefits of a system like D&D is that results are usually clear, and the DM is rarely left uncertain about how to narrate them.
D&D accomplishes this with a lot of tables, and math (see the jump skill). Nothing wrong with that - some people like books and calculators with their RPGs. I think it's possible to achieve more immersion and faster gameplay with a more vague system like I've suggested. Complications are one way to do that. One of my motivations, though, is to take out critical fail and critical success from the ends of the outcome line, because these make degrees of success implicit. And... (now for a little math)
...if a PC has a 40% chance to accomplish a certain task, each "success" die roll should result in the exact same outcome, since (on a d20, anyway) each die roll has the exact same chance of occurring. The same goes for the "failure" rolls. A 1 is a 3 is an 11. Now, this still presents a glaring problem: there are two types of outcomes on the d20, and they are essentially -complete opposites- (succeed or fail). If you change those outcomes to favorable or unfavorable, they're still opposites, but I think the difference is softened a bit. If you roll an 11 or a 13, a 1 or a 20, your outcomes could be very similar but with the key difference: one is favorable, the other is unfavorable.