This success to failure ratio has been reflected in our Wild Talents game, as Pat noted. I'm the GM, and Pat's right: these high level characters succeed a lot, like 95% of the time. I've generally found this to be pretty satisfying so far. This dynamic allows the players to drive the story forward where they want it to go by choosing what types of rolls they want to make then, and then having the story travel down the avenue opened up by their success. This also is fun for me, because I get to make fun stuff up about exactly how they succeed and thereby steer the story within the frame I've come up with. (I particularly like it when the scientist succeeds on any of his multiple different types of knowledge skills, because I get to make up wonky sci-fi explanations about why the world works the way it does.)
But we are missing the element of uncertainty that flows from the clattering of the dice. One of my friends who I used to play with (in law school, of course) is a bit of a gambler, and the only reason he played was to throw those bones and hope to catch lightning in a bottle. Without the uncertainty, the electricity's all gone.
Here's our theorized solution (with a prelude): We're using the one roll engine mechanics, which not only tells you in a single roll if you succeed, but also how precisely or efficiently you succeed. Moving forward, I'm going to focus on the narrative consequences that flow from inefficient or imprecise success. If pc's don't succeed precisely, consequences will flow. For example, if the diplomat manages to gather facts for an upcoming trial, but does so imprecisely, I'll say that he pissed off some other diplomats...and that'll be something the players will have to deal with.
From all this discussion, you may be able to tell that this game has really been challenging for me as a GM. I'm very sensitive to the railroading issues Pat discusses below, and I try to avoid them at all costs. I'm also trying to let the pc's shine because it's a high level game, but I want to put them in some tough situations (which I haven't yet, but trust me, it's coming soon - third session is coming up.) And there's other usual GMing stuff, like keeping everyone involved, giving each character a chance to shine, being fair, keeping up pace, etc. It's a fine balance, and it's become clear to me that context really matters. I think the system we're using is really cool, but it's definitely making me think about GMing from new angles.
Failure's never been so much fun.