The Role of Failure

Failure doesn't happen very often in D&D. It's not really something we enjoy, after all. Players feel like a combat encounter was close if one or two of the PCs was unconscious and dying for a round or so -- nevermind whether the actual outcome of the encounter was ever really in doubt.
Here's another statistic from my trademarked "Bureau of Totally Believable But Made Up Statistics": less than 1% of all combat encounters of D&D that are played result in a true "failure", a TPK (Total Party Kill).
Think about it. How often have you, as DM, wiped out your entire party. If you've done it more than once or twice over the years, take a long look in the mirror, you might be some sort of asshole. I'm just saying.
When skill challenges were first introduced when the 4e was first released, they were quite difficult -- to have a reasonable chance of success, the PCs needed to be able to assist each other and focus on their best skills, etc.
When the rules were revised, the odds of success were much higher. And a party of players who were in the habit of supporting each other should have very little trouble succeeding at a skill challenge appropriate for their level -- just as they should have little or no trouble completing a combat encounter with monsters of their level.
That's probably the right choice to have made. In general, players don't like to fail, they're going to have a better time if they win their battles and succeed at their skill challenges.
At the same time, we don't pit the players against combat encounters that are always "fair". We regularly challenge our PCs with combat encounters that are two or three levels higher than their character levels -- and frequently a major boss fight will be higher than that. We will throw in complications like terrain that favors the monsters without worrying that it's going to make things challenging for the PCs -- they find overcoming those sorts of complications, and beating those overclocked encounters, fun.
And if they don't succeed beating that overclocked encounter, they don't feel so bad for failing because they know the deck was stacked against them the whole time.
We need to feel free to present the same sort of occassional, overclocked skill challenges (assuming it's reasonable in the game's story). Convincing the king to loan the party a war galley to take on an adventure should not be a level approrpiate challenge for all levels. It's probably the sort of challenge that should be a low-paragon level challenge at least.
Think about it: using the skill challenge rules we might be tempted to design an encounter in which a party of 5th level adventurers must convince an innkeeper to let them borrow a wagon and an empty beer barrel, and a second encounter in which they must convince a king to allow the party bard to marry his daughter. Even if we decided that both would be hard checks, the DCs would still be the same.
So, obviously, one of the things we're going to need to think about is how to set the level of the challenge -- and we need to be prepared to create a variety of those difficulties for the same reasons we do a variety of encounter difficulties.
I'll write more about success and failure -- about how to frame ande define the success and failure of a challenge, and how to use degrees of success to enrich a skill challenge, in later entries.
As a final note, I want to make sure I make it clear that one of the cornerstones of what I think we need to embrace as we develop our ideas about what skill challenges look like is the idea of variety. There are no "right" answers that fit all situations. So, when we ask the question "Should skill challenges be set up to make it likely the PCs will succeed?", the answer is not "yes" or "no" -- its "maybe" or "that depends." Not every combat encounter is a pushover, but some are. We should be able to say the same for skill challenges.
So, the better question to ask yourself, as you prepare a challenge, is "Would this challenge serve the story best if it's easy, hard, or somewhere in between?"