Extract from GM Mastery Ezine, because I just keep agreeing.
The whole article is quite huge btw, so I've bumped chunks to comments and still left stuff out which is mostly tabletop relevant.
Guide To Adventure Writing
By James Edward Raggi IV
http://lotfp.blogspot.com/ Success and Failure
The most important thing to remember when constructing an
adventure is to not assume that the PCs will succeed at any
point during the adventure.
As a referee, your job is to be completely impartial during
game play. You have absolute power at the game table and can
bequeath success or mandate failure at any time. Doing
either of those things ruins the game, as both give no
incentive to play well.
Do not fudge the dice. Ever. Luck is a part of the game, and
the dice are there for a reason. Resist the temptation of
sparing characters that fail or even die due to "bad luck"
or a "stupid die roll."
Would it be acceptable to tell a player that just rolled a
stunning success that you've decided, just because it's more
fun, that the die roll doesn't count and he instead failed?
I don't think so. So why would ignoring the dice in the
players' favor be acceptable?
Good game play will tip the scales of fortune and those that
rely on pure luck deserve what they get - either way. At the
same time, if an incredibly lucky roll derails the entire
adventure and gives the players a quick victory, it should
stand. It needs to work both ways. When the dice go badly
for the players, they should be thinking of how to not let a
roll of the die be the sole determiner of their fates. And
when the dice go a little too well for the players, the
referee should note what he needs to do to prevent a single
die roll from determining the course of an entire adventure.
Traditional games are all about the players (and referee)
learning to play better over time. The characters'
experience gains are secondary. Demand and reward player
excellence and the game will be more challenging in the long
run.
So what are the consequences of deciding to play this way?
The party is just lost and sitting around because they
didn't find the secret door that leads to the next section
of the dungeon? Tough. It goes unexplored.
The party missed a vital clue and has no idea where to turn
next in a murder investigation? Tough. The killer gets away.
There are too many options to choose from, and the players
are disorganized and can't agree on an option and look to
the referee for guidance? Tough.
This only works if the referee is willing to realize that
sometimes, all his work on an adventure is going to be
wasted. The players are sometimes going to be unwilling or
unable to see it all. The referee must contain his ego and
resist the urge to introduce some way of being able to show
all his work off. And the referee must not take the unused,
unexplored parts of his adventure and plug them in
elsewhere, as this negates the choices the players have made
that led to them, intentionally or not, failing to explore
the areas in this particular location.
Playing this way also means that the game can "stop" at any
time because a battle wipes out the PCs, or some other
disastrous result that means the mission will come to an
abrupt end. Oh well. Of course success is always more fun
than failure. But if failure is not an option, then the
success is but an illusion, it's fake, it's a lie. And by
taking the attitude that the end result determines the fun
of the game, then suddenly the process of playing the game
is not fun in and of itself.
I don't need to say anything about how stupid that is, do I?