A Gamasutra article by Ben Schneider about placing game players into defeat situations has sparked an interesting question.
Why is it so hard to plan for gamers to fail and present them with meaningful gameplay from that point?
I'm guessing a lot of it is the branching path effect. There are simply too many variables, too many ways for a player to fail.
Thinking about WoW, you die and you are a ghost. You can run to your corpse and rez or resurrect at the graveyard with a stiff time-limited penalty to your stats. This isn't a bad way to handle it. The player is presented with a meaningful choice, there is a penalty to death whichever case you choose (some people pshaw over a corpse run, but it's still a stretch of time where you aren't earning xp or honor). But it's hardly dramatic.
Maybe on death you'd have a random chance to enter a different narrative section. Let's say you're fighting in a humanoid camp. You get killed. Instead of ghosting your screen fades to black. You wake up in an instanced cave and have to fight or sneak your way out. Dying at this point ghosts you and you can only rez outside the instance.
It would be something different, at least. Yetis could drag you deep underground try to sacrifice you to a crazed merman.
Let's start making lose conditions interesting.
Prey handled that aspect pretty well. There's no death, just a strange mesa world with floating spirits. If you arrowed the spirits they would recharge your health and spirit (naturally), giving you an advantage when you resurrected. It wasn't really fun, but better than quickloading.
One of Ben Schneider's final paragraphs is a keeper:
"What is needed is the creation of a language of game drama, in the academic sense: a set of established conventions that will allow the player to read a setback as good storytelling, and not a slip-up on their part. A grammar of subtle cues to create the distinction. In any game with a story, where the player has sufficiently broad control of their character, those controls can be used to participate in story events as well as in gameplay. You can run madly to help a friend on the battlefield, only to arrive at the moment of his death; or desperately pull your troops back the moment you recognize that it’s an ambush; or struggle vainly as magical tendrils grab at your ankles and wrists, resulting in inevitable capture. In each of these examples, the player could read the failure as a skill failure or a dramatic twist. How are they going to know the difference? The answer is a coherent, consistent grammar. The cues are ours to invent. A change in camera angle. A musical motif and a change in the lighting. Excessive motion blur. Voice over, either inside the character’s head or from a narrator. A gentle slowing down of time."