Player character death is usually a bad thing. That’s why in story- or character-driven games I run the following rule is in effect: Player characters will not die unless the player explicitly risks the character’s life.
By character death I mean any event that makes a character unplayable for extended period of time (at least two sessions) and is not easily reversible.
Character death is often harmful to the game, because
- Character generation takes time, especially in heavier systems.
- It often stops the game, either permanently or temporarily.
- Players don’t like losing precious characters.
- GM probably had hooks, plot opportunities, bangs, and so forth planned for that character. They all need reworking or become totally useless.
Random character death is even worse. In addition to the above, it is usually anticlimatic.
My solution: Justification
When players must risk the characters specifically for there to be a chance of PC death, they are never (by definition) random. In many situations there won’t be a crushing anticlimax (maybe a mild one). And the player can only blame themselves for the loss.
My solution: Implementation
So, how do players risk their characters and why would they ever do it? Usually the risk is clearly defined; for example, in most of my games combat is a rare and serious affair, and hence entering it always means a risk of PC death. Escalating some conflict may mean it: The noble told you to be gone and stop bothering her daughter. Then he had you thrown out of the ball and declared an outlaw. If you still persist and try to sneak to meet her, failure means the guards will slay you first and not ask any questions after that. If you negotiate well, they might let you meet the noble, and if you flatter well, you might get away with your life. Fail and you will be executed. I think this avoids the problem of random death: The death, if it happens, is clearly significant because it shows that the character really cares for the noble’s daughter and is willing to risk death in the pursuit. Also: the rule of three may or may not be involved in making this seem not arbitrary. Something worth thinking about.
The (slightly more) tricky part is in making the game interesting with reduced chances of PC death. Something must still be threatened and the PCs must still be capable of doing something for it. Personally I have discovered two methods of makings PCs and players care about stuff. First is in asking the player what the character cares deeply about (the rules of the game may or may not help in this; if they do, they are called flags; another fine opportunity to shill the Burning Wheel which has such a feature). Loves, hates, is dutybound to, is tied to with ties of blood. Second is forcing such a relation, either out-of-game (make characters who hate this bastard) which can seem too forced, or through the game (play NPC so as to inspire fear, love, friendship, loyalty,…), which is far from reliable and can misfire.
Threatening the stuff that is important to the PC and the player can make some players turtle. Hence it is important to make it clear that it makes the game better and that you won’t arbitrarily destroy such things without the PC having a chance to avert the destruction (but maybe sacrificing something else in the process, like her life) or also give equal opportunities to gain ore good stuff (which can later be threatened). This is all good stuff for plain old dilemmas. A shadowy figure offers you great power and wealth and magical knowledge if you betray your friend. Allows the character to get something for losing something else. Few players will start turtling if they get to make relevant decisions.
My solution: Additional benefits
I won’t have to fudge rolls to prevent pointless character death. Players who actually are okay with their chars biting the dust at dramatic moments get to do that, which makes martyrism possible. Players get more control of the game. From my point of view, everyone wins.