Continuing the somewhat derailing conversation from this thread
My opinion is that the GM should be wary of killing off player’s Relationships. Obviously, sometimes a Relationship stumbles into harm’s way, and when that happens, death is sometimes an appropriate cost for failure. What I feel strongly about is when the GM should put a player’s Relationship in harm’s way. I believe that there are really only three acceptable circumstances for this:
[li] The Relationship (or something having to do with the Relationship, e.g., “the innocent”, “my family”, “the citizens of Townsville”, etc.) is written into the player’s beliefs. By writing the Relationship into their beliefs, a player is telling the GM to use the Relationship as a way to challenge that belief. This doesn’t always have to incorporate putting the Relationship in danger, but that often works.
[/li][li] The players’ failure would reasonably put the Relationship in harm’s way. Preferrably, this should be based on established facts about the relationship. The example I used in the linked thread was the players failing in their roll to incite a riot: the mob of townsfolk becomes out of control and this places a Relationship (who has been established to live in/be in the town – no dragging in someone’s estranged uncle from the other side of the kingdom just because a player failed a test) in harm’s way.
[/li][li] The Relationship is physically acompanying the players. Any time the players are in harm’s way, so is the Relationship.
It’s unfair to the players if the GM endangers their relationships without the players’ consent (i.e., agreeing to a roll when you know the cost of failure, writing a belief about them, bringing them with you).