How To Kill Your Party

I ran Candlewick last Saturday. I think it’s the closest I ever came to killing one or more PCs.

I felt deeply conflicted. I don’t have a problem with PCs dying through bad luck or incompetence. In Candlewick I don’t have a problem because it’s supposed to be a brutal, tragic setting – even if the PCs are children. In fact, especially because they’re children.

On the other hand I did have an NPC practically drag them into the place of doom. Normally the PCs could run away, or at least have their eldrich horror friends come to their rescue. In this case their options were severely limited because of where they were. They were up against something that was bigger and rolled more dice than them.

I Cast Detect Plot

I’m guilty of making players make rolls to find stuff out, assuming they will be successful. What if they’re not successful? Do they just blindly wander by the plot? I’ve played in those games, and I’ve run them. I prefer narrativism that thrives on players getting the right information at the right time, without having to roll dice.

But consider party death. It’s very hard to make PC death narrativistic without being arbitrary or contrived. Most deaths should be tragic and avoidable, the result of bad luck or judgement – not the GM suddenly declaring “rocks fall, everybody dies”.

To avoid death being arbitrary, you actually need a healthy does of simulationism and gamism. The former because you need an objective view that the most likely outcome is a death. The latter because you need to give your players a chance to cheat death. If they succeed or fail, Drama will follow in the aftermath.

Why Don’t You Just Die?

Killing PCs is not just rare, it’s pretty hard to do in a way that’s balanced. Consider the examples on Wikipedia’s TPK page – it notes CoC and Paranoia as high party attrition games as a consequence of the genre, but in these cases anything that can kill a PC probably will because it is overwhelmingly strong.

Balanced simulation/gamist systems can be either abstract, needlessly complex, or just very tiresome to play through. D&D, Exalted and others suffer from exactly this problem. I firmly believe that the harder and more complex a game is, the less it feels like a dramatic combat.

I was actually quite pleased with the way MaoCT handled combat and harm, and it’s partly because damage is in low single-digits applied directly to the player’s stats. I’m not sure how the players felt, but they certainly got the message. If they’d have been crossing off boxes on a wound track (Exalted, World of Darkness) or hit points, would the consequences have been as immediate?

Time and Place

There’s nothing wrong with killing PCs, but it has to mean something. That means the players have to be able to visualise where their PCs are, and the circumstances surrounding their death.

In one of the best fantasy games I’ve ever played I lost my first character in the 3rd session. At the time we were escaping some foreign troops. I hadn’t realised where my character was – there was a lot going on – and all of a sudden, I’d been shot with a crossbow (through the spleen, as I remember).

I rationalised it as the character being a monk, not a warrior, and just being unlucky with being suddenly confronted by troops. But if the GM had been clearer about the lay of the land it would have seemed fair even if there was no way to win.

It’s not necessary to pull punches but it is necessary to give players a sense of how deep they’re in, even if the characters aren’t.

Final Words


p>Some games (CoC, Paranoia) carry the expectation the players will die; they make this easy by making the odds of dying transparent and high. Other games give the expectation that no-one will die; they often do this by having no really clear mechanic for injury and loss, as well as abstractive mechanisms. OK, you’re going to die if you run out of hit points, but the attrition rate is so slow that it drags the combat out and robs the scene of drama.

Very few games really deal with recognising and appreciating risk. Crucially there’s a step that’s overlooked when players are confronted with a risk – a contract between player and GM that they understand they are risking something, and will pay the price if they fail. That may seem implicit in all games, but the way we play a lot of games it really isn’t.

On the other hand, in a game like Candlewick where the characters are children, missing that step out is valid when the character’s innocence hasn’t taught them to be wary yet. In fact, these experiences are the hard knocks that will teach them to be more careful in future – if they survive.

  • Nice article!

    I’m not sure I agree, though, with “To avoid death being arbitrary, you actually need a healthy does of simulationism and gamism”. Consider Dogs in the Vineyard, where death can only come as a result of (a) agreeing to stake your life on a conflict, or (b) deciding to take risky fallout dice rather than yield the current conflict to your opponent. There’s no objective modelling in there, nor any tactical decision – it’s a question for the player: would you be happy for this to be your character’s death scene? If not, don’t risk it.

    Another great example is Trollbabe, where the GM literally *cannot* kill the PC – the only way the PC can die is if their player (runs out of increasingly-dangerous re-rolls and) decides they are not happy with the outcome of the conflict and would rather narrate their character’s death scene. Again, there’s no simulationist risk-modelling to that mechanic, nor a gamist tactical decision. It’s simply up to the player to decide what’s dramatically appropriate for their character.

    In both cases the death cannot possibly be arbitrary because the player can avoid it if they choose to – not through good strategic judgement or management of risk, but through choosing other dramatic outcomes instead.

    – Nick

    • Both of your examples are fair – I should have said “you need simulationism, gamism, or player consent.” But the third case is a red herring. I’m talking about a case where the decision rests solely with the GM to kill the character, as opposed to the GM asking the player whether their character should die at this point. In that case I expect the character to be retired against the player’s wishes.

      In fact, although I really like a narrativist style I’m not keen on the players coming to terms with their PCs’ death before it happens, which is pretty much what happens in the games you describe.

      Also, with the DitV example there is a case of gaming there – either the character stakes their life before the dice are rolled, or (as you say) they take fallout dice which are still a random element (IIRC, have only played once).

    • Is it not the case that if you yield to an opponent during a gunfight in DitV that means they kill you if they want to? That’s the way the group I was in were playing it and that’s definitely not with player consent.

      Indeed, DitV combats are incredibly brutal as a result. I’d compare the system to playing WFRP with zero fate points left in terms of how suddenly and arbitrarily you can die. (The deviation on the dice pools is massive, so it’s basically impossible to be sufficiently good at fighting that it’s not insanely dangerous.)

    • Gunfighting follows escalation of other combats. They can be brutal but something must have happened to raise the stakes so high in the first place – so I think Blackrat’s assessment holds up.

      On the other hand it seems you don’t think DitV makes a good strategic combat game, and I agree. But the point of DitV is not to game but to know how the direction of the conflict is going, which it does pretty well – forcing the life-or-death decisions on the players.