Sunday, December 18, 2016

Total Party Kills, Death Saves and Character Mortality

I was recently exchanging comments with a Dungeon of Signs reader regarding the use of dragons and other extremely scary high-level creatures in low level adventures.  I'm for it, but my interlocutor made the point (a decent one really) that for new players character death, especially a total party kill is a really discouraging thing and might discourage someone from tabletop roleplaying.  This got me thinking about "Total Party Kills", and I realized that I've never run one, and never experienced one as a player, at least since returning to games as an adult. The very next time I ran a game though there was a furious animated furniture based massacre of the entire party, and everyone felt pretty good about it.
Edger Samuel Paxson - Custer's Last Stand (1899ish)
This general lack of 'TPKs' may seem shocking, especially as I run and play games using older systems, or retro-clones that have a reputation for being exceptionally deadly to characters (Like LOTFP - which actually has lenient death rules).  It's also not to say that characters haven't died with frequency and finality in games I've run and played with - even beloved characters, nor to say that this isn't upsetting.  I was quite aggrieved when my Hill Canton's character, the Eldish renegade "Tizzird" died from the single blow of a strange fractal demon thing.  His replacement "Killer" Ponzi the mob enforcer is somewhat less enjoyable.  Yet despite frequent character deaths, things never seem to come to the dread Total Party Kill, and I think this has a lot more to do with play-style and GM attitude then ruleset.  

In the past I've suggested that the death of characters and even entire parties is positive to the tabletop roleplaying experience - and I stand by the idea that character development is more fun when it happens through play cooperatively with setting development, and that the death of characters is part of this development.  This may seem contrary to a lot of players and GM's experience, and I've seen plenty of discussions about how character death is campaign destroying.  So the question becomes, why one would want a game where character death is a regular risk and how to do so without having problems or spoiling the fun of the game.

I think the why is very clear.  Games are more fun when there is a way to lose or when there are setbacks, and for a story of fantasy adventure character death is a clear loss condition.  It is also a way of signalling a loss that has little impact on the other players in a group.  If loss results in a negative effect to a character (turned to stone, sucked into a dimension of punishment) reversing this effect or rescuing that character almost always becomes the other players' goal, unless the player whose character has been negatively effected insists otherwise.  This can be a fun element of a tabletop game, but it should be rare and not the dominant result of in-game failure, because it prevents the players from completing or working on their own goals and plans.

The Saint of Killers is a vary boring Player Character
One of the key things to legitimizing character death is to make sure that everyone at the table thinks it's fair, and recognizes that it's the result of informed player decision or risk taking, rather then GM fiat, trickery or malice.  This means that where there is high character mortality, non-adversarial play is even more important then ever.

What do I mean by "non-adversarial play"?  It's pretty simple, playing tabletop role-playing games as a cooperative game or story telling venture between everyone at the table, but especially between the GM and the players.  The GM is not trying to 'get' or trick the players into losing, but rather attempting to create and environment/setting for the players to explore and adventure in.  The best way to do this while including terrible monsters, deadly traps and  this is to be a magnanimous GM and remember that the players are operating only based on the information you as the GM are giving them.

First it's good to assume that when a player has their character do something seemingly self-destructive (say leap into a very deep well) that they misunderstood your description.  Always double check and even re-describe the danger.  Something like "The well vanishes down into the blackness, it looks like a really long drop - do you really want to jump down it?"  More often then not the player will reveal that they misunderstood your description (i.e. they thought the well was a five foot drop to water or something) and change their plans.  If the character still does the seemingly suicidal activity, then it's not as if anyone at the table can reasonably feel the GM tricked a player into killing their character when the inevitable happens. That's being fair to one's players and it goes as far as giving the players reasonable clues about traps and how dangerous monsters appear, even when they don't specifically ask.  Remember that as the GM you control the entire subjective experience of the characters.  If you say "There's this big lizard in the room, and it looks cranky" the player may envision an iguana, while the GM knows this thing is closer to Godzilla.  Use description, and if you're terrible at that, even provide a clear statement that the enemy is dangerous. Remember the character certainly knows a 30' long lizard can likely swallow them whole, even if the player hasn't been given enough information to recognize that fact.

Adversarial, "killer GMs" have another terrible habit beyond withholding information about dangers, they also demand that the players provide 'perfect' responses to dangers.  The classic example of this is a swift moving underground stream. A killer GM will make the players roll a series of saves to avoid being swept away and drowned unless the players carefully describe their efforts to cross the stream - doffing armor, setting up ropes, using poles to prob the stream bed etc.  Now a non-killer GM might make the players make the same rolls to avoid being swept away, but only after letting them know about the risks, because their characters, being competent dungeoneering types, would spot that the swift moving water was dangerous to enter without removing their armor, or taking some other precautions.  GMs should assume character competence. Does a beast look diseased?  The characters will notice the acidic puss dripping from its jaws, before they have to save v. death when it bites them.  Does the rope bridge look ancient and frayed?  The characters will notice that it might be unsafe, before plunging into the depths. Is the narrow bridge above the lava made of slick obsidan, characters will notice and cross slowly and carefully - without the player telling the GM. Again, by giving both players and characters the benefit of the doubt, and accepting that errors in player observation are most likely errors in GM description (boxed text makes this worse as often neither players nor GM pay attention to it).

This same tactic works for traps and similar engines of character destruction - the key is being consistent.  Whenever a player says "I open the door" for example, I always confirm - "You reach out and grab the handle to pull it open?"  Usually the player adds something else, usually about using a 10' pole with a hook on it.  That means they survive some simple door traps, but are far less likely to complain when the door opens to reveal a howling vortex into the depths of space because as a GM I have been playing fair with them (howling space vortexes make for very chilly doors if you ever need to check for them).

Both of these ideas, confirming if players want to commit to potentially dangerous actions (or any serious action really) by making sure they have all appropriate information, and assuming character competence, go a long way toward making sure that player choice leads to character death not confusion, GM vindictiveness or bad GM description.  Yes, as a GM you can present situations where almost all choices are deadly, and this isn't adversarial - as long as you provide hints, clues and signs of danger, and don't prey on player's failure to explain simple precautions as the mechanisms that lead to character death.

More then making character death or a total party kill feel like part of the game rather then the end of the game, one needs to treat them as such.  This is one of the dangers I see with narrative based campaigns.  If the heroes of the story die it is very hard for the story to continue.  Yes, sandbox players develop plans, discover world-wide NPC schemes and build backstory - but they do so without the baked in expectation that these elements are the focus of the campaign, and because the setting is designed around setting, not narrative, changes, including character death, are much easier to incorporate.  This isn't an attack on narrative play, or it's not intended as such, and it's not something novel - Dragon's of Despair famously requires the GM to keep canonical characters alive to keep the adventure path's story moving forward.  It's just that in a narrative based adventure structure (where adventure moves through a branching set of story options) character death is extremely undesirable, and forces unanticipated, drastic rewrites of the core game structure - the narrative.  This seems especially true in D&D based games where the mechanics aren't really built around narrative progression as much as exploration.  It may be much easier to work out in different systems that are built for narrative play.

Location based (or sandbox) adventure may sometimes lack novelistic story arcs and character development (sometimes it happens - but I'd be the first to admit that it's somewhat rare), but the death of a character, or the party doesn't derail the campaign, new characters can be created, and a party reconstituted easily enough based on the character of the player who had a cold the last session and a few henchmen left in town (now promoted to characters).  The narrative may change, the party may elect to follow different goals, but the world, because it has factions and events independent of the player characters can go on with minimal interruption and provide exploration, adventure and stories for a new party, while the old one becomes another 'lost' expedition (and potentially an adventure hook assuming they had valuables or important information on their persons).

What Murderhobo doesn't aspire
to Fantasy Jesus Malverdehood?
The pang of a Total Party Kill can be mitigated even more, by folding it in to the setting, players generally enjoy seeing there characters lives (and here deaths) reflected in the setting, and its easy enough to think of ways that a dead party of adventurers might create change - an especially pious or helpful priest might become a folk saint, with a small chapel appearing a few months after his death and peasants ascribing miracles to the dead adventurer's intervention.  Warriors and other characters who leave behind henchmen or followers may have inadvertently founded a mercenary company or local militia that views the dead PC as a guide for their ethos and/or tactics.  Wealth has a life of its own as well, and a town where the party kept their plunder is likely to experience growth when it works up the nerve to confiscate it.  In general these sorts of small changes and 'Easter eggs' act as a memorial for dead characters and have the useful effect of building the setting up with less GM effort then normal.

As I mentioned above in an aside, Lamentations of the Flame Princess (a decent B/X retro-clone) has a reputation for being extremely deadly to characters.  Mechanically this isn't really true, at least not compared to the Moldvay Basic/Expert D&D rules that are its source.  In LOTFP characters don't die until they reach some number of negative HP, they bleed out and provide options to save them to fellow party members (a good mechanic actually, ratcheting up decision making and risk.  What gives LOTFP it's reputation for character death is the world that it creates through player expectations and adventure design. 

LOTFP adventures like Death Frost Doom and Tower of the Stargazer tend to have dangerous traps and tricks that can put characters in very dire situations, while other adventures tend to use monsters that are very dangerous.  Played on its mechanics alone however the system isn't particularly unforgiving, which is perhaps good given its ethos. However, mechanics do play a roll in how often characters will die, and the mechanics that come with early editions of D&D are especially unforgiving.  Drop to 0 HP or less and the PC is dead.  Combined with instant death saving throws this makes for extremely sudden and deadly game.  Personally I don't use rules that are quite as brutal, I don't want any 'Blackleaf' situations, but more importantly I have no reason to kill PCs as much as possible.  Still, there's nothing wrong with the early D&D death rules though, they just mean that character survival is always precarious and to make the setting and mechanics harmonize the GM needs to be aware of this.  It's absurd to me that one would even contemplate a railroad adventure with these sorts of instant death rules, but TSR did with the Dragonlance modules and the tortured efforts of the adventures to mitigate AD&D's mechanics are one reason they have a bad reputation now.  The mechanics that run counter to the setting and player expectations cause difficulties, but mechanics that harmonize with setting and player expectations aid the GM in running the sort of game that feels right for the setting.

Death rules are one example, and I think an important one, where there are many options for setting building through mechanical design.  If one wants to avoid character death at all costs, or to make character death something that only happens when a player chooses for their character to die for the benefit of the plot, this is an option - but one that is unlikely to appeal very much in a game where combat is one of the primary sources of risk for the players.  I don't tend to play that sort of game, but some other loss condition should be included.   From what I can gather below are various mechanical means of handling death.  Some GM's combine multiple methods for more complexity or to give the level of risk and penalty they want for their game:

A) Death at 0 HP - Characters and monsters are the same, when they fall to or below zero hit points they die.  This is the original way of handling death and it is simple, but offers no reprieve, making it especially final and potentially unexpected in games with random or 'swingy' combat and injury mechanics.

B) Death at -X HP - A common way of handling character death is to provide a buffer of a certain number of negative hit points before death occurs.  Minus ten seem common for some reason, which provides quite a buffer.  Under this system PCs at 0 HP are simply unconscious, and often PCs with negative HP will continue to lose hit points over time and slide towards death if they aren't healed or aided after sliding into negative HP.  This system can be as deadly as the GM wants it to be, death at -2 HP is likely (especially in a game where damage totals are high like 5E), while death only when the character reaches negative their full HP value is very lenient, especially as the character levels.  More over this sort of system depends a lot on the availability of healing.  Where magical healing is common (or available at all), anything that allows it to be applied before death will make for much more forgiving death rules.

C) Death Save - When the character meets certain criteria (0 or negative HP - but it could be certain special attacks as well) the player makes a saving throw, and on failure the character is dead.  Often in older D&D based systems this is a "poison or death ray" save, but saves against constitution are common in newer editions.  Fifth Edition allows for multiple death saves, combining the concept of Death at negative X with the save concept and making it hard for characters to die (unless foes keep attacking fallen characters).

D) Death and Dismemberment Table - Either as a kind of 'death save' or as an addition to the death save depending on what's on the  table involved.  A death or dismemberment table contains all sorts of injuries that inflict either gruesome deaths or permanent injuries to characters.  There's something fun about these tables in that they add permanent injuries and scars to characters making survival of a near death experience a character building experience. They also add complexity (though not much) and I find them to be either too punitive or too lenient. Death and Dismemberment tables are often very "swingy" themselves with the possibility of a minor concussion warring against that of suffering an internal injury that deletes half of the character's permanent HP.  Like a death save they have the advantage of providing a separate mechanic that can be brought in by GMs for special circumstance or attacks that are terrifyingly dangerous and they can be easily combined with other mechanics.

My H.M.S Apollyon rules use (B) and (C) mechanics above. At 0 Hp a character is considered unconscious, and at negative hit points a character must roll a death save.  Unlike many versions of this save, it's a separate save number, starting at 10 or above on a d20 that increase by 1 point each time a character makes the save. This is a fairly lenient death mechanic that makes lower level characters more likely to survive then high, but isn't complex and it is hard to meta-game by boosting CON or some other stat (though the berserker subclass is almost entirely built around surviving death saves).  If a character survives a death save they are at '0' HP for the rest of the session, but may act normally with the understanding that any future injury will cause instant death without a save.

None of these methods really matter in a total party kill as there will be no characters left to rouse their comrades and it might be assumed that most monsters eat their victims after finishing off the injured.  Not all foes though will massacre a downed enemy, so there is a strong possibility that with a more lenient death mechanic some party members will live to be captured (by intelligent foes), or crawl away later (after losing to some creature that simply ignores downed enemies - something like a golem or some undead) and it seems important for the GM to think about these things when faced with a possible total party kill, as the narrow survival of some characters through luck or the oddity of their enemies makes for a plot defining event, just as a Total Party kill does.


  1. I've also never had a TPK, but I've always seen it as an enormously creative seed and not really to be avoided. If your party dies combating the villain, his plan succeeds and you're adventuring in a new setting based on the aftermath.

    Your previous party could be folk heroes and an inspiration as the last people to defy the tyrannical evil which now controls the nation, and I'd definitely include some kind of quest to recover their items, have beloved NPCs make cameos etc.

    My current campaign is set twenty years into the future (and a substantial geographical distance) from the campaign I'm playing in, and the players love any reference to their other characters' adventures.

  2. The two factors you identify as determining whether character death feels "fair" or "appropriate" are spot-on. I especially haven't seen much discussion as to plot type (narrative vs. sandbox) in determining whether character death adds to or subtracts from player enjoyment. I usually run games between the two poles, but am currently running my first full sandbox game - fifteen sessions in and no one has died, due more to careful players than lenient death mechanics (I'm using a death save mechanic, with some dismemberment-like consequences on survival)! I'm not sure whether to be proud of my players or disappointed in the obstacles I've created - probably both, haha.
    Your berserker subclass sounds interesting - do you have it up on here anywhere?

    1. A) No the berserker sub-class is not up anywhere.

      B) I'm not really trying to suggest that a scene based campaign will make death feel more 'fair' as much as character death presents more problems for adventure design and continuity in a scene based campaign.

  3. The only TPK I can remember GMing happened when half the group insisted on heading into an encounter that I pretty clearly signposted was beyond their ability. I’m confident I clearly signposted that because the other half of the group absolutely refused to participate in the encounter. The half who went in got slaughtered. Then the half who realized they had no business going in decided to “avenge” their fallen comrades, and got slaughtered. As a GM it sucked.

  4. I ran 5e for my inlaws (the first time my wife's parents had played d&d or any rpg) and got a TPK in Mines of Phandelver (I think it's called).

    1. I almost got a TPK running Place of the Silver Princess in 5E some years ago.

      It's absolutely possible, especially at low level - but reading through 5E forums and WotC adventures (especially newer ones) it seems very much against the popular play style?

      Also always nice to see you around Trey - and to hear you are running games! I still want to see more Lands of Azuth.