The Miserycrawl, Negadungeons and Killer GMing.

THE COMPLAINT

I've occasionally seen some table top gamers and bloggers denigrate "miserycrawls" as an OSR aesthetic, usually this is largely the typical litany of boring and traditional complaints and stereotypes about the OSR as a pack of neckbearded haters of innovation who relish in the unfair slaughter the player characters and the gruesome aesthetics of heavy metal album art.  As dull as this critque is, it did get me thinking about setting feel, adventure design and why I like both low fantasy settings and high lethality games.

I advocate for a low fantasy settings when using a Dungeons & Dragons based system, especially the early editions, where simple, quick combat that tends to leave even strong characters dead sometimes because these settings work with these mechanics.  When one is using the lethal saving throws, low hit point totals and the high probabilities to hit that one finds in white box style D&D (negative AC is an artifact of later editions and it can ruin mid-level play) letting players know that they are not invincible, or even the toughest people in the local area, tends to be important - hence the low fantasy, or grim fantasy setting.  The feedback between mechanics and setting are important for letting players understand how to 'play' in the world and help build player comprehension of both ruleset and setting expectation. Low fantasy settings (though whimsical 'gonzo' settings also work very well as failure becomes part of a narrative of slapstick black-comedy) become especially important now, as they reinforce the rules and set player expectations that character death is more likely then in more modern systems, or games with player facing narrative.

Retreat from Moscow - the Scenario (Painting by Adolphe Yvon - 1856)

ACROSS A WASTE OF RAZORS - THE MISERYCRAWL & THE NEGADUNGEON

The specific complaint about "miserycrawls" suggests a sort of game where the sandbox (really the setting) is about barest survival, and the characters are disadvantaged and ill-equipped against the wilderness and the horrors within.  While Deep Carbon Observatory has been specifically cited as such an adventure, I'd suggest that it is very different - a disaster scenario, much like Dragonlance's Dragon's of Flame, but with the key difference of not presenting a a strict railroad that ignores the suffering of 'commoner' NPCs and one where the writing is both evocative and filled with strong imagery.  Carcosa, Black Sun Death Crawl and DCC's Perils of the Purple Planet may fit the bill in that they are all settings where the survival aspect is emphasized, and character mortality is expected to be high.

To some extent this idea of the high mortality survival sandbox is deeply rooted in D&D's past, though it's notable that rather then the Fantasy Wildlands sort of setting (including it's dinosaur filled swamps) the products listed above take an explicitly post apocalyptic sort of approach (and even a contemporary Mad Max aesthetic, with a strong dose of the weird).  Now setting design is a subject open to debate, but to me steering clear of vanilla fantasy settings with strong Tolkien roots is important as it helps circumvent player expectations that the game will resemble a fantasy computer role playing game and allows both player and GM imagination to take the fore rather then encouraging the players to think like some stereotype from Fellowship of the Ring.  Similarly the DIY D&D/OSR tendency to reskin or redesign all monsters to make them unique is an effort to break free of the accumulated player knowledge about monster ability, so that monsters can be wondrous and frightful rather then an exercise in players' ability to memorize old monster manuals.

In other words I'm not sure what the distinction between a non-vanilla survival based sandbox game and a 'miserycrawl' is.  Perhaps the critique of the miserycrawl is a larger critique of dark fantasy settings - 'crapsack worlds' or the 'dung-ages'.  As someone who has run a long crapsack world game, I can see that an endless litany of depressing situations might be irksome, especially as any game setting that asks for complete player seriousness is dysfunctional.  All tabletop games lead to a fair amount of joking and irreverence towards the game world or become bathos drenched impossibilities. If one tries to focus on a world of downtrodden peasantry, depression and endless mudflats with an eye toward encouraging players sadness, hopelessness and despondence ... one will fail, and fail in an ugly manner because the dumb jokes and friendly banter of people sitting around and playing a game will always overcome a GMs efforts at serious pathos and convert it into bathos.  Yet the crapsack world can itself become the 'straight-man' for the humor of the game, especially when there are gonzo (i.e. funny because they are referential or absurd from a player perspective, but not a character one) or weird (because strange, hopefully evocative imagery tends to overcome mock seriousness) elements.  The basic complaint then seems to be either one of tone or one of character mortality, but these critiques overlook both the inherent silliness of tabletop games, the importance of system and setting feedback, and the emphasis on moral play and player freedom that such settings proivide

A couple of years ago there was a similar brief interest in the discussion of the 'NegaDungeon' that had similar undertones at times.  The NegaDungeon, exemplified by LOTFP's "Death Frost Doom" and defined as a dungeon that plays on player hubris and genre conventions to offer a negative result (world changing in Death Frost Doom's case) for player action, is an interesting idea, and one that was subject to the same sort of critique as wallowing in player misery, defeat and trying to derive fun only from the sadistic destruction of the characters.  I don't think the NegaDungeon is this really, rather it's a way to try to tell a horror story in a tabletop game, rather then and adventure story, and the player fun/success is derived from escape from the story or survival rather then retrieval of treasures or victory over enemies.  This seems like a viable playstyle, but perhaps a jarring one to spring mid-campaign in a traditional fantasy world and something that may work best as a one-shot or campaign starter (why are a band of skilled killers, thieves, sorcerers and religious fanatics wandering about without patrons or family - because NegaDungeon.)  Yet the complaint about NegaDungeons sometimes also includes complaints about the importance and ease of character death in these settings so it has a lot of similarity to the rhetoric around miserycrawls.

LETHALITY AND SETTING?

In an old school game, the mechanics mean you're the guys at the bottom of this painting.

The problem then that seems to remain with any critique of "miserycrawl", beyond simple genre dislike, is one of excessive or unfair lethality to player characters.  Indeed, the deadlines of older style D&D is often commented on, but this in some ways ignores an underlying question.  Obviously if one simply believes that characters should never die in games there's no way to answer this question to your satisfaction in a game where character death is possible, and if one likes games where character death only comes after an agreement to use it as a way to advance a story narrative, death due to random factors might look both pointless and unfair.

However, assuming that one is playing a game where character death is possible, how does one best make it feel both fair and not overwhelmingly terrible for players.  In a system where character death is known to be a possibility or likely result of going on dangerous adventures presenting the world as dangerous feeling, and so potentially as a 'miserycrawl', is actually helpful.  The idea that dying to everyday horrible like banditry, bar fights, ergot madness, food poisoning and disease is commonplace suggests that dying to poisoned spikes and the rusty weapons of the unquiet dead is at least somewhat glamorous and perhaps a better fate then what awaits characters who decide to stay at home.  Yet this isn't itself the big issue - most players playing a system with high lethality are doing so because they feel that character death is a reasonable penalty for failure. What makes even these sorts of players cranky is a character death that seems unavoidable and worse when it seems to provide the GM with glee. It's an easy matter to kill characters as a GM - just force them to repeatedly fight things far more powerful then the characters, at a tactical disadvantage and/or scatter deadly traps all over the place. This is adversarial, bad GMing.

Assuming a game with an open/sandbox world, lacking challenge levels or similar tailored difficulty, a system that provides for dangerous and potentially lethal combat and a save or die mechanic for other perils it is not hard to wipe out a party of characters.  Often this is the result of bad GMing or an adversarial approach to the game.  In a system where death is easy, the GMing can't be a partisan for it, and can't be adversarial towards the players' efforts.  Power is stacked with the GM, and it's especially important to be a fair even forgiving GM, because while players may realize that their characters can and will die, most find it 'unfair' if this seems like the result of pique or a GM on a power trip trying to show off that he can outhink the players. Character death should be the logical outcome of bad player decision making or bad luck, its possibility lightly telegraphed and certainly avoidable in hindsight.

HOW NOT TO BE A KILLER GM

Assuming that one isn't simply an awful person who likes playing petty tyrant to one's friends at the gaming table (and really if that's the case nothing I say is going to be good advice), no one wants to be a killer GM.  Now there's always a bit of back and forth between player's and GMs about rules and the activities of their characters, especially when something bad happens, and the GM should hold to their positions in these kind of discussions - not letting players have everything they want all the time is good GMing, even if what they want is sometimes cool.

Killer GMs though aren't just trying to maintain the fictional integrity of the world, but actively and antagonistically attempting to thwart player actions and plans.  I suspect there's a couple of reasons for this, from a spiteful desire for power and control to the need to tell a specific story creating an impulse to punish players who depart from that story.  This second set of behaviors was actually encouraged by some editions of the game (or at least some product lines - like Dragonlance), and still is popular in some sorts of GM advice.  Beyond any distaste for eliminating player choice, even the impulse to gently correct players back into the story can be very destructive in a high lethality game with random mechanics.  For example, in a game where the party encounters a snide but powerful NPC (say that classic slightly mad wizard) who wants to give the players a mission, but the players become annoyed and their characters attack.  In a sandbox game the wizard will make his demand, but the players will know that there's a good chance this NPC (or really faction) can wipe them out, and that there's no story dictating a next act.  They will weigh and scheme and wheedle, but are very unlikley to attack.  In a railroaded game the players have likely become convinced that there are 'combat encounters' and 'story cutscenes' and chaffing from this limitation, annoyed by the story pushing them to work for an arrogant quest giver they will likely attack.  Pushing the game off the GM's invisible rails is a fun time for some players, and leads to disruptive play because the player thinks that every event is already set. A GM that suddenly changes style, usually gently rerouting player actions that depart from the story, but occasionally making the logical and horrible consequences of player decisions to depart from the GM's plans will be seen as an unfair killer GM, because that GM has made his expectations of playstyle vague.

In a less deadly system, or one with narrative protections this sort of misunderstanding is less likely to occur, and when it does happen it's likely to be less final, because defeat and character death aren't the same.  Bad in character decisions driven by frustration at the lack of player choice or player inattention (really why pay attention to the GM's monologue if any response will always lead to the same place) is mitigated by a forgiving system. There's no mitigation in most older systems of D&D, and one of the defining features of a sandbox setting is that it's very easy to get into situations or in conflict with enemies that are far too powerful for the characters to overcome. So the key element to avoiding players feeling that the death of their characters is to avoid misunderstandings about what's happening in the game world, without completely warning off the player from doing certain things.  Additionally it's to make it so that when characters do fail and die these failures feel like the logical result of player decision and risk taking.

The setting and mechanics can both set expectations, and so help a GM and players avoid misunderstandings about risk and reward.  For example, the combat mechanics of Basic D&D are pretty unforgiving, and transparent about it, which should give players the idea that combat is dangerous and likely to result in character death. This is fine, even desirable for an exploration game where player freedom of action allow many alternatives to combat, but in a setting where the characters move from fight to fight without opportunity to avoid a die rolling frenzy of death or set up other factors in their favor, these deadly rules quickly become a hindrance as character after character falls to probability.  Of course players coming into such a game with expectations built around different more combat friendly mechanics (say 4E or a "World" game) won't know that fighting is a dangerous and poor decision unless they really read and comprehend the rules and they may end up fighting every encounter, with characters dying a lot and the player thinking "This OSR thing seems to be punishing, filled with boring simple combat, and attractive to GMs that love killing characters." 

Setting can help.  The descriptive power of a good GM can really get across a lot of clues about setting expectations, and when one wants to emphasize that combat isn't always the best option (even if it's clearly stated in the rules) a miserycrawl or low fantasy world of horrible injuries and high mortality provides a good way to do so.  It is also a reason that a lot of OSR/DIY GMs and players like evocative and weird settings - the cliches of vanilla fantasy have become the cliches of fantasy video games, where avatar death is increasingly uncommon and where players already come to ever encounter or location with a set of pre-existing world knowledge (perhaps goblins are weak enemies) which will serve the player poorly in a system where even weak enemies are potentially dangerous, and where bullying them into becoming allies is likely a better plan then fighting them.

The alternative to building a setting that emphasizes or provides clues and justification to or for the unforgiving nature of the rules is to focus on the inability and incompetence of the characters.  This works okay in extremely humorous settings, where players can be amused more easily by their character's failings - Paranoia seems like a classic example.   This technique is less helpful when one wants to run something that isn't gonzo and is closer to classic fantasy with more potential for heroic characters, as when one makes the characters seem fragile and weak it tends to detract from the sense that they might do anything in the open game world.  The other answer is to make the setting feel like a dangerous place, a match for even the toughest most competent of characters.  This second answer works best to encourage player freedom of choice as well - because the player has some belief in thier character's abilities. 

CHARACTER COMPETENCE AND THE MISERYCRAWL

The idea of character competence at the heart of not being a killer GM is that the characters are competent at what they do.  The first level fighter is a 'veteran' - a warrior and slayer of men on distant dusty battlefields, hard to frighten, adept with tactical stratagem, knowledgeable about ambush and survival. The first level thief knows every dire mechanism that the merchant houses use to protect their wealth, has a fair grasp of standard poisons with the danger sense and ferocity of a startled alley cat. Magic-users are learned, filled with powerful secrets and observant and clerics blessed and protected by divine favor.  Yet the hazards of the underworld are far more dangerous then back alley brawls with hardened toughs, and the depths monstrous strangeness far more terrifying then breaking the final desperate shield wall of a band of sea raiders pursued back to their longship. The alternative is to view the characters as complete incompetents, weak willed, unskilled at the arts they profess and incapable of basic survival without specific player input.  I prefer the first way of viewing the game world, and this in some way leads to "miserycrawls", because the mechanics are still high lethality, the open game world offers the potential for "unbalanced" encounters and the focus on player choice means that players can make very bad decisions.

Player choice means this is what your 3rd level party is likely to be.
Why not make the world equally or more unpleasant?

To avoid running high lethality mechanics without being a killer GM, one must take the player's word that their characters are not incompetent wastrels and act accordingly.  Danger is already all around, and the mechanics punish mistakes harshly already - there is no need for the GM to use their power to interpret player action harshly. When the players describe their actions, don't look for specific actions they have overlooked which create a dangerous situation, always assume that if a common sense precaution would prevent a tragedy that the characters took it.  For example, if the players are climbing a long rope and have the time to do it, the characters know what they are doing and have cinched their belts to the rope as a means of preventing a fall, even if hit by a crossbow bolt from the dark they won't automatically fall off.  In other situation where there's a couple of risks at play and more then one logical simple answer, the GM should ask and even go as far as suggesting  the more obvious precautions. 

A classic example of this is the fast moving underground stream.  Wading across icy water in torchlight can be dangerous - especially in heavy armor, which will lead to sure drowning if the armored character is swept away. When players say they are going to cross such an obstacle, a good GM should mention that there's a special risk to armored characters, giving the option of removing their armor and assuming that if they do they properly bundle up the armor so it won't get lost.  Players have a choice here, because obviously plate mail will take a turn to put back on upon crossing the river - an unarmored turn where a random encounter would be especially deadly. It's worthwhile letting the players know the potential risks of the logical courses of action is fair as competent characters would recognize these things. When a player says "I open the door" in a world where doors are often trapped with terrible deadly things, it's a good idea to ask "So you're gonna just pull the handle open, or are you taking precautions?" and to do this consistently.  Assuming that characters are competent goes a long way towards letting your players know that you aren't out to get them, and assuring that when they do die in spectacularly horrible ways it's recognizably the fault of player mistakes or decisions and not the fault of the GM's adversarial approach.

WHY THE MISERYCRAWL WORKS

This is how I run systems with unforgiving rules and how I especially run dark fantasy or low fantasy games where the world is obviously dangerous and grim.  Older high lethality systems and grimmer settings compliment each other the setting emphasizes ruleset and playstyle, allowing the GM to run things fairly without the perception that they are trying to make things harder or more dangerous for the characters.  This is I think the positive genesis of the "miserycrawl", which if GM'd right creates a synergy between setting and rules and provides players new to the concepts of old school play with  clear clues about wise game play conventions and differences with newer systems.  However, a key to making these settings work is to make sure that the mechanics and setting work together - running Deep Carbon Observatory in 5E might be possible (and 5E is easy to turn back into a far more deadly system), but without the rules reinforcing the setting and providing notes of concern or seriousness (characters in actual peril and character deaths) the GM might be forced to push too hard on the grim aspects of the setting, creating the sort of bathic nonsense that gave rise to the miserycrawl name.