First Impressions: tremulus

Continuing my tour of roleplaying systems that started with the previous post, today I look at tremulus. You may have heard of it’s progenitor, Apocalypse World, or of its sibling, Dungeon World. tremulus is a “storytelling game of lovecraftian horror”, and is, from what I understand, a fairly straightforward hack of AW, with some bits screwed on top. The result is called “haiku”, and is said to also draw from FATE and Fiasco.

Still, much of what I’ll be saying here likely applies to AW as well. Perhaps I’ll get around to playing it on its own one day. As for my experiences with tremulus, we’ve played a 5-session game, that was initially meant to finish in 3. And it was amazing, one of the best games I’ve been a part of. But how much of that was due to the game itself? I am, as ever, dubious. Lets dig in.

Quick note first: I don’t think you can currently buy the game, as it was a result of successful kickstarter. But as the finalized pdf was recently made available, it’ll probably be up for sale soon enough.


Seemingly everyone has been praising AW/DW lately, so chances are, you already know how to play them. Feel free to skip this part in that case.

AW Engine (AWE for short) has two elements to it: a beautiful core mechanic, and a bundle of GMing principles.

Core Mechanic

The core mechanic is simple: whenever PCs try to do anything, they roll 2d6 + relevant stat. 10+ result means they’ve succeeded; 7-9 is a mixed success – they get what they were after, but something else happens as well; and a result of 6- means not only do they fail, but something happens, either as a result or reason of their failure, or even just in their general vicinity. Something always happens, there is never a “I rolled a check and failed, guess that was a waste of a roll”, so frequent in other games.

The most common actions players undertake are written up as basic moves. Things like Act Under Pressure or Poke Around. The real purpose of these is to ease the cognitive load on the GM, as they come with pre-described results, often letting player or GM choose several options from a list. As you’re struggling with a cultist, do you choose to suffer little harm, or to pry the dagger out of her hands? Each character playbook (class, basically) also has several special moves to choose from, and these serve to differentiate characters, allowing them to do something no one else can, use a different stat or providing a bonus to certain actions.

And then there are DM moves. These range from general, like “Separate them” or “Offer an opportunity, with or without a cost”, to ones specific to situation or hazards PCs face – more on those in a bit. These moves are to be used when dice indicate that something other than utter success happens – those results of 9 or less. Even more so than character moves, these serve purely to illustrate what DM’s moves are like, and to offer some inspiration at a glance.

Of course, these codified moves can have the exact opposite effect, as both players and GMs immediately get stuck looking at the short list in search of something to do, despite the game explicitly warning against this. And after the first couple of sessions I’ve basically ignored the GM moves section, going with what felt appropriate.

GMing principles

These, again, are codified, an actual list of do’s and don’ts.  “Ask provocative questions, build upon answers”, “Make the characters’ lives interesting”, “Play to find out what happens”, “Introduce the strange, the weird, and the alien at every opportunity”. They are by and large rephrasing what GMs have been always doing, but putting it in such a light as to force reevaluation, often leading to revelations. They also do a good job of focusing reader’s attention on how the game should be played, what should be emphasized, what the creators thought was important. Trying to keep them in mind as I ran the game has definitely affected it.

Another thing worth noting is the focus on PCs. GM never rolls any dice, instead describing a situation and always, always asking “what do you do?” All the moves are written with that in mind.

The last significant element in the game’s approach to GMing is the way it suggests adventures should be structured and prepared. You make up several hazards, the trouble PCs will have to deal with. They have a type, and an impulse dependent on it. Both are codified (the game really likes its lists), providing a bunch of starting points to choose from. To give you an example, one of the hazards in our game has been Hope Springs sanatorium (type: Location, impulse: Maze, to trap, to frustrate passage).

Several hazards are then combined into a plot thread, and each thread is given a Tragic End (what happens if PCs fail) and a likely progression of events. One hazard is nominated to be a linchpin, which is purely to help GM think about it, just like everything else in this methodology. Combine a couple threads, and an adventure is ready.

tremulus-specific bits

At least I think they are. Rolling a 6- on Poke Around results in GM earning a hard move, a sort of currency, to be used on the PCs at any point. It differs from regular moves in that it’s not in any way preventable. It’s a terrible thing that happens, and PCs have to react to its consequences. “A cultist stabs you, take 2 damage and you now have  a knife in your gut, what do you do?” as opposed to “a cultist goes to stab you, waving her knife menacingly, what do you do?” PCs are expected to fear these hard moves, and they did. It probably helped that I used small Cthulhu statues as tokens for them.

On a roll of natural 12, or as a result of some moves, PCs earn a point of Lore, another currency of sorts. It comes with Dark Insight, a bit of mythos trivia, and is used to power rituals, as well as special Lore moves that each playbook gets. Its a thematic little mechanic, though we’ve found some Lore moves much more useful than others. Even smaller Cthulhu statues were handed out to symbolize these.

Finally, there’s Ebon Eaves generation. Ebon Eaves is a small town, the default pseudo-setting of tremulus (more are coming as a result of kickstarter stretch goals). Players are asked to pick 3 options that most appeals to them in two questions: what they think to be real about the town, and what weirdness they’re aware of. Resulting combinations independently define Local Color and Town Lore, both of which come with an open bit for players to read, and secret info for GM to consider, which includes hazards. It’s an interesting way to quickly generate a town, though not really replicable. It’s basically a brute force approach to setting generation, where each possible combination (and there are quite a few) requires a couple paragraphs of text. Probably good for convention play, though.


Over a thousand words later, we get to the actual meat of the post. How does it all work together?

It is, in fact, narrative

There’s little to none mechanical mastery involved, no way to position oneself to get a better chance of success, no character builds to speak of. There is, however, narrative mastery. Being able to phrase the action in a way that allows to use the better stat, to frame the situation such that possible outcomes are less terrible is key. As a direct consequence of this, GM fiat is the name of the game, to the point where dice and rules become almost irrelevant, a thin facade that allows for pretense of impartiality.

How many dice rolls does it take to get out of a burning house? In D&D you might sketch out a map, figure out what checks and difficulties will be involved in breaking down locked doors and dodging collapsing roof, and use movement rules to see how PCs do it. Win conditions are set and everyone knows the criteria of success. In AWE it’s entirely down to the feel of the scene and the will of the GM. One roll or ten, horrid burns or glasses knocked off.

This is not bad as such, it just removes the safety net rules tend to provide in games. And just as there’s a lot of pressure on the GM to make it right, it’s possible for players to lose the  sense of immersion and feel cheated as their actions and their rolls keep getting thwarted. Understanding and trust between players and GM is paramount.

AWE/tremulus allows players to have great agency, but it is always and forever agency granted to them by GM.

It is also quite meta

Even though it’s shy about admitting it. This fact is most obvious when looking at the Poke Around move, which uses Luck, a meta-stat on its own. If a player decides there might be something useful in a place and Pokes Around, they may find items, clues or even a hidden passage. The act of successfully Poking Around creates those items or passages. The GM may have thought it was just a regular room of a regular house, but a player decided it could be interesting, and in doing so (and rolling well) imposed that idea upon the world. Suddenly, there is an old amulet hidden in a drawer under some papers, which probably changes the identity of the owner of the house and casts further implications.

Similar things happen with other moves as well. When looking for ways to interpret outcomes of rolls, a GM will come up with something, as is their job. Something happens, even if there wasn’t originally supposed to be anything happening there. The act of rolling those dice caused a change to the world that is most often related to what a PC was trying to accomplish, but not necessarily caused by it. Outcomes are merely correlated to dice, not caused by them. And therefore outcomes are merely correlated to stats and to characters themselves. Which means that having a low stat might cause someone to have been a cultist all along.

Sadly, the book doesn’t explore this, and I think there’s a lot of discoveries to be had here. I wouldn’t mind playing around with a better defined meta-system based on AWE. Remove character stats in favor of character tags, similarly to how objects have them, and treat them the same way. So when a Frail character fails a roll involving physical activity, it might be due to their frailty, or to something else entirely.

The core mechanic works beautifully… unless it doesn’t

Being an investigative game, some of the actions PCs often undertake don’t have immediate consequences. It’s easy for GM to imagine what might go wrong if a PC wants to climb a fence, but what if she’s just Puzzling Things Out, A move PCs can undertake any time a new piece of information becomes available to them. That’s the question I asked on RPG.SE, and tried to incorporate the suggestions I received (many of them meta) into our game. And it still felt forced at times, and I still had to remind players that “something always happens”, even during the 5th session. If by that point the fundamental philosophy of the game doesn’t quite click, perhaps it’s not so fundamentally applicable after all.

This is where tremulus cheats a bit, by introducing hard moves as a consequence. Sure, something happens, but at GM’s convenience. Interestingly, only Poke Around results in GM gaining hard moves, perhaps as a deterrent, to prevent players from constantly finding stuff. Still, I’m surprised Puzzle Things Out doesn’t carry the same risk, as not only does it allow players to learn plot  by quite literally asking the GM, it even grants them an ongoing +1 bonus to rolls relating to the answers, the only easily obtainable mechanical benefit (not that we ever used said bonus, the answers never really felt appropriate).

On the subject of hard moves, I’m not sure how well they work. As a limited resource for GM, it can only serve to limit GM’s options. I sometimes found myself wanting to spend them, because it felt fair and to signify just how badly PCs were screwed. And if I didn’t have any, I would still do the thing, because that would make for a better story. So what’s their purpose, again? Smoke and mirrors.


Speaking of mechanics, the bit I liked least about it are the damage rules. I have my doubts about damage in AWE in general, but in tremulus there are extra issues to contend with. To start with, the rules are all over the place. Say, you’re about to get damaged. If it’s shock, you have to consult page 116 and make a sanity check move, which might reduce the shock you take, or give you an ongoing penalty. Having done that, or having received harm instead, go to page 29 to see if this damage had any extra effect on you, such as losing your footing, letting go of something or even suffering extra damage. You may then wish to flip a few pages forward to see if you can take a debility instead of damage – as long as you remember that’s an option. Oh, and if you got the damage as a result of Resorting to Violence, you may already have had to choose between taking less damage, inflicting more damage, grabbing something or frightening your enemy.

That’s at least two rolls to resolve one action, involving choices that end up cancelling each other out, and requiring flipping through the book back and forth. And instead of driving damage home, it feels like GM being mean to player. GM could say that struggling with the enemy and failing results in your gun being knocked out… Or you could take the damage, roll again and still lose it.

Like a chisel to the stone

That’s the metaphor I’d go with for describing how the game works. You start with a giant boulder of mystery. Inside it, there are veins of hard plot, things you know are true or that you know will happen unless thwarted. The hazards and the framework they build. As players interact with the story, they feel out the stone. As they make a move and roll dice, they take a chisel to it. With each swing, each roll of the dice, chunks of stone fall off, revealing the true plot underneath. That’s the principle of “play to find out what happens” in action, the result of allowing meta-consequences. Clink. The mayor is hiding something. Clink. There’s a doorway behind the fireplace. Clink. There is fresh blood on the blade. Clink.

Players choose where to apply the chisel, GM determines where the crack leads to. As the game progresses, the plot takes shape, and there is less and less of the stone to chip away at, fewer and fewer opportunities to introduce new elements. At least if you’re trying to have a finite story that will have a proper ending. As long as your stone of mystery is not the size of a mountain – and there should definitely be some inaccessible, unsolvable mysteries in a game of lovecraftian horror. So as you’re approaching the end, and everything falls into place, the game loses some of its appeal. You’re just putting finishing touches on the story, polishing off final bits. The true horror is already revealed. Hopefully, by this point you’ll have enough investment to carry you through.

That’s what the game wants to be, and even if I don’t think it quite gets there, it’s pretty damn close.

6 thoughts on “First Impressions: tremulus

  1. That was an extraordinary review! I just discovered that tremulus exists. Your analysis makes me want to order the book and play… if only to find out for myself WTF it’s all about. ;)

    1. I’m not sure it’s possible. The freedom to introduce new plot elements is the strength of the system. Once the story is being wrapped up and no new elements are desired, the system has done its part, and its up to the plot you’ve already created with its help. Which is to say, it doesn’t need fixing, it’s working as intended.

      I can just about imagine an entirely different system, perhaps based on Fate, where the difficulty of introducing new plot elements (setting aspects?) grows with each one already introduced, but the already introduced plot elements have mechanical weight (again, aspects). A natural transition from exploration of plot to its utilization.

  2. I’m totally confused about the “take -1 harm” choice from Resort to Violence. I simply don’t get it. -1 harm from what? Another attack?

    1. Resort to Violence is handled differently from the “attack action” in many other games. When you Resort to Violence, you and your target deal damage to each other simultaneously – it’s a brawl. By picking the “take -1 harm” option, you reduce the damage you take as a result of this brawl, sometimes to zero.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.