Thursday, March 17, 2011

Trail of Cthulhu, games as stories, Tales of the Arabian Nights, and why Mansions of Madness doesn’t really work

I’ve finally had a chance to play the Trail of Cthulhu role-playing game, which uses Pelegrane Press’ GUMSHOE game system. The GUMSHOE system is very interesting, for both boardgamers and role-players. To explain why, I need to back up a bit and lay some groundwork.
What differentiates the sorts of games we like, be it RPGs or boardgames, from other sorts of games is that they tell stories. They may be boring, short, or thin stories, or the story may not be the most important element of the game, but if stories weren’t important, we wouldn’t get pasted-on themes, nice art, or miniatures. The story can be something that is more abstract and visceral, as in Knizia games like Ra, Through the Desert, or Ingenious, but these games still have a narrative arc of buildup, tension, and release that is the stuff of storytelling. Plus of course, there is a large segment of the hobby – which Fantasy Flight is trying to corner – for which the story the game tells is the key thing.
The Cthulhu Mythos is well-travelled thematic ground, with many board and role-playing games trying to capture the flavor of Lovecraft’s popular creations. As always, trying to take a literary story and re-tell it in game format is not an easy proposition, and failures vastly outnumber successes. To see why it’s hard, let’s look at one particular game system that, while popular, is to my mind clearly not a success: the RPG Call of Cthulhu.
Call of Cthulhu is, at a system level, a very traditional roleplaying game. Ever since D&D, the core of role-playing games has been a task resolution system. While the details may differ – the game may use a d20, 3d6, d100, or a pool of d6 or fudge dice – the vast majority of popular RPGs are set up such that whenever players interact with the world of the game, it’s a conflict or a task at which they succeed or fail with measurable probability. When a character wants to accomplish something, we pick a character trait to use, figure out a difficulty number, and roll some dice. The variance between the systems is in the choice of what skills to define and what kinds of probability curves to use.
This is great, but this core system of task resolution simply can’t tell a wide range of stories that people who play RPGs happen to like and desperately want to game. The most obvious are, unfortunately, mysteries, horror, and epics (I use the term “epic” as Stephen R Donaldson lays it out in his monograph Epic Fantasy in the Modern World).
The problems with telling mystery stories are straightforward, and fairly obvious if you’ve ever tried to run a mystery in Call of Cthulhu. The narrative structure of a mystery story is that there is a trail of clues that the characters must gather and piece together to figure out what’s going on. That trail of clues drives the narrative arc. The characters start out with a hint, follow the leads, and over time the truth is revealed. There are all sorts of conventions to the mystery genre which allow readers or viewers to engage with them, but this is the core. This is an incredibly common narrative format, used by H. P. Lovecraft, Stephen R. Donaldson, the X-Files, and Law & Order as well as many – probably the majority – of the episodes of Star Trek or Buffy: The Vampire Slayer. Even the Harry Potter novels are, from the point of view of narrative format, actually mysteries.
The problem of course is what happens when acquiring a clue requires success at a task which the players repeatedly fail? What if there is a witness holding out on them and the players can’t make their Intimidation check to save their lives? Or if there are documents hidden in a room and the players can’t pass a search check? The GM then has to resort to ever-more-improbable ways to get the players the information they need to follow the trail of clues. OK, you blew your search, maybe the contents of the documents was known by an NPC and you can try diplomacy. Blew that too? OK, maybe the documents were in another location. Still not making that search roll? Eventually the documents end up lying in the middle of the road where the PCs trip over them. This is immensely unsatisfying because a) why are we rolling all these dice and jumping through all these hoops when the conventions of the genre of story we’re trying to tell requires us to get this clue?, and b) in the system we’re using, which is all about tasks and succeeding and failing at them, why are we not being punished for all these failures? Because the players are failing all these checks, they can clearly see the hand of the GM coming in and granting them the information they require. To look at it form a narrative point of view, you never have a scene in Law & Order where the detectives execute a search warrant and no information comes out of it. Searching the apartment was a scene in a sequence, and the narratively interesting thing is not whether or not the detectives’ skills were up to the task of finding anything, but what they found, how they went about finding it, how illuminating the information was in light of other clues already gathered, and what they do with the information to move the narrative forward.
This is not to say that good mystery stories have not been told by many talented GMs using the Call of Cthulhu game system. But their success in doing this is in spite of the system, not because of it.
To divert briefly into epic tales, you don’t have to go very far into Tolkien to find story elements that stymie RPG-standard tools of skill checks and difficulty levels (or traditional boardgame tools of resource management, risk, and positional tactics). The epic confrontation between Eowyn, Merry, and the Witch-King cannot be gamed using any sort of task-based system. Tolkien has just spent the last three books building up the Witch-King and the Nazgul as terrifying and powerful, so in gamer-land no rational player who can look at their character sheet and know their odds of succeeding at various tasks is going to resort to direct conflict to take him down. And if they do, and win, does it feel like a victory, or like the GM resorted to fiddling the dice or making stuff up to let them do it, a far less satisfying outcome given the entire structure of the game is based around tasks with predictable odds? There is something else going on here. This is an epic scene where characters go beyond themselves, tying in with previous plot hints, and as such is hard to imagine how it could satisfactorily be done in a games which are driven by probabilities and specific knowledge of capabilities.
To get back to the main topic of mysteries, the GUMSHOE system sets out specifically to tell mystery stories. It recognizes that to do this, a systemically different way to define characters and drive narrative is required. So it defines characters partially in a traditional conflict-based way (because mysteries have fight scenes), but simultaneously in a more narrative-focussed way. Your skill with firearms will be familiar, but your rating as a forensic accountant is different. If you have skill in accounting, the system says that you are sufficiently skilled that no narratively critical clue that can be unearthed using accounting will elude you. Your rating in these skills are not skill points, but narrative points, and reflects the importance of that skill to your character’s narrative. If you have some rating points to spend in accounting, your character can move the narrative a bit if the player can come up with a way of weaving the skill into the story. If so, the character can unearth clues which, while not the core clues that allow the players to solve the mystery in a baseline sort of way, will expand the character’s understanding of what’s going on and perhaps make piecing together other clues easier. It’s important to mention that the GUMSHOE system is not a collaborative storytelling system like Fiasco or Polaris; 3 points in accounting doesn’t give you narrative prerogative to skip the suspect interview and hit the books. But it does allow you to weave the storyline if the GM can figure out how to get you interesting information from your proposed course of action, the more detailed and persuasive the better (perhaps you could use Legal to get a search warrant for a suspect’s banking records, then Accounting to track down information that the GM had originally intended to come out via Intimidate or Reassurance in an interview).
Because it’s such a different way of looking at characters, and because task-based systems are so ubiquitous, this definitely takes some getting used to. A 3 rating in Evidence Collection is not more capable than a 1 rating in Evidence Collection. Instead, the character with a 3 rating has a little more latitude to expand the narrative – the rules refer to it as “spotlight time” – than the player with a 1 rating, if he can effectively weave it into the story. Either character will discover the clue that will get the group to the next scene, but the player with the 3 rating can spend some points to try to direct the narrative a little bit and gain information that, while not critical, will be helpful later or give more detail to the grand picture. So, for example, Evidence Collection may turn up three shell casings, some fingerprints, and a bloodstain, but a 1-point spend might additionally tell you (with some narrative associated) that that the shell casings have been sitting there for four days, even though the crime scene is only a day old. In both cases the players get the two critical plot hooks, leading them to identify the fingerprints or take the shell casings to the lab, but the player who had the spend has some information which may make the picture make more sense as it develops and will make the scene more narratively satisfying. So, we have a systemic way to develop the story in interesting ways that relies on player ingenuity in the application of their skills, but not on crude skill checks. This means that GUMSHOE is very good at the specific types of stories it is trying to tell. It focusses on information, how (and not whether) it is obtained, and what the players do with it, which is the stuff of mystery stories.
Boardgamers actually have had something analogous to this for some time: Tales of the Arabian Nights. In this game, the players choose what skills and traits their characters have – Appearance, Weapons Use, Magic, Piety – and then how to respond to encounters, whether by Negotiation, Robbing, Courting, and so on. Then through the magic of a lot of cross-referencing and a book with 2500-odd paragraphs in it, the narrative of the encounter emerges. Instead of choosing how to use your resources and abilities to navigate an existing narrative successfully, your choices (along with a dose of luck) define the narrative which allows it to be, at times, epic in nature. Like in Trail of Cthulhu, Seafaring is not going to get you out of an encounter with an angry Djinn if there is no water in sight, but your skills and your choices nonetheless help shape the story. This is what makes Tales of the Arabian Nights narratively satisfying, while Betrayal at House on the Hill just feels like a fire-hose of disjointed random events.
This brings us, finally, to Fantasy Flight’s most recent weighty box of plastic and cardboard, Mansions of Madness. I’ve only played it once, so I’m not going to judge too harshly. But, like Call of Cthulhu, Mansions of Madness is trying to tell stories that are narratively mysteries while using the standard boardgame tools of conflict, risk, and resource management. In my opinion, this is a case of “when the only tool you have is a hammer, everything looks like a nail”. The common core mechanics we have in boardgames (and RPGs) are simply not amenable to mystery stories, and Mansions of Madness ends up being a nail, in this case a glorified dungeon crawl. Which is fine, but all the trappings of mystery – the extensive intro text, the flavor of a path of clues – are squandered and can actually detract from the gaming experience, since they may misdirect you into thinking the story is something it isn't.
If we want to tell different kinds of stories, we need to expand our toolbox. Arkham Horror is not a tale of mystery or horror, it’s a tactical game of resource management with the narrative structure (to the extent it has any, which is not great) of an action-adventure with characters being led through set-pieces over which they have no control. By contrast, Castle Ravenloft – which is fundamentally the same game as Mansions of Madness – may not be a classic game, but it’s more narratively satisfying because the tools it uses are appropriate to the story it’s trying to tell and it gets the critical structural bits (pacing and tension management primarily) pretty much right.
Worlds like Tolkien’s Lord of the Rings and H. P. Lovecraft’s mythos are notoriously hard gaming problems, done badly so many times, and these are the reasons why. The very few great games we have work because they’ve limited themselves to portions of the story that can be told with the mechanics available. The brilliant bit of the classic CCG Middle-Earth: The Wizards was to focus on the years between The Hobbit and The Lord of the Rings, when stories could be adventures of risk and reward and not epics. Knizia’s Lord of the Rings works in part because it focuses in tightly on the hobbits, who are modern characters who become immersed in an epic world which is not their own, and also of course because Knizia is a design genius who is keenly aware of how tension management and tight pacing can produce strong narrative structures in games. What success Lovecraftian boardgames have had, they have when they focus on the pulpier, action-oriented face of the mythos at the expense of the core stories that the readers love (it’s interesting to contemplate how much of the veering of Lovecraftian material into pulp is a direct result of a gaming fandom which lacked the conventions to tell the real stories). Clearly there is room for innovative new systems and mechanics that will help us tell these other kinds of stories in enjoyable and satisfying ways. RPGs are leading the way with serious, envelope-pushing titles like Trail of Cthulhu, Polaris, and Fiasco, all designed to tell specific types of stories that would be extremely challenging (at best) to do using more traditional systems. There is no reason these trains of thought can’t be extended into boardgames where the differences between the two blur.


  1. I'll point you to a third way of handling mystery stories, much closer to Fiasco or Polaris in structure – Seth Ben Ezra's Dirty Secrets.

    [Full disclosure: I edited Dirty Secrets and I'm just beginning my third editing job for Seth]

    In it, the narrative is built scene by scene and when a culprit is revealed semi-randomly the players narrate an appropriate explanation of the crime based on what has gone before. This glosses over a lot for the sake of brevity, but for many players it will be much more satisfying than traditional RPG mechanics *and* GUMSHOE.

  2. I think it's important to remember the roots of RPGs, mainly they started as tactical wargames. With that in mind it is easy to see why task resolution by die roll came to dominate game systems. Non-combat resolution almost always feels shoe-horned in as a concession that not everything is about combat.

    That said, I think you exaggerate the failings of many RPGs to handle mysteries. Any good game I have seen punishes you for failure. It's what makes PnP better than video games. You can fail and your failure has consequences on the world. I would hate a DM that makes sure the players always win.

    Secondly, a good mystery (or really any good adventure) requires some fudging of the rules. However, D&D for instance has things like take 20 which will make sure you don't miss that important document. And even if you fail to intimidate a witness, this happens in fiction all the time. I don't know which cop dramas you watch, but the ones I have seen have many tense scenes where it's not clear if the suspect is going to crack. Sometimes he doesn't and they have to go at it from a different angle.

    You make it sound like players are destined to solve a case and any interruption if their beeline to the finale is harming the narrative. To me it's more organic and ultimately fulfilling if things are not so inexorable. Yes, it can suck to DM a game where you prepare something that never gets used, but it's just as rewarding to fly by the seat of your pants as you react to your players.

    Also, I find myself skeptical of systems that allow you control of the narrative as part of the game mechanics. For one, it's not really game design. I can do the same thing with D&D by just relaxing strict adherence to the skill rules. Second, and maybe I am not understanding correctly, GUMSHOE either lets players interject new story elements or discover small bits of information a GM has already prepared. If the former, I wonder how often the piece of information fits with the GMs planned narrative. I suspect the information gleaned is mostly inconsequential. If the latter, this isn't really breaking new ground.

    Bottomline: it puts the onus on the players to create an interesting narrative. But the responsibility was already on them, so what does this really change.

    P.S. Arabian Nights bored me to tears. None of the "stories" were remotely interesting and it felt like a big pile of random.

  3. The tricky thing about Lovecraft is that his stories aren't typically conflict-driven, and they're not really mysteries either; they're more like "terrible discovery" stories (The distinction being that in a mystery the reader could theoretically guess what happened, whereas in discovery you are held in suspense waiting to see what sort of fantastical thing the author has invented).

    How do you approximate that in a board game? The best way I can think of off the top of my head would be a "capture-the-flag" type design where the players must travel into a situation (whether by analogue via pawns or by the playing of cards), uncover something, and then try to escape with their lives. It would be neither a competitive game nor a cooperative game; whoever gets out wins, whoever doesn't loses.


  4. On a couple of points:

    First, collaborative storytelling (like Polaris or Fiasco) vs GM-driven games like D&D, there is obviously a continuum. In Fiasco the players as a group make up a story that makes sense and obeys the rules of the genre (as strongly supported by the game's rules and structure) as they go along. In D&D, the GM drives the experience. But even in D&D, it's still collaborative storytelling: the GM sets the scene, the players tell the story of how they try to navigate it, and the GM adjudicates success or failure, sometimes following rules, sometimes using a process called "making stuff up". GUMSHOE is pretty close to D&D on this particular scale, where the GM runs the story, and the players navigate their way through it. It's a well-made point that GMs do the sort of things done by GUMSHOE already, discarding the rules and improvising and adapting to tell particular type of story. What GUMSHOE does is give you a *systematic* way for the GM to do this adaptation and improvisation, and for the players to help feed into this process, and there is a lot to be said for having a system.

    To pick just one, whenever you pick up a book, some of how you read it is going to be your expectations about the genre. If you're reading a mystery your gear up in different ways and interpret fragments of narratives differently from other forms. A lot of D&D's (and other generic systems like GURPS) problems come from communicating the types of story the GM is trying to tell. If the GM is trying to tell a mystery and the players interpret it as a traditional dungeon-crawl, wires get crossed and problems ensue. It's a valuable to have a system explicitly geared to the type of story you're trying to tell, both because it's likely to work better, and so the players can't miss the context.

    Second, the types of stories the Lovecraft tells actually span an interesting range and can't all be pigeonholed. I am not by any means a Lovecraft guru, but stories like The Call of Cthulhu and The Dunwich Horror are of a different character, one of the reasons Trail of Cthulhu supports Classic, Purist, and Pulp idioms.

    Lastly, some of my frustration with Mansion of Madness is that you've got this "GM"-type player, the Keeper, who is "running" the game, sort of. If you want to tell a mystery story, you have this player resource available to do something interesting with setting up a path of clues or managing the storyline tension. But Mansion of Madness does so little with this player; he's just another player position who does little more than run the monsters and try to defeat the players. Taking advantage of the fact that this is a person who could manage the narrative would seem like an obvious place to start.

  5. The short version of my response, Chris: I agree. For the long version, click here over to my blog (where I cross-posted to here).