Monday, February 6, 2012


So I was reading through a mostly very good supplement and I came upon the following paragraph in the DM advice...

"Be wary of plot construction that demands characters accept captivity to gain crucial information. Many players would sooner 
have their Investigators trepanned by the fungi from Yuggoth than accept even a brief sojourn in comparatively cushy confinement. Unfortunately, with this player type, you won’t get very far by pointing out that getting captured is a Pulp staple. Their attitude is rooted in a deep-seated desire to maintain emotional control, and is not typically susceptible to argument."


Players do not rebel at the idea of scenarios that "require them to accept captivity" because of some psychobabble.

They don't like it because, after the scenario starts, if there is any situation that they cannot get out of simply because the story demands that they not be able to get out of it (remember they are there "to gain crucial information" not "because they fucked up and logically would not be able to escape") then they are not actually playing the game for the length of time it takes to endure that situation.

It's like a cut-scene in a video game. Fine, have it, but don't make me press "talk" "talk" "talk" until the NPC dispenses the information.

This is why people don't like railroading. It isn't playing, it's listening to the GM tell a story, with maybe a little "Ok, when I say Free you say Bird" going on.

Now many players may like having the GM tell them a story, but it is not, strictly speaking, the activity many of them signed up for, so it is as much an interruption in an RPG as the GM stopping play to juggle for ten minutes. Hey guy, juggle, maybe I'll like it, but if I don't, it's understandable since I showed up to play, not watch you juggle.

This is followed by some not necessarily bad but frustratingly--and all-too-familiarly for GM-advice books--vague language about avoiding railroading…

--Structure can be difficult to achieve in the roleplaying medium. Guide the players too little and they lose the thread, resulting in a loose and sloppy narrative that provides none of the neat, order-making pleasure the genre is meant to provide. Guide them too much and they feel that their freedom of action has been taken away from them, and that they’re merely observers moving through a predetermined sequence of events. (As you probably know, this latter syndrome is known in roleplaying jargon as railroading.)

And then there is the typical cop-out…

--The trick to successfully running investigative scenarios is to strike the right balance between the two extremes.

No, the trick is not to balance. (Secret aesthetic law: balance is almost never the answer to anything.) The trick is to:

1) clearly define the term railroading for the GM, and then
2) provide the GM with tools to avoid it if s/he wants to.

So, how do you know if you're railroading?

Its a long fucking story, but basically it's easier to tell when you're not--here's how:

In a non-railroad the PCs are given choices that result in the PCs encountering substantially different (yet still interesting) scenarios than if they had made other choices--preferably scenarios they could potentially have seen the dim outline of while making those choices*. In a railroad, they are not.

In edge cases, it is best to ask yourself How different could this have worked out if the PCs had made different choices? The larger the difference the less of a railroad you got going on.

Anyway, there seems to be a popular myth that while yeah, it is easy to avoid railroading in a location-based D&D adventure, it's almost impossible to not do it a little in Call of Cthulhu or likewise horrorish scenarios and these things are really more about atmosphere and set-pieces anyway so why worry about that?

Well, honestly, if you're happy with that, cool beans and carry on.

If you are not, I am going to describe a very simple but effective way to totally not railroad people in a horror game (or really, any investigation-based game).

Note this is not the only way, but it is, like a dungeon or a sandbox, a simple, durable structure you can use that has meaningful choice built into it.

I call it Hunter/Hunted.

-The idea is simple and comes from about a million horror and cop stories: sometimes a scene happens because Sam Spade has found out about a baddie and sometimes a scene happens because the baddie has found out about Sam Spade. And, there, aside from a few stops for bourbon and kissing, is the plot of everything from Lost Boys to Blade Runner.

-Most investigative scenarios advise breaking things up into "scenes"--the idea is you have a scene, find clues in it, these clues lead to the next scene. They then usually cover their ass by saying either "if the PCs don't do this or find this clue or go to the wrong place give them a bunch of hints or a prophetic dream or otherwise nurse, nudge, or nullify them until they go to the next scene" or just give some vague advice like "hey Venice is interesting, think of something"

-Not so here. Or not exactly: Basically we keep the "scene chain" structure. If the PCs go from clue to clue in a timely fashion like good investigators they follow the scene chain. However, we also give each scene a twin situation, this twin is what happens if the PCs don't follow a given clue, follow it up the wrong path, or otherwise take too long (in-world game time) to follow the clues. In this twin situation, typically, the PCs have taken long enough to figure out what's going on that the enemy has noticed their efforts and started hunting them.

-In the twin situation, the GM decides how long it will take (in-world) and what circumstances are necessary for the enemy to be able to track the PCs and starts the clock running. The enemy then catches up with the PCs at whatever point it logically would if it was using whatever tracking abilities it has (whether this be psychic powers or enhanced smell or because Twitty says she saw Mac D talking to a cop next to the sub shop). The GM may drop clues as to their hunted status along the way.

-In situation 1, the PCs are investigators of horror or violence, in situation 2, they are victims of it. This matches rather nicely the twin roles nearly all horror and investigator protagonists have in these kinds of stories. So at each stage of the adventure you have an "i" version and a "v" version of each scenario.**

-The enemy has a locus: usually in physical space. Like: that island Cthulhu is under or the gangster's safe house.

-At each "Investigator" stage, you move closer to the locus (or finding out where it is). At each "Victim" stage, the enemy sends someone or something out from the locus to get you.

-IMPORTANT: Usually the enemy comes in a different form at each stage. You don't meet actually Cthulhu or actually Al Capone until the last stage. The first few stages feature their agents or otherwise indirect confrontations.

-Also important: at each stage, a clue is left. The clue leads to the "i" version of the next scenario.


Opening: PCs are on their way to England on a boat to investigating the Proverbial Mysteriously Dead Uncle

1i--the nurse is clearly a disturbed woman (in the cafe) you don't know why
1v--terrifying dreams of faceless people (wherever you are) that might make you lash out at your fellow PCs (dream anagram of a doctor's name)

2i--the doctor's office is full of strange specimens (SAN check!) and tomes
2v--a(nother?) nurse is following you and attempting to drug you(wherever you are) (has distinctive uniform from a certain lab)

3i-the doctor's laboratory is filled with horrific creatures (and the doctor's in here)
3v-the doctor appears (wherever you are), a strange creature emerges from his featureless face (wherever you are) --a decent add on here is a chance that the doctor drugs a victim and drags him/her back to the lab or, if injured, runs back to the lab, leaving a blood trail. Either way the encounter at the lab--if it happens--is significantly different than if the PCs had just gone straight there and surprised the mofo

(this adventure brought to you by the lyrics to Sister Morphine)

-The "victim" role may seem passive, but in a sense the victim has more freedom than the investigator in terms of defining the terms of the conflict. Let's say we have the simplest scenario available--the PCs are being stalked by a werewolf. Even if the PCs are totally passive--they have the entire planet to walk around on. Will the werewolf attack them in a taxi, on a plane, in a hotel room, underwater? Depends entirely on where the PCs themselves go. You could even say to them: "The sun has gone down, according to what you've seen, the werewolf will be active within half an hour…". If the PC is in, say New York City, the number of resources s/he can bring to bear or changes s/he can make to the battlefield are infinite.

-Victim scenes require you to think like the monster: you are an insane doctor bent on injecting people with a "sacred" serum that makes their faces come off. You've discovered a flapper, a librarian and a drug addict are trying to interfere with your plans, What Do You Do? Where are you? What tools do you have access to? What information on them did they--via their actions to date--give you access to? Use it.

-In many adventures, there is no reason this system should not be transparent--making it clear through events and other characters that "it is vital to find Whatsitsnuts before it finds you".

--Broadly, any structure which has meaningful choice works like this: there are interesting choices leading to events, each event gives the players a unique resource or unique problem they did not have before--even if they all lead to the "same" place in the end, players end up there in the Final Chamber with different tools, perspectives, advantages or disadvantages.

-How do you avoid even the appearance of railroading? Just remind them how they got where they are. Go ahead and tell them where they'd be had they done otherwise.

-Another avoiding-railroading tip: include problems that are clear but open-ended, like: this puzzle requires you get us something old, something new, something borrowed and something blue and shove it in this slot. A real-module example is the "tooth door" in James Raggi's "Death Frost Doom"--you need to drop a tooth in a basin to open a certain door. Where you get the tooth from is your business, but the place is fairly isolated. There are a lot of graves around…but also a few other PCs…and that NPC down the mountain...

*i.e. "left door or right door" is little better than a railroad, whereas "left door that, if listened at, reveals a chittering noise or totally silent right door" is definitely a non-railroad.

**Important clarification:
The Victim scenario is not nursing the PCs back on track by doing an ass-pull. In Hunter/Hunted the "Victim" scenarios are entirely logical, predictable-using-internal-gameworld consequences of the PC's failing to find the foe. i.e. if you don't show up at the safe house in 3 days, that's enough time for the goons to track YOU down. No alteration of the gameworld's internal logic or consequence chain is required. Pure cause-and-effect. the villain is not running into you by coincidence, it's because you dawdled and they managed to track you (thus the ticking clock).


-C said...

(Secret aesthetic law: balance is almost never the answer to anything.)

SHHHHH!!!! It's a secret to everybody!

Seth S. said...

This is pretty great stuff and something I can definitely use for my modern fantasy game.

There are definitely sessions that have more investigation than anything. Mostly I've just been throwing together the most basic structure and just winging it, but this system should ensure my decisions are logical (logical for insane demon princes anyway)

Zak Sabbath said...


Cool, glad to hear it wasn't just "Duh, Zak, we knew that"

huth said...

So does that mean your group is "prone to flailing about"? Heh.

huth said...

Most scenarios in the game do already do that though, although without the explicit or specific mirroring.

huth said...

(published ones I've read, I mean, not, like, every random thing I scraped together from scribbles.)

Zak Sabbath said...


I think describing the specific mirroring is important since (like I quote) game designers--especially with horror--seem to get sooooooooo vague about what is and isn't a railroad and how to keep the scene going without doing it.

And, no, actually my players tend to follow the clues more often than not.

huth said...

The thing is, that entire section is mostly hand-holding for (apparently) people used entirely to planning location-based scenarios.

I mean, if they write

If you’ve prewritten an adventure and the PCs veer off, that’s fine,

then it's obvious that this is a pep talk to avoid failure due to reticence, not a systematic critique of a bullshit term.

Zak Sabbath said...


"Railroading" means an extremely specific thing, like "hit points". If it is addressed, it should be defined.

In this case, it was not defined, and the hand was held poorly.

huth said...

This was making sense to me until you brought up hit points, which never make sense.

Aaron F Stanton said...

The first time I ran Castle Ravenloft, the players took one look at the castle looming at the end of the winding rocky path and said "Aw, hell no" and proceeded to scout out the base of the thousand foot cliffs behind the castle. Then the monk with gloves of climbing and swimming skimmed up the sheer face with the intent of lowering a rope and bringing the party up to invade the castle through the back.

Of course, Strahd leaned out the window and fireballed the monk when he got close enough...

Try hanging onto the side of a cliff when you're busy dodging a fireball. Thud. (Lucky for the monk, he was high enough level to minimize the falling damage.)

Point is, even when a module railroads players into walking through the front door, they can and usually will adapt. A decent GM will adapt to their tactics as well.

Zak Sabbath said...


definitely, but the perplexed sometimes need help

huth said...

There's only so many ways you can say "Chill out and roll with it, man." Maybe 5e should include some relaxation techniques and soothing GM mantras.

Welcome to Dungeon! said...

Also, there are always more and more perplexed people given that the vast vast majority of GM advice for like 25 years falls into, basically 'how to railroad,' 'how to make your railroad inobvious,' and 'how to get your players to like being railroaded.'

mordicai said...

I set up my notes as a checklist-- events that will, could, or might come to pass-- & let the players sort of meander through as seems fit. Or go off book. Or skip them. Or whatever.

In general, I find players don't mind being in long as they can DO things. You gotta give them alliances to start making in the cell, or letting them fiddle with the "ancient & forgotten chamber" that the bad dudes have started using as a make-shift cell, or SOMETHING.

Matthew Miller said...

Awesome post, Zak. I'm a big fan of investigative adventures, and I'm always interested in tips for making them more sandboxy and less railroady, so if you want to kick off a series of posts on how to do investigative RPG right, you'll get no complaints from me :-)

In the past, I've used a tip from GURPS Mysteries which recommends that investigative scenes should be alternated with action scenes. That way the players won't feel like they never get out of the CSI lab. Your suggestion of prepping twin "investigative" and "victim" scenes dovetails nicely with this. Justin Alexander, in his excellent essay on the Three Clue Rule offers similar advice. If the players don't have the information they need or they're using the information in the wrong way:

"The solution is to have something active happen to them. Raymond Chandler's advice for this kind of impasse was, "Have a guy with a gun walk through the door." My typical fallback is in the same vein: The bad guy finds out they're the ones investigating the crime and sends someone to kill them or bribe them."

richard said...

first, yes. Your victim/investigator scheme is helpful, thanks. I will use it more consciously in future (I guess I was kinda using it before but hadn't thought it through). And I've felt validated in Masks by beating the bad guys' clock rather well, and I've even relied on the bad guys bringing the fight to me, so unmasking this process seems to be a Good Thing mostly.

But. a) the psychobabble rings true for me because I usually play war, not sport, and getting captured is a terrible violation of war principles, while b) being comfortable with the genre conventions of hunter/hunted seems kinda sporty to me - that is, the enemy probably won't treat you in a war way, because if they did then they'd use overwhelming force straight out of the gate.

I hope I haven't compressed that too far for it to make sense. please say if it's unclear.

Zak Sabbath said...


I get what you are saying (and am impressed by how we are all using Daztur's war/sport terms already) but:

A) i don't understand what being captured solely because the GM's plot requires it has to do with "war principles. Or what any of that has to do with "maintaining emotional control". You can interrogate prisoners. There are good reasons to take prisoners in a wide-open no-railroad game, there is never a good reason to railroad.

B) I don;t think you have to assume mystical entities and gangsters fighting cops would necessarily "use overwhelming force" immediately because, well, look at the millions of reasons not to that screenwriters have figured out over the years. There are many reasons not to risk that until the enemy is at your door.

fauxcrye said...

How do you feel about complete failure on the player's part or a sequence of events that lead outside of this structure? In fact to other structures entirely. Though I suppose at that point you are just jumping on a different chart. Where they find in the end they were too late or made too much noise and the enemey has gotten wise to them and changed their locus? This change of locus can occur to avoid conflict with the heroes or because the villians have successfully hatched their plot. A successfully hatched plot can really change the focus of the players for a while. In a campaign setting you can lead this to another story using the same enemy or misdirection.

* Heroes are offtrack, Villian leaves clues that misdirect the trail to someone else. This misdirection could help further the goals of the big bad if heroes don't realize they are being used.

* Villian escapes to cause problems another day. The most natural form of continuing villian is one that actually thrives due to player actions and not artifical DM means.

I use similar model for my campaigns in general. And often have to revise them as players change the status quo around themselves or pay more attention to something else. Sometimes allowing ignored things to come back as a different threat.

It is a good device and you really outlined it well.

Zak Sabbath said...

Since this structure assumes the enemy doggedly pursues the PCs if the PCs do not pursue the threat, "complete failure" just means the PCs die, which is always acceptable.

The only outside-the-structure situation I can possibly imagine is if the PCs build an impenetrable barrier to be pursued yet ALSO decide not to pursue the enemy, which seems so boring that I can;t imagine one of my own players doing it.

thekelvingreen said...

I could have done with reading this a week ago!

Rasa said...

Wow. You read my mind, and wrote up a post that answers a ton of questions, and codifies a lot of nebulous thoughts I've been having. Having this in mind will help me out so much, so thanks again Z.

fauxcrye said...

I guess it is called hunter/hunted for a reason since it outlines that particular model with only those possible outcomes. I was reading a bit more into it I fear. But still a good model that can be expanded upon.

Mr. Blue said...

Maybe it's a cultural thing, I mean, compare the hated Amtrak to the beloved Eurail system. Some cultures just like railroads.

I'm glad I live in America. Driving cars is way more fun.

richard said...

@Zak: by "violation of war principles" I mean that, if the investigators and the bad guys are both "at war" then getting captured is never good. The players have to maintain control over their characters' freedom of action - if they fail to do so then they've already lost (and that can work too - then they are in the bad guys' power and it's up to the bad guys to do something interesting if you want to reveal more mystery, including the mystery of "why didn't they kill us?"). Accepting being taken prisoner as an "in genre" thing implies sport expectations.

Now if the bad guys do allow the PCs to continue investigating having "won" in this way, that implies either they're up to some mysterious thing involving the PCs, or they're playing according to sport rules, while the players are still at war. I've actually played in games where that was the case and it's weird and annoying and less satisfying, IMHO.

Zak Sabbath said...


you haven't addressed the issue i put forward of how this relates to justifying the "psychobabble" ("emotional need" is not the issue--sitting there doing fuck-all because the plot assumes it has to happen the issue) or what I said about how there are perfectly good cause-and-effect reasons to take prisoners.

Basically, please start over, because this isn't making any sense to me.

Will B. said...

I like this structure, will probably refer to it if/when I get around to running Dark Heresy. But...

If the players become aware that this is the structure you're using, do you think they might learn how to game it? What I'm thinking of is if they know that they can sit on their asses and will always get their quarry coming to them, then why not just forget about investigating and spend your time on building the best possible trap for when the villain attacks?
For example, in your Werewolf in New York scenario, my first thought would not be "Shit we have to find that werewolf soon!" but "Let's get some silver bullets and wolfsbane and find a suitable place for an ambush."

Zak Sabbath said...


Good, ambush the werewolf! That's smart thinking, that's the way it should be.

However I would not reveal the structure unless it was revealed in-game for some reason.

I also would not use it over and over without varying it or throwing curveballs. This is just a basic outline of how to do these kinds of plots without railroading--there are infinite baroque variations.

Seth S. said...

I wouldn't mind a post on variations you've used and thought up. Even specific examples would be fine.

Like I said earlier this is pretty relevant to me.

John Johnson said...

Great post. I especially love the advice about investigation/mystery, because structuring that type of scenario has always been difficult for me.

I think that lots of railroading can be avoided if the GM simply remembers that the NPCs have lives outside of whatever the hell it is that the PCs are doing.

If the PCs are hired to find the guy who's sending death threats to the mayor and they stumble around in the wrong part of town the mayor will end up dead and now the session becomes "Let's find the guy who did it" instead of "let's stop the guy from doing it".

Jack said...

This is real clever and I wish I'd read it before I started my Call of Cthulhu campaign.

Black Vulmea said...

My basic rule of PC/NPC interaction: If the PCs are winning, then someone else is losing. There's always someone unhappy about the adventurers' success, and if they're unhappy enough and have the means, they will try to do something about it.

Hunter/hunted strikes me as similar to that.

richard said...

@Zak: re "emotional need" is not the issue--sitting there doing fuck-all because the plot assumes it has to happen the issue

I agree with you, that the central issue really is that when you're railroaded, you're not playing, and that there is never a good reason to railroad.

I think there's another, supplementary point to be made, though - a supplementary problem - which I think (interpreting here) the psychobabble tries to address, to do with how such railroading plays out emotionally at the time it happens.

I've been in this situation, where the DM assumed we would be captured and was eager to get on with the plot reveals contingent on that. And we players acted the way the psychobabble predicted. Partly, I think, for the reasons you stated (we disliked being railroaded and wanted to play, not sit passively) and partly (most immediately, viscerally) because our war style of play was based around avoiding falling into the enemy's power, maintaining maximum freedom of action, and defeating the enemy before they defeated us. So the DM was fighting against two powerful frustrations on our part: of not being allowed to play, and of watching our own failure and not being able to avoid it. And I for one tend to act desperately if that happens - I don't just go "this sucks, I can't play," I try anything I can to avoid the outcome because I imagine that perhaps there's something I haven't tried yet, that might avoid defeat.

richard said...

I am so sorry this is so long. Right now I can't think how to make it shorter.

Re taking prisoners: Yes, sure, this can happen and often does, in history and in games. Sometimes people even choose to be taken prisoner. But all those cases are special in some way - you have to believe that you're not committing suicide by being captured, and IME the DM has not communicated that to us adequately, generally I think because they wanted dramatic tension. For Bond, that moment when he's surrounded by men with guns means an uncomfortable but reliable short-cut to the villain's lair, but Bond is a self-aware, knowingly "sporting" setting. In RPG combats that I've been in, getting surrounded by guns has meant tactical failure, probably to be followed up by TPK, and I for one will pretty much always hope to get lucky with the dice and escape, rather than hoping that they don't shoot in the next round (even, funnily enough, when playing Victory's Bond RPG).

And it doesn't help that IME, even if a DM says "come quietly and we won't hurt you," it's rarely a good idea to believe them. Indeed, IME capture has generally meant something worse than simply being killed - you might be replaced by a ringer or some other sneaky trick, which is otherwise hard to pull off. Try telling Telecanter's players in the Tumbling Dungeon that they might want to try being captured, after half of them have been brained in the dark by witches and the one guy who has been captured and returned is no longer trustworthy. I've mostly played CoC, with implacable, unreasonable enemies who, if they offered to parley with you, were probably setting you up for a devil's bargain. The potential fallout from a captured-and-released PC is too tempting for the DM and too dangerous for the party - half the time I think we'd probably kill anyone who was returned to us.

And if we, the party, captured a cultist/conspirator we would probably kill them right after we'd got all the information we could get out of them, and then burn the body to prevent it coming back as undead. Or we might use them as bait, or any number of other sneaky war moves, but we would make damn sure they didn't escape the way we would hope to as PCs. Maybe that experience is not typical and other tables have more frequent and sane capture situations.

Re not using overwhelming force immediately: yes, you're right. I withdraw what I said. I was really thinking, as an adjunct to "don't railroad" I would add "keep your villains rationally villainous" or, more generally, "don't nerf your villains." But somehow I wound up saying something else.

Tedankhamen said...

That is why I always liked the UK Cthulhu supplement 'Green and Pleasant Land.' You had a mad doctor summoning Hunting Horrors to attack his targets and the PCs but leaving clues as to his whereabouts. I've long lost the book, but as I recall all the adventures had this hunter/hunted aspect, including the wonderful fiction about a haunted house at the end.

August said...

Trail of Cthulhu contains a lot of advice to GMs on how to avoid railroading. But I constantly get the feeling that it's really about how to maintain the illusion that you're not railroading. Thanks for giving some advice about how to do mystery scenarios where the characters' actions actually matter.

Steve Lawson said...

What you have outlined, Zak, reminds me of the mechanics and advice in Mouse Guard, when it comes to PC failure. The emphasis there isn't on avoiding the railroad, but rather on not letting failure become a dead end or TPK.

So in Mouse Guard, if the PCs fail to investigate a scene properly or fail their rolls to notice the map they need, the DM can either deplete their resources by making them angry or tired or sick (which has mechanical consequences) or (and this is explicitly preferred in the rulebook) introduce a plot twist, such as having a hostile NPC burst in and demand the map, or having the PCs wander far afield from their goal before they meet an NPC who gives them the information or whatever.

Zak Sabbath said...

@steve lawson

There are extremely important differences:
insulating the PCs from TPK, having an NPC show up and give the PCs the information they failed to find, or introducing plot twists just because the PCs lost the thread is called "nursing" and is essentially a railroading technique.

In this scheme, becoming 'hunted" is the _logical in-world consequence_ of PC failure in the actual situation, the fiction does not contort to prevent failure, it has the possibility of failure built in.

However, since Mouse Guard is (as I understand it) kinda written for little kids, I can see how nursing might be the best solution in that game.

Steve Lawson said...

Yes, Mouse Guard is built to insulate the party from a TPK--it's not that kind of game, and the rules are clear that mortal combat should be rare and not a surprise.

My example of the NPC giving them the information was admittedly lame--I meant for the main point to be that they'd have to go out of their way and encounter other challenges that they could have avoided if they'd just found the map in the first place.

I'm not sure I understand the difference between "introducing plot twists just because the PCs lost the thread," and giving "each scene a twin situation, this twin is what happens if the PCs don't follow a given clue, follow it up the wrong path, or otherwise take too long (in-world game time) to follow the clues." You don't think that "plot twist" and "twin situation" are virtually synonymous?

Zak Sabbath said...


This is very important:

In Hunter/Hunted the "victim" scenarios are entirely logical, predictable-using-internal-gameworld consequences of the PC's failing to find the foe. i.e. if you don't show up at the safe house in 3 days, that's enough time for the goons to track YOU down. No alteration of the gameworld's internal logic or consequence chain is required. Pure cause-and-effect. the villain is not running into you by coincidence, it's because you dawdled and they managed to track you (thus the ticking clock).

The way you describe the "plot twists" it sounds like the PCs could miss the point of the scenario and then, just by concedence, something appears to pull them back in to patch up the hole.

Steve Lawson said...

I think that falls into your "get a better DM" logic, and I'm the mediocre DM due to my poor explanation. I'll try one more time; if this doesn't convince you, I'll stop beating a dead mouse.

In Mouse Guard, a world that is essentially hostile to mice takes the place of the Cthulhu-style horror in your example. So if the PC mice fail to find the NPC mouse they are looking for, perhaps it is because they took too long and in the interim the NPC mouse has been eaten by a snake--entirely logical, as mice have to constantly be on the alert for predators. Now the PCs have to deal with the snake, and in dealing with it, have the chance to find the NPC's diary or whatever information they need.

Zak Sabbath said...



A: you know the "V" scenarios ahead of time and they make perfect sense given the bad guy, or

B: you made them up because the players didn't do what you expected.

If B, you are in danger of (but not necessarily) railroading. If A, you are definitely not.

You don't adjust the internal logic of the game to make the game easier for the PCs. You design it, wind it up, and let it go.

Do not stop posting if you don't understand, this is important.

Steve Lawson said...

Yes, in MG, you should prepare the twists ahead of time so they make sense, though I suppose a particularly hapless group of PCs might exhaust the prepared twists. In MG, the PCs are supposed to have goals and the obstacles they encounter are supposed to threaten those goals. If the twist makes it so it really doesn't matter if they beat the obstacle or not, then I'd say that was a bad twist and a pointless obstacle.

I believe I understand your emphasis on things making sense and also letting the chips fall where they may. But it also seems that, like Mouse Guard, you are using your I/V structure to keep the game from hitting a boring dead end if the players miss something important either through role-playing fail or dice fail. In your game, if things stall out, the GM pulls in the bad guys in a logical way. In MG, if things stall out, the GM pulls in weather, animals, or other mice in a logical way. I'm not sure they are the SAME thing, but I do think they are very similar ideas for keeping a game from screeching to a halt while also introducing consequences and complications that could have been avoided if the PCs had succeeded in their initial task. I don't know if it's in the MG

So what would you say to someone who looked at your diagram above and saw two parallel railroads, the "I" line and the "V" line?

Steve Lawson said...

Oops, just ignore that sentence fragment at the end of the second paragraph.

Zak Sabbath said...


The person who saw "2 parallel railroads" would be delusional or illiterate because I defined railroad in the post and the situation I describe does not match that.

There's nothing stopping a PC from hiding from the railroad in a silver railroad car and sniping it with a silver bullet, or building a hot air balloon and flying away or any other thing their heart desires.

Zak Sabbath said...


As for Mouse Guard, I think this basically butts up against the design problem of "You can either have situations that challenge the characters' beliefs or you can have open-ended choice but it's hard to have both" that makes most Burning Wheel & Story-Games people not the same as sandbox D&D people.

However, I can imagine an "ethical sandbox" where the PCs are outlier revolutionaries in regard to any possible part of the universe they inhabit, and so their beliefs would constantly be challenged merely by attempting to survive. Or a vampire scenario. Anyway I presume it is possible. Just not easy.

And, for my group, not fun.

Though we are still waiting for Connie to run Mouse Guard.

Steve Lawson said...

I can't wait for you to play Mouse Guard. You will hate it. :)

Zak Sabbath said...


re: parallel tracks:

The Hunter/Hunted scheme is like saying "everything in the world is either a dog or it isn't".

That's a simple idea, but it's not restrictive. And it's true.

thekelvingreen said...

I'd love to see a sample scenario, as I sometimes have trouble following your pure theory posts, Zak. I think I get what you're saying, but a longer example -- if you have time, which you probably don't -- would be helpful.

I just finished running a published Call of Cthulhu campaign. In terms of plot and event sequences it was very linear as written, which was a bit of a worry, but on the other hand, there was enough detail in there on the antagonists and their plans that I felt I could chuck out the plot.

As the players bumped into the antagonists' plans I tried to then respond according to what I knew of their personalities, and it seemed to work quite well; I think this might be close to your hunter/hunted model as I understand it, in that the NPCs reacted to and against the players' actions, which led to quite a deviation from the published storyline, for the better I might add.

Things got a bit sticky when the players broke the plot, but even then I was able to offer up a viable -- if not as impressive -- "V" scene.

Jens D. said...

Very nice! It's a good concept to keep in mind as a GM. I try to implement something (I think) very similar into my campaigns: the PCs may do what they want, but the things, they took responsibility for, escalate every time a little bit more when they decide to stall. This may not have direct consequences for them (like, for example, the cult the PCs are trying to stop, claims another victim), but it might mean any particular bad guy might take an interest in them and act accordingly (the cult finally finds out who is responsible for the desecration of one of their shrines...). This way I have no static setting and the PCs are able to do as they please. Add any number of sidequests and stakeholders and you pretty much avoid playing a railroady game (as I understand it).

huth said...

I just finished running a published Call of Cthulhu campaign. In terms of plot and event sequences it was very linear as written, which was a bit of a worry, but on the other hand, there was enough detail in there on the antagonists and their plans that I felt I could chuck out the plot.

Many of the later ones suffer from that same problem—they want to end with a big scene of a monster coming out of someone's face, or a suitably gruesome irony for the villain, but have a hard time layout out the maybes or means of pulling things back toward the intent. That's why I like Pagan Publishing adventures from about 1997 on (although The New Age is occasionally guilty of pre-plotting); they're (deliberately, I gather) chunkier in exactly the detail of antagonists and intentions than scripted scenes compared to many Chaosium adventures from the same era (I'm thinking of BtMoM and Dead Reckonings in particular).

huth said...

I have some sort of visceral, hard-to-describe issue with some fundamental assumption of this scenario structure, and I'm having trouble putting it into words.

I think it has to do with the lack of twist and deception—as a Keeper, I feel like I have to be actively fucking with my players enough that the optimal/suboptimal columns should be inverted; if they kill the doctor or investigate the obviously creepy nurse then they're falling into the gruesome demise line.

Zak Sabbath said...


well let me know when you figure out exactly what that is

huth said...

I'll try this: The tactical element in an investigation isn't whether you stab a dude or the dude stabs you. It's assembling the history of whatever happened correctly. Whether or not the dragon eats you at the end is irrelevant.

huth said...

So if you figure out the mad doctor's secret and he still kills you, you've still beat him.

Zak Sabbath said...


I will re-quote myself and you can decide whether what you are tlking about matches this or not:

In a non-railroad the PCs are given choices that result in the PCs encountering substantially different (yet still interesting) scenarios than if they had made other choices--preferably scenarios they could potentially have seen the dim outline of while making those choices*. In a railroad, they are not."

David S. Goodwin said...

I like this.

I also like that you defined railroading; I worry sometimes that I railroad, but I meet your criteria of informed choices leading to radically different outcomes from any scene.

It occurs to me that the Ultimate GM would naturally know how every NPC responds to every scenario; there would be no railroading, since every result would arise naturally from the fiction. But since the Ultimate GM doesn't exist, we need cool tools to approximate that result. Poorly designed tools = railroading; well-designed tools just reduce brain energy required to model the internal logic of the fiction.

huth said...

@ David

Getting the players into internal fictional logic versus D&D fictional logic (wherein capture = living death instead of capture = just another problem) has about thirty years of oral culture training to fight against.

huth said...

**internal genre fiction logic

huth said...

@ Zak

Why do you need a bifurcated linear scene chain or multiple twinnings to prep that instead of a floating pile of foreshadowed antagonist actions though?

Zak Sabbath said...


Who the fuck said "need"?

Again, I will quote myself and you can decide whether this answers your question or not:

...I am going to describe a very simple but effective way to totally not railroad people in a horror game (or really, any investigation-based game).

Note this is not the only way, but it is, like a dungeon or a sandbox, a simple, durable structure you can use that has meaningful choice built into it.

Anonymous said...

“The "victim" role may seem passive”
Yeah, I’ve tried this hunter/hunted structure a couple of times over the years and it can fail just because of that. Sometimes it does work but I’ve more-or-less given it up for now.

huth said...

@ Zak

It doesn't, because I'm specifically wondering why you need to bother establishing the chain of events beforehand instead of prepping three player reaction steps, one-to-three NPC actions, and just dumping each clue on them once, and letting it arrange itself in play? What's the specific benefit to prepping the investigation that comes from restricting potential access to information at the beginning, turning it into a chain instead of a sandbox? Especially in regards to prepping antagonist reactions: why not establish a consistent modus operandi, early on (when they find the dead uncle's body), and then derive overt/covert reactions from that? (I don't mean that rhetorically. What are you looking for when designing a serial killer?)

Relating it back to the dungeon, dumping them in a room with three secret doors and a hidden trap at the beginning seems like a better bet at getting them to find one than just one door and just one trap.

Zak Sabbath said...

"What's the specific benefit to prepping the investigation that comes from restricting potential access to information at the beginning, turning it into a chain instead of a sandbox?"

Your solution is fine as well, it's just some people are too inexperienced or time-constrained to do that or deal with the potential consequences of doing that.

Plus, y'know: variety.

There is no "pre-arranged chain of events" however--if you think that you are reading the post wrong. Merely locations and situations that (as in life) trigger certain kinds of events.

huth said...

@ Zak

It's just that, were I to actually prep that way, it would end up as a non-choice—since the only thing I can really predict is that the players will definitely interpret it wrong. I know they will. So what I have to prep, practically, however player-choice-empowering or not, isn't really what happens if they do it wrong but all the different possible ways they can do it wrong. Otherwise, I'm going to be flailing about the entire session.

Zak Sabbath said...


" the only thing I can really predict is that the players will definitely interpret it wrong. I know they will. "

Well then your whole thing is predicated on conditions unique to your group and I am not psychic so I didn't write a post directed at them.

Zak Sabbath said...


or maybe not "unique", but certainly conditions that are not necessarily broadly applicable.

Zak Sabbath said...


or maybe "ok, broadly applicable, but not to my group or groups in some respect like it which, in the end, are the only people I can talk about with any authority"

huth said...

@ Zak

#3 leads me to think my issue with it is a local version of a pretty broad problem. A more, uh, empirically-treatable version of the discussion would have to be "how much information is enough to hit/trigger/whatever the choke-point"? I'm sure DMs with more years of experience than I can weigh in; I can only shrug and say that I'm going to give them a freebie at the start.

Zak Sabbath said...


It is rare that I come upon a comment which I fail to understand in any way but it has happened.

That link went to something unrelated and the #3 refers to--what?

I am very confused

huth said...

You're right, though, my group is probably a little unusual compared to others, since I only started playing when I was 17ish and most of my players started in their 20s with this group, which was the first time I DM'd more than one session. There's a whole oral culture of playerdom which we're not part of.

huth said...

@ Zak

Haha, it was a link to your orcfanticide list post.

Zak Sabbath said...


an inability to do what one wishes to do is, i suppose, regrettably common. This post is for people who can't even decide what they wish to do.

huth said...

To try and rephrase, my question is: when or how does the GM know, or how to they decide, that the players know or understand enough to be making the choice proactively and definitively not reactively?

Zak Sabbath said...


It is a matter of knowing your players very well.

huth said...
This comment has been removed by the author.
huth said...

I should've left my out-of-sequence post because now it just looks like I said something dickish and deleted it.

Paul Staxx Spraget said...

Just stumbled upon this great post from the Story Games forum. In the past I too have been frustrated with the clunky 'breadcrumb trail' of clues present in investigative RPGs such as Call Of Cthulhu.

I recently discovered this thread on The Forge about approaching the problem from a completely different perspective, and it blew my mind and made my subsequent investigation scenarios about 500% more satisfying and naturally fluid. I recommend you check it out.

Zak Sabbath said...

Linking to a whole thread by multiple posters doesn't really clear much up--which part is the part you found useful?

Anonymous said...

To be specific, my forays into investigative games like CoC have been centered around a “big reveal”, and I believe it’s that focus on a “big reveal” in-and-of-itself — as opposed to sandboxier fare — that I have had problems making work at the table. My apologies (two years later...) for placing the blame more on the hunter/hunted structure, and I misunderstood your take on hunter/hunted. Thanks for sharing your GM:ing techniques. I’m going to stay away from CoC and similar games for a while.

Zak Sabbath said...

Hey that's neat that you came back to note that, 2097