This post originally appeared on between 2004 and 2007.

A column on roleplaying

by Robin D. Laws

[Ed.This was originally an internal design document, but it should come in useful for anyone interested in GUMSEHOE background creation.]

The GUMSHOE system departs from standard RPG design practice in a couple of significant ways. Neither of the two extant rules manuscripts, Esoterrorists and Fear Itself, expends much precious space explaining the theory behind these choices. To design rules add-ons for GUMSHOE, though, you have to think in the way the system demands. This document shows you how think GUMSHOE.

Tediously Obligatory Disclaimer

Before we start, please note that just because GUMSHOE makes a certain game design choice doesn’t mean that we’re saying that all games should be this way, or that these design choices are objectively better than others in all cases. They’re right for GUMSHOE. It is meant to perform a specific job. Other games built to achieve other ends might arise from completely opposite principles to be ineffably awesome. A crunchy, rules-driven, determinative, simulationist, integrated game could rock. It is not this game, though.

Design Watchwords

The design watchwords for GUMSHOE are:

  • Emulation of narrative structure (not simulation of imaginary reality)
  • Technique (not rules)
  • Simplicity (not crunchiness)
  • Modularity (not necessarily integration)

Emulation: The ultimate goal of GUMSHOE is to foster play that feels like a mystery novel, TV procedural, or occult adventure comic. Precisely what’s being emulated differs from kit to kit. The first question when designing a rule is: “How would this happen in the source material?” Supplementary questions include: “What structural effect do scenes involving this have on the story? What effect are they meant to have on the audience?” The logic is literary and structural, not literal or reality-oriented. If you design a rules add-on and its result is to encourage behavior or activities that characters in this sort of investigative fiction never engage in, you’ve gone off track, substituting extrapolative logic for dramatic logic. Do not attempt to introduce more verisimilitude than the source material requires.

Also, respect the power of clichés. Sometimes they are required to allow the machinery of genre plotting to work. Many players engage in roleplaying to get up close and personal with their genre expectations. A new spin on a tired trope can be fun, but if your add-on allows only a revisionist take on the material you’re emulating, you’ve created something eccentrically limited. Sometimes it might be appropriate to be self-aware and ironic about the clichés that come with your territory—in a Scream-type scenario, for example. More often you’ll want to find ways to make clichés feel fresh and powerful again.

Because we’re emulating narrative structures, not simulating an imaginary reality, scenarios should not call on GMs to make random determinations for anything that matters. Don’t tell us that the guards will react violently if they roll X and peacefully if they roll Y. Tell us that the guards will react violently under condition X or peacefully under condition Y. Better yet, make these two conditions dependent on player choices or the use of their general abilities. Give us decision trees for GMC actions and reactions, depending on how the PCs change the situation. (General notes on GMC plans and motivations might be preferable in many cases.)

Theoretically there is an investigative sub-genre which, to emulate properly, requires you to ignore any of the other pieces of advice given in this document. If so, do it—but be clear that your deviations from the norm address only that sub-genre.

Simplicity: When designing a new rule, challenge yourself to find the simplest possible expression of it. The urge to complicate is powerful, but must be resisted. Avoid crunchiness creep. Other games put rules front and center during the play experience, and that can be cool, too. But here we want the rules to get out of the way of the GM and players. Episodes of rules use should happen quickly, and take up only a small percentage of any given game session. Just because a rule is cool, doesn’t mean that it is necessary. A rule is never an end in and of itself; if it doesn’t justify itself, it’s so much mental smog.

Technique: The best way to keep a rule simple is to have no rule at all. A technique is a structured way of playing, for GMs and/or players, into which numbers and die rolling do not enter. The flashback concept from Fear Itself is a prime example of a technique. It shapes play in a distinctive way and refers to a narrative technique players will know from fictional sources. It is purely a novel way to perform interactive scenes, without a mechanical reward or consequence.

Another example of technique would be the stereotypes from Fear Itself. Where another game would realize its desired archetypes by giving them rules properties—making them templates for character creation, directly determining your game statistics—this is a simple list that you can take or leave. It is a springboard for player creativity. Again, it gets the players thinking about the source material, but leaves them free to realize them in whatever way, and to whatever degree, they want.

Modularity: First edition AD&D is a modular rules system; sub-systems operate independently of one another. No particular effort is made to make PCs and monsters conform to the same scale and list of capabilities. When Gary and company needed a new rule, they thought, “how do I make this work?”

Third edition D&D is an integrated design; all of the rules systems interrelate. When the designers came to each rules subset, they asked themselves, “How do I make this work in a way that’s congruent with the rest of the system?”

Design integration is considered an important goal for state-of-the-art crunchy games. Integrated rules are aesthetically satisfying and ought to be easier to learn and remember.

GUMSHOE is a modular system, with a twist. Where you can maintain congruence with the existing rules and still emulate the source material, you should do so. However, emulation takes precedence over congruence. The key example here is the way that the game handles abilities completely differently, depending on their relationship to narrative structure. Investigative abilities work one way; general abilities use a completely system.

Aesthetic neatness never takes precedence over function. Note how in GH some of the Psychic abilities use the Investigative mechanics, and others use the General. Again, this depends on their story role: whether they are used to gather information, or to handle threats. In the first case, failure is not permitted. In the second, it is.

There are also actions that use different abilities (and rules sub-systems) depending on their narrative consequences. In GH, you might use Investigative Procedures to find a hidden item that provides information, or Sense Trouble to find one that endangers you. In Esoterrorists, you’d use Explosive Devices to find a bomb whose placement doesn’t threaten you, but does provide a clue. If its primary purpose in the narrative was to threaten you, you’d use Surveillance instead.

Reassurance (gaining information) and Shrink (healing psychic damage) provide another example. Similar according to real-world logic, very different when you look at narrative effect—and therefore treated with different mechanics.

Another split: PCs are treated differently than supporting characters. GMCs need general abilities but aren’t actively investigating mysteries and don’t need investigative abilities. In most genres, important antagonists don’t need Stability scores — though text defining their areas of knowledge and mental states could be very useful.

These distinctions can be counterintuitive, so don’t introduce them for their own sake. When necessary, though, swallow your aesthetic qualms and embrace them.

When designing new rules, the configuration of other rules is important but is not a starting point for your thought process. If you need to devise drowning rules, don’t start by looking at the falling rules and extrapolating from there. Ask yourself how drowning works in the material you’re emulating and go from there. You’ll want to eventually look at the falling rules to see how they match up, and if a previous designer followed the same assumptions you did. If they solved the same problem you tackled in a more elegant way, then go back and tinker. If your solution works better for your situation, stick with it.

As you design new kits, modularity may inspire you to swap out portions of the core rules for something that works better for the material you’re emulating. Trail Of Cthulhu might require a different way of tackling Stability. Many other theoretically possible investigative kits, from Scooby Doo to Agatha Christie, would dispense with it altogether.

GUMSHOE is meant to be a tool kit, from which GMs can mix and match add-on rules to create the settings they want. Encourage this mind-set by indicating what other sorts of investigative games your add-on might be good for. Be clear which add-ons are suitable only for your sub-genre, and which ones have broader applications.


When attempting to design systems that facilitate the GMs and emulate narrative structure, you may find it useful to consider creating an anti-rule.

Even gamers who think they know otherwise will over-rely on any rule you put in a game book. Any Call of Cthulhu player will tell you that a good GM doesn’t make you roll for the really important clues. Yet when we play a conventional investigative game of any type, we do have the players roll for clues all the time, because the rules provide for it. Like a gun on the table in the first act of a play, if you introduce a rule, it’s going to go off. GMs who know better will use it anyhow, out of reflex. GMs who don’t know better will cause countless hours of bad entertainment.

An anti-rule is a rule that exists purely to prevent the GM and players from doing this. It looks like a rule and walks like a rule, but really its main function is psychological. It gives gamers the comfortable feeling that there’s a rule guiding their behavior, giving them permission to engage in organically creative play. Like a rule, it provides structure, but unlike a rule, it doesn’t determine what happens in play.

The investigative rules of GUMSHOE are a prime example. The entire rules structure exists to prevent you from rolling against an ability to get a clue. It’s a rule to tell you you don’t need rules. The point-spending for evocative but nonessential clues adds a comfortable and satisfying gamey element to the experience. It allows you to use a rule now and again, but safely, so that the rules don’t get in the way and spoil everything.

Scenario Note: GMCs Making Rolls

Something I should have thought about sooner: whenever possible, it’s best to take situations in which a supporting character makes a roll and turn it around to one where the PCs make a roll against a difficulty. [Ed: We call this approach “player-facing.”]

It’s not so much an issue in combat and physical situations, where both PCs and GMCs typically have about enough points to last through one confrontation. But in situations like perception, PCs and GMCs are not really congruent. PCs have to space out their point spending through an entire adventure. GMCs are usually there for only a scene or two, and so can spend huge chunks of points on a roll. Having the GM making this tactical decision for them suddenly puts her in an adversarial situation that doesn’t really gibe with the spirit of the game.

In some cases you won’t be able to get around it, but whenever possible, turn these situations around. Instead of having the GMC roll Sense Trouble to see through an impersonation, set out a condition which, if the player makes the wrong choice, triggers her suspicions. Instead of having the GMC roll to search the PCs and find weapons, have the PCs make an Infiltrate roll to hide them so well the frisker doesn’t find them. This is not only fairer to the players but makes them more active participants in their own adventures.

Thought Process

In conclusion, when confronted with a rules problem, ask yourself the following questions, in the following order:

  • How does it work in the source material?
  • Is there a way to do it as a technique, and not a rule?
  • If I need a rule (or anti-rule), how simple can I make it?

Having already worked out the narrative consequences of the action I’m trying to model, have other designers already tackled similar problems in a way I’ll find instructive?

Do I label it as a universal add-on, or specific to this sub-genre?

Eternal LiesTell Me Lies, Tell me Sweet Little Lies

Converting Pelgrane’s ‘Eternal Lies’ Campaign to Call of Cthulhu

By Andrew Nicholson

Part One: Fools Rush In

It started as a way to persuade a group of gamers to try Trail of Cthulhu instead of their more usual fare.

“Hey”, I said, “…it turns out Pelgrane are looking for people to playtest a new campaign they’re hoping to publish – wouldn’t it be cool to see it first?”

There were nods and grunts around the table, accompanied by the supping of cola and devouring of curry … and thus it was decided.

I was more than ready for something a little different. I’ve been a keen Cthulhu Keeper ever since picking CoC up in about 1984, and after two years of only getting to run the occasional scenario, I was seriously looking forward to running a dark horror campaign.

A short while later (and a name drop or two) I had in my hands the playtest documents for Act One of Eternal Lies… and that was it, I was hooked.

Fast forward a few years, and Pelgrane announced the campaign was ready for publishing.  People all over the internet were asking about it, and playtesters were enthusing about what a great campaign it was… but there was one repeated question:

“Does it have Call of Cthulhu statistics? If not, will there be a Call of Cthulhu conversion?”

Pelgrane’s Trail of Cthulhu range is, without doubt, excellent. The GUMSHOE system has many fans. However, much like many family relationships, there are those who will always prefer its older and better known brother. Some do not like GUMSHOE’s spend system. Others are fond of ToC – but just plain love CoC.

Having enjoyed the first act of the campaign so much, I felt it was a huge shame people might miss out on it. So, looking at the playtest notes for the first act, I did a rough conversion on scrap paper and thought: “I can do this – I know both systems pretty well, I’ve got the experience, and it shouldn’t take long either”

A couple of conversations with Simon at Pelgrane later, and we had it agreed. I would write a conversion of the campaign, and we hoped to have it ready in time for release at the same time as the book.

And then the files of the campaign arrived. I devoured them greedily, and it slowly began to sink in.  Act One was only the introductory act. Act Two was big. Very Big. I also realised the way I had originally intended to write the conversion notes just wouldn’t do; it would have been fine for those with a passing familiarity with GUMSHOE, but would not give enough support to those who had never played it.

Also, we only had a few months to do it. My conversions would not only need working out and writing, but some playtesting to make sure they weren’t completely unreasonable.

In a word: Eeek.

But there was no way I was not going to do this. I had given my word, after all – plus, was I going to let an opportunity to do something cool for the Pelgrane slip away?  Hell, no.

The First Decision : Keeping the Flavour

I read through the system conversion ideas in the back of the Trail of Cthulhu rulebook.

I sat down with a cup of tea, and gave it some thought.

I decided that Rule number 1 was: Preserve the flavour of the campaign’s encounters where at all possible.

By this I mean that players using either system would get as close to an identical experience playing the narrative of the campaign as could be managed. Tough encounters should remain tough. Easy encounters should remain easy. Some encounters should require sacrifice of irreplaceable resources. Investigative encounters should require the players to think, but not necessarily be dependent on random dice rolls.

Trail of Cthulhu approaches investigations slightly differently from Call of Cthulhu. In particular, there were couple of distinct tenets that I wanted to preserve: “Make the game player focussed”, and “It’s not rolling dice to find core clues that is important – it’s what investigators do with them”.

Player Focussed

Player focussed games aim involve the players as much as possible whenever anything happens. The vast majority of the dice rolls are made by the players, and even in situations where they are opposed by NPCs the Keeper rarely rolls a die.

The theory is this keeps the players feeling involved and also means that if something happens they feel they had some input into it. There are few things worse for a player than to feel that something nasty happens “automatically” to them without them getting to at least roll dice. It may be a mainly psychological reaction, but it’s one I’ve seen time and time again.

So, to preserve this, I approached most situations in the conversion in the same way. When potential conflicts arise like an investigator hiding from an NPC, rather than call for opposed rolls (something that Call of Cthulhu doesn’t have the best system for anyway), I would put the onus on the player rolling against their skill, with a negative modifier to account for the skill of their opponent. This was a quick and easy conversion to do, still fits the Call of Cthulhu system – and I could still list the NPCs skills for those Keepers who wanted to go with the opposed roll approach.

The exception was in combat. While it was easy to make non-combat situations player focussed,  combat would require far too many spot rules, and would not feel like Call of Cthulhu.  Combat, therefore, shouldn’t be messed too much with.

Core Clues

Trail of Cthulhu also uses the concept of core clues – clues that are so vital to the investigation that without them the investigation will stall. Its been a long accepted flaw that in some Call of Cthulhu scenarios, one failed roll can mean the entire adventure stalls to a halt unless the Keeper arranges for some sort of work around.

One of the things I like about Trail of Cthulhu scenarios it clearly identifies these clues, and Eternal Lies is no exception. However, the question becomes – how do we approach this in Call of Cthulhu?

I looked at the suggestions in the main Trail of Cthulhu rulebook. They seemed serviceable… but I thought I could do better.

I had been using a house rule in my Call of Cthulhu games for several years, derived from Trail’s core clue concept. It had been playtested on numerous occasions, and seemed pretty effective. An investigator still has to roll against a skill to find a clue, but failure on the dice doesn’t mean failure to find the clue; it means that extracting or deciphering the clue becomes more difficult – the exact problem being decided by the Keeper based on the circumstances. I find this works very well, and keeps the acquisition of clues interesting for the players while still keeping the investigation on track.

However, while I knew this would be an approach many Keepers would be happy with, I also needed to account for those who are less comfortable with improvising, or prefer either more rules orientated approach. So, we playtested some alternatives solutions, and a couple of them (along with the rulebook’s suggested approach) are discussed in the conversion.

Having dealt with the more general rules issues, now I needed to start looking at specifics…

(end of part one – part 2 will follow next month!)