Go to Part 1

Now we have a basic framework for the urbancrawl: A map divided into districts. Content keyed to each district.

What we need now is a default action that will allow the PCs to engage with that keyed content. In dungeoncrawls and hexcrawls, that action is geographic movement. In the urbancrawl, it’s the investigation action.

In my discussion of hexcrawls, I advocated for the “hidden hex”: The hex is an abstraction that’s useful to the GM for keying and managing content, but which has no meaning for the PCs. In the urbancrawl, however, I believe that you will generally be keying content to districts that are meaningful to the PCs: They know where that ward or neighborhood or landmark or street is located.

So the basic investigation action ends up being pretty simple and very transparent: The player points to a neighborhood and says, “I want to investigate there.” The GM looks at their key and tells them what they’ve found. (If the GM is using multiple urbancrawl layers, then they can randomize which layer has been discovered with this investigation action.)

GENERIC vs. SPECIFIC INVESTIGATION

In actual practice, though, you’re quickly going to want to add additional utility to this basic action.

The first thing you’ll need is the ability to distinguish between generic investigations and specific investigations.

In a generic investigation, the PCs are just looking for anything interesting without any particular agenda. This generic investigation is identical to our basic interaction above.

In a specific investigation, on the other hand, the PCs are poking around with a particular goal in mind. This is where the urbancrawl layers become significant: They’re not just looking for anything of interest, they’re specifically looking for a patron. So instead of randomly selecting content from your available layers, you’ll key the specific content tied to the patron level of your urbancrawl.

(You might also include a chance of them finding something other than what they were looking for. For example, you might roll 1d6 and on a roll of 1 they get information from an urbancrawl layer other than the one they intended.)

I’ll also argue that this is the point where you should start hiding the abstraction again: You’ve arranged the content of your city into urbancrawl layers because that’s a convenient way of organizing it and interacting with it as a GM, but those layers don’t really have “meaning” in the game world. The players shouldn’t be saying things like, “I’d like some content from the Heist urbancrawl layer, please.” You want them to be saying, “We need another big score real bad. I’m going to go hit up my contacts in Oldtown and see if I can find something lucrative.”

(At that point you might look at your key and note not only the Tablets of Shandrala that are being held by the local Sheriff of Taxes (as keyed to the Heist layer of your urbancrawl), but also that Count Ormu on the Vampire layer of your urbancrawl in the same district holds the onyx crown jewels of the Lich Queen of Rasang.)

At this point, you might also be thinking about prepping a random table for selecting specific districts if the PCs attempt city-wide investigations. That’s certainly an option and I’m not going to say thee nay, but I suspect you’ll lose some feeling of the ‘crawl if you do that: Instead of crawling to a specific part of town and poking around (exploring the town and everything it has to offer as they do so), the players are just kind of generically asking for a content handout.

(For similar reasons, I wouldn’t run a dungeoncrawl by randomizing which room they explored next. This is why I think making the district structure of the urbancrawl explicit is the right way to go: It encourages that specific engagement with the geography of the city and requires that meaningful exploration choices are being made. On the other hand, I could certainly be completely wrong about this.)

MECHANICS OF THE INVESTIGATION ACTION

Instead of having the investigation action result in an automatic success, you might want to resolve it mechanically. Obviously the exact nature of the mechanic will depend on the system you’re using, but for the sake of argument let’s assume we’re using D&D 3.5.

The most obvious mechanic here would be a Gather Information check. You could set a universal DC for the check (DC 15 to perform an urbancrawl investigation); or you could vary it by city (DC 15 in the City-State of the Invincible Overlord, but DC 20 in the more tight-lipped City-State of the World Emperor); or you could vary it by urbancrawl layer (DC 10 for the Random Encounter layer, DC 20 for the Vampire layer); or you could define it for each key entry (DC 15 to find the blood den in Midtown, but DC 22 to find the den in the Nobles’ Quarter).

You could also allow for exceptional successes to generate additional information. If you were using a One Roll Engine system like Better Angels or Reign, for example, you might generate “soft rumors” for each additional set after the first. These soft rumors wouldn’t give the PCs a specific hook, but they would tell them about general areas of interest that are available in the local urbancrawl. (For example, you might tell them about rumors circulating through the Temple District about people showing up with puncture marks on their necks. They won’t find a specific blood den that way, but it might prompt them to go looking for it.)

You could similarly use the height of an ORE set to determine if the PCs get additional pieces of actionable intelligence. For example, any simple success tells them where the vampire blood den is. But if they score additional dice in the set you might tell them how many vampires are nesting there or give them a floorplan. (This could also just translate into additional rolls on an old school rumor table about the location.)

CONTEXTUALIZING

Starting off with generic Gather Information checks is probably a good starting point, but I’m guessing it won’t take long for the PCs to start attempting investigatory actions that would be handled better through alternative skill checks.

If we think back to our survey of old school city supplements, we might recall that Pavis had some interesting guidelines for PCs researching information in the city.

Gather Information checks can be used to handle the collection of gossips and rumors from taverns and markets.

Knowledge skills could be used when searching through records and the like. Pavis suggests the records of cults and guilds (diaries, receipts, letters, ledgers). It might also represent tax records or newssheets or libraries or any number of other things depending on the particular setting. Gaining access to these records might require Bluff, Diplomacy, and/or the paying of a fee or bribe.

Knowledge (Local) could be used to generate “soft rumors” like the ones discussed above (suggesting potential avenues of investigation).

Specific types of urbancrawl layers might also suggest other skills. For example, maybe an Appraise check would make the best fit for the hypothetical Heist layer I keep using for examples. Players are also likely to suggest all kinds of specific hijinks that could trigger other skill checks (like a Forgery check to gain access to property records or a Decipher Script check to figure out the graffiti patterns the local gangs are using).

The point here is that, like any action resolution, you want to contextualize the investigation action and you also want to respond to the contextualization provided by your players: Don’t just tell them “the vampire den is on Highborn Street”, tell them how they learned that piece of information.

A good technique for this is to make the skill check and then (assuming success) frame the scene just before the information is acquired. For example, if they make a Gather Information check to figure out where Don Carlo is holed up don’t just tell them, “You talk to Don Carlo’s driver and he tells you that he’s got safehouse on the east side.” Instead, cut to the point where they walk into a seedy bar and find the driver half drunk. That lets you play through the moment, which can set the mood, allow for memorable roleplaying opportunities, or lead the investigation in unexpected directions.

Go to Part 12: Exploring the Advanced Urbancrawl

Since we’re talking about urbancrawls lately, my attention was immediately arrested when I saw a couple of pretty amazing city generators over at Last Gasp.

Both of them are dice-drop generators, which reminds me of the fun you can have with Vornheim, but they’re integrated with prodigious random tables. The first is a generator for small, Lovecraftian-style fantasy towns: The dice drop generates a layout for the buildings and the faces of the dice are used iteratively to form factions, feuds, leadership… It’s fantastic. Check it out here.

The other generator creates a particular iteration of the ever-shifting magical city of Cörpathium, which draws its inspiration from M. John Harrison’s Viriconium. The dice drop in this generator determines the relative locations of city districts and then (this is the cool part) generates a whole bunch of conditionals based on which districts are present and how they relate to each other. You can find it here.

Go to Part 1

The pair of questions I keep coming back to are: Why are the PCs urbancrawling? And what are they actually doing when they “crawl”?

And I think the reason I’m struggling with those questions is because, in an urban environment, their answers are very dependent on the specific context of the PCs’ actions.

This dependence is the result of the city being an extremely dynamic place: In a dungeon there’s generally just one interesting thing happening in a room. In the wilderness the interesting thing is separated from other interesting things by miles or leagues of scenery. But in the city there’s so much activity so densely packed that any given block (or even building) will often have dozens of different things happening in it. The question of which of those things you’ll end up engaging with is highly dependent on the experiences that you choose to seek out.

Having recognized this Gordian knot, we now have to seek the sword that can slice through it. And I think the key here is to stop thinking of the city as a monolithic entity and start thinking of it as being made up of diverse parts. We need to manage the dynamic nature of the city by breaking it apart into distinct layers. The city is not a single urbancrawl, but rather a multitude of urbancrawls that lie on top of each other in a simultaneous coexistence.

LAYER 0: THE GAZETTEER AND THE MAP

Let’s start with the base of the city: The gazetteer. This is the Baedaeker’s travel guide version of the city. It’s the list of useful shops, taverns, inns, and Important Public Locations.

The gazetteer isn’t an urbancrawl. Although it might be interesting to build some of this stuff into an “Explore the City” urbancrawl layer for those completely new to the city, the stuff in the gazetteer constitutes the elements of the city that will generally be visited through targeted travel. For ease of reference, putting these locations in a gazetteer format makes the most sense.

The other thing you’re going to want forming the foundation of your city is the map. And you’re going to want to split that map up into naturalistic divisions. For ease of reference, I’ll refer to them as districts, but in the game world they could be anything: Neighborhoods, wards, sectors, gang territory, streets, or whatever. You’re aiming for districts that make sense to the characters actually living in the city (they’re labels or divisions that they would recognize and talk about). But you’re also aiming for a districting concept that scales to the amount of material you’re planning to include in each layer of the urbancrawl. (This is just like a hexcrawl: If you find yourself frequently keying multiple entries into a single district, your scale is probably too large. If you find yourself with a lot of empty districts, your scale is probably too small.)

THE URBANCRAWL LAYERS

Over the top of this foundation – the gazetteer and the map – you’re going to layer in your urbancrawls.

Unlike a dungeoncrawl, the goal of an urbancrawl doesn’t default to treasure hunting. It defaults to finding something interesting. If the PCs are fairly ignorant of what the city has to offer (or are simply looking for new opportunities), then this idea of “finding something interesting” can remain fairly generic. But as the PCs learn more about the city, the action will inevitably become contextualized: Instead of saying, “Let’s see what’s going on in the Longbotttom neighborhood.” they’ll start saying things like, “We need to find out if there’s a blood den near Powderhorn Park.” or “Maybe we can figure out what the Halfling Mafia has been up to.”

Each urbancrawl layer basically boils down to one way in which the content of the city can be contextualized. In a given city, for example, you might have separate urbancrawl levels for:

• Vampire blood dens.
• Patrons who can give them jobs.
• The activities of a criminal gang.
• Potential targets for lucrative heists.
• Purely random encounters that provide “color”.

Each of these urbancrawls would (ideally) have interesting material keyed in every district of the city. So if the PCs go poking around the Longbottom neighborhood they might find the local vampire den. Or get contacted by agents of Lord Melbourne. Or run into mafiosos hassling local businesses. Or discover that a local merchant family currently holds the Neferelli Diamond. Or get their pockets picked by goblin urchins.

(Having five full layers like this would probably represent a really dense urbancrawl. It would require a lot of prep, but it would also deliver hundreds of hours of play. My guess, though, is that you probably only need 2 or 3 layers to get a really dynamic urbancrawl started.)

THE THIRD DIMENSION

This basic structure of urbancrawl layers is probably sufficient for running a simple urbancrawl. But I’m going to propose that you can add significant depth to your city by extending its urbancrawl layers into a third dimension.

You’re going to take one of your existing urbancrawls and you’re going to add layers to it. These deeper layers won’t necessarily be complete (in the sense that they’ll fill every district in the city with content), so it may be more convenient to think of them as “hidden nodes”.

The idea is that these hidden nodes can’t be directly or immediately accessed by anybody ‘crawling the city. Instead, they can generally only be accessed in one of two ways:

First, districts in the “lower levels” of the urbancrawl may contain clues that will point directly at these hidden nodes. For example, PCs raiding a vampire blood den may discover correspondence from Count Ormu implicating him as a vampire lord.

Second, these hidden nodes can be “exposed” to people ‘crawling the city if certain conditions are met. These conditions would generally take the form of “clearing” lower level nodes. For example, if various adventuring parties take out three of the blood dens in the city, Count Ormu’s network may be sufficiently disrupted to expose his involvement.

This second condition is particularly important for open tables because it solves the “I didn’t get the clues from the first half of this mystery” problem: If you’ve got the clues pointing at Count Ormu, great. If you don’t, but the blood den networks have been sufficiently disrupted by other groups, then Count Ormu becomes available to you through general ‘crawling.

In conceptualizing these hidden nodes, it may be useful to reflect once again on Kenneth Hite’s Conspyramid from Night’s Black Agents, which provided the most immediate inspiration for this added dimension:

You don’t necessarily need to engage in the same rigid hierarchy or chains of communication (the geographic component of the urbancrawl will cover a lot of the same bases), but it also can’t hurt, right?

Of course, the third dimension of some urbancrawls will be more conceptual rather than organizational: Different targets for heists, for example, may not be directly connected, but you can still add additional levels to a heist-based urbancrawl layer (representing the attention of more powerful clients or security arrangements which have been exposed or simply a pacing mechanic for heists over time).

Go to Part 11: The Investigation Action

Go to Part 1

By 1982, urban RPG supplements had pretty much universally transitioned to become narrative-backdrop travel guides: The modern gazetteer format that generally features a history of the city, a description of notable locations, and a cast of important NPCs. (Vestigial rumor tables hung around here and there for a few more years, but generally faded away until the OSR began bringing them back into vogue.)

Which is not, of course, to say that there aren’t some truly fantastic city supplements. I actually ended up surveying a lot of great stuff while researching these posts: The City of Greyhawk boxed set, the truly prodigious combination of FR1 Waterdeep and the North with the Forgotten Realms: City System Boxed Set, Greg Stolze’s City of Lies for L5R, Monte Cook’s Ptolus, Chicago by Night, City of Freeport, and so forth. It’s just that they’re being designed for a narrative-based game structure that’s not particularly illuminating when it comes to urbancrawling.

Recently, however, we’re starting to see a resurgence in games that are willing to get a little experimental with their game structures. (This has been particularly true among STGs, but it’s also happening with RPGs.) The result has been a handful of “new school” urbancrawls.

DRESDEN FILES

When I first started chattering about urbancrawls, a lot of people pointed me in the direction of the Dresden Files. This game has gotten a lot of buzz for its robust city-creation system and I was told it might be exactly what I was looking for.

Unfortunately, it’s not.

The city-creation system in Dresden Files is really a campaign creation system in which the creation of the city is tied to the creation of the PCs and the players share narrative responsibilities in defining the themes, threats, and locations which define the city. It’s a nifty approach (and I recommend checking it out), but the focus is still on creating a backdrop for narratives.

NIGHT’S BLACK AGENTS

Kenneth Hite’s Night’s Black Agents has an international focus, but it features two very clever systems for running conspiracy-based sandbox campaigns that I think may prove useful in our thinking about urbancrawls.

First, there’s the Conspyramid. When you’re prepping your campaign you draw up a Conspyramid with six levels of power: “Each ascending level has fewer, more important nodes.” So, at street-level power you’ve got six different nodes. Bump it up a couple levels to provincial powers and you’ve got four different nodes.

The Conspyramid is useful because it simultaneously shows the GM how the organization of the conspiracy works (in an abstract way) and how the PCs can investigate the conspiracy.

The Conspyramid is clever because Hite also ties it mechanically into the game mechanics: As the players fill out their adversary map (i.e., figuring out how the conspiracy hooks together), they gain dedicated pools of points to spend on ops targeting connected nodes on the Conspyramid. They can also use Human Terrain and Traffic Analysis skills to figure out the connections between a node they know and other nodes (i.e., generating leads).

That’s a default goal, a default reward, and a default action.

Hite then adds a second track in the form of the Vampyramid:

He describes this as an “escalating response algorithm” which provides the vampire conspiracy with a naturalistic response to the PCs: So the Conspyramid represents a largely static ‘crawl; the Vampyramid provides easy-to-manage active responses.

It’s the most innovative, creative, and gobsmackingly brilliant work I’ve seen on an RPG game structure in over a decade. Hite’s a genius and you should check it out.

(UPDATE: Hite has informed me that his own work on NBA was inspired by Elizabeth Sampat’s Blowback. I recommend checking that out, too.)

While these structures cannot be directly applied to the type of urbancrawling structure we’re looking for, where I think the NBA systems are extremely informative is the intersection of investigative “layers” combined with default, mechanically-driven investigative actions. (We’ll come back to this idea shortly.)

VORNHEIM

Vornheim by Zak S. proffers a quote about Moving vs. Crawling which is so extremely useful that I’m going to provide it here in full:

In a dungeon or wilderness adventure everything is hard – navigating, finding food, getting a decent night’s sleep, etc. – and so everything is part of the adventure. Adventuring in a city is different from adventuring in a dungeon or wilderness because cities are actually meant for habitation. In most cities, many things will be easy and therefore not part of the adventure and the GM has to do a great deal of deciding when to “zoom in” and deal with the situations in more detail. For this reason we’re going to create a distinction between simply “moving” through the city and “crawling” through it. (…)

“Crawling” occurs when:

• The PCs are being chased.
• The PCs are in a hurry.
• A large number of elements in the city are actively hostile to the PCs (such as during an invasion or plague of madness).
• The PCs are systematically searching a small area of the city for something.
• The PCs are trying to avoid running into someone or something.
• It’s night.
• The city is transformed in some way such that it ceases to function like a city (post-nuclear bomb, etc.).
• The PCs don’t really know where they’re going.
• There’s urgency attached to the PC’s decisions about how to proceed for any reason.

A lot of the Old School Renaissance has largely spent its time regurgitating the forms and content of the ‘70s and early-‘80s. (And, don’t get me wrong, produced a lot of good material doing it.) Vornheim is a prime example of the OSR being a little more daring, grounding itself in the old school material, and then innovating.

For ‘crawling, Vornheim creates a pair of simple structures: If you’re crawling from neighborhood to neighborhood (i.e., trying to traverse the city) you generate one random encounter per neighborhood. If you’re crawling within a neighborhood (i.e., they’re trying to find something in the neighborhood) he uses a method of rolling 2d10 and using:

• The relative position of the dice to determine where the goal of the ‘crawl is relative to the PCs.

• The number on the die to determine the layout of streets between them and their destination. (Literally. You can check out the diagrams of how this works here.)

It’s a very clever and quick system. Where it comes a little short is in providing structure for making the journey from Point A to Point B a meaningful/interesting one.

What makes Vornheim truly invaluable in any discussion about urbancrawling, however, is the plethora of incredibly cool, incredibly useful, and incredibly original tools that Zak has designed for procedural content generation. (This is something I talked about in my Fun With Vornheim series, which you should check out for some awesome examples of what the book is capable of.)

TECHNOIR

Technoir is another system I’ve talked about quite a bit. It’s got an incredibly clever resolution mechanic, but what makes the game truly exceptional are its plot-mapping mechanics.

The short version is that the game is built around “transmissions” which each describe a city of the future. Each transmission consists of six connections, six events, six factions, six locations, six objects, and six threats. These are organized into a 6 x 6 master grid which allows you to randomly generate elements and add them to your plot map.

By itself, that’s a nice little procedural generator. But Technoir takes it one step further by including explicit mechanics for how elements are added to the plot map, and the primary method is controlled by the PCs: Whenever they hit up one of their contacts for information, an element is generated based on the contact’s current relationship to the plot and it’s connected to the map based on situational mechanics as well.

(A lengthier example of using the Technoir system can be found here.)

The result is a robust improvisational structure which has the delightful property of allowing the GM to discover the “true conspiracy” of their ‘noir adventure at the same time that their players are investigating it. (It’s also a system which could be very easily translated to any genre or setting.)

In terms of urbancrawling, the key insight from Technoir is the ‘crawl action itself: Hitting up your contact.

What makes this notable is that this is not a decision about geographic navigation, but it nevertheless fulfills the same exploratory function. The only limitation is that this is a mystery-based structure and, as you’ve probably gotten sick of me saying, mysteries don’t work for open table play. Technoir solves the problem of being unable to solve the mystery if you missed the clues in the first half (by utilizing a structure which constantly manifests new clues), but you still have the problem of players experiencing the first half of a mystery and never getting the satisfaction of its solution. (But if you want an urbancrawling structure and you don’t need it to support open table play, then I enthusiastically recommend Technoir.)

Go to Part 10: One City, Many Urbancrawls

Go to Part 1

This will be a rather eclectic round-up of some additional old school city supplements I surveyed while doing my urbancrawl research.

BOOT HILL

The original Boot Hill was primarily a wargame for running gunfights in the Wild West with a light roleplaying element draped over the top of it. (This is probably because it was released in 1975 and people still hadn’t quite figured out the whole “roleplaying game” thing.)

It does contain some advice on setting up a town: “Each building used (excluding the small ones with half-moons cut into the doors) in a town must have a complete plan drawn for each of its floors, including any cellars or basements.”

Hmm… Probably not that useful.

EMPIRE OF THE PETAL THRONE

Also released in 1975, M.A.R. Barker’s Empire of the Petal Throne had a much firmer grasp on the whole “roleplaying game” thing. It also includes some intriguing guidelines for urban-based play (exactly the sort of stuff missing from OD&D and AD&D):

• “Initial Encounters in Jakalla”, which serves as a random generator for people looking to hire the PCs and the missions they want to hire them for.
• “Erecting and Buying Buildings”, which calls to mind Arneson’s Blackmoor as described in the First Fantasy Campaign.
• “Advertising”, which the City State of the Invincible Overlord was also providing guidelines for almost simultaneously. (It seems likely that this focus is the direct, organic response to actual players.)
• “This will require (a) a terrain map of the area, (b) a surface map of any city or cities to be developed, and (c) maps of the Underworlds of those cities. A detailed terrain map of a single hexagon can also be drawn to scale, if the details of a particular region are required (e.g., for a military campaign).”

THIEVES’ WORLD

The Thieves’ World boxed set was released in 1981 as a generic supplement based on the series of fantasy anthologies and featuring stat blocks for a couple dozen different RPGs. Let me lead off with a quote from designers Yurek Chodak, Steve Perrin, Greg Stafford, and Lynn Willis:

While fighting, arguing, and other sorts of adventuring are important and have their place, all adventures conclude, and even the greatest adventurer goes home to spend his profits. If he is like most of us, he will try to impress others with his success. Those other people are found in cities, as are rooms, markets, taverns, libraries, and gyms.

The city is the natural home of every adventurer. The ebb and flow of city life gives opportunity for every type of character. A city is the only environment proper to a full-time thief. Only in cities do many strangers meet, and only in cities is information plentiful. The bigger the city, the more readily are exotic items found or sold.

Earlier I described the early Judges Guild’s approach to cities as a mish-mash of color, information gathering, and carousing. This quote seems to sum it up elegantly.

Also of interest is this quote:

Contrary to most fantasy role-playing games, the gamemaster should not try to run too many players at one time while running a city. We find a mix of 1-2 players to be optimal, with 4 being a very real maximum even with experienced players. If more players are run, someone will always be bored, since the G.M. can only interact with one player at a time. Conversely, the number of characters is not particularly important, as only one character can be active per player at any given moment.

The notable assumption here is that city-based adventures will typically feature characters acting independently instead of collectively as a group.

In terms of practical material, what interests us particularly in Thieves’ World is the rich and robust encounter system. The tables are designed to be used every 10-15 minutes (a timeframe which appears to refer to the real world and not the game world).

A significant number of encounters start with “accidentally bumps…” This represents the jostling the characters would receive in any busy street and helps to prevent thieves from becoming too obvious.

The encounter tables are then broken down by neighborhood and some specific regions within each neighborhood are given specialized tables. The exact procedures for generating encounters are not particularly notable, but the sheer scope of the system is an escalation of the old Judges Guild methods (which the Judges Guild was simultaneously abandoning with City State of Tarantis).

Also of note are the Business Generation rules, which the GM can use to randomly stock the largely unkeyed map of the city.

PAVIS

Released in 1982, the overall approach of the Pavis boxed set for RuneQuest is purely narrative: It’s explicitly designed for use in an “episodic campaign” and it includes a 64-page book which consists entirely of adventure scenarios for you to build your campaign around.

Despite this, Pavis does include a couple of interesting resources for use in urban play.

First, there are guidelines given for seeking long-, medium-, and short-term employment. These caught my eye in part because they presented a unique angle on expectations in an urban campaign, but also because they emphasized the ability to handle time in the campaign on different scales. (Something which can be very useful in managing open tables.)

Second, some really interesting guidelines are given for different ways in which PCs can research information in Pavis:

• Collecting gossip and rumors from taverns and markets.
• Searching the records of appropriate cults and guilds (diaries, receipts, letters, ledgers).
• Hiring a sage.
• Paying a fee to gain access to a library.

It’s interesting to note, in general, that these elements – which would have probably received mechanical sub-systems just a couple of years earlier – are instead rendered as merely guidelines (except for the hiring of sages).

Go to Part 9: New School Urbancrawls