How do you build a campaign? Can you generate all or most of it on the fly? If you play with whoever shows up in a casual, open table style, then how do they get started on an adventure and how do extended adventures get resolved by different groups? What is the nature of “the truth” that you build for your players to adventure in? Why would it matter if you just move things around in that “truth” if the players don’t know about those things yet?

The Start of the Journey

I began role playing with Dungeons and Dragons original white box in 1974. The first thing I did was to make a dungeon, populate it with monsters and treasure, tricks and traps, and let the players have at it.

After a short while that type of play became a bit stale and we started playing Metamorphosis Alpha, where the characters are primitives on a vast colony starship that is out of control in deep space, full of mutated creatures and techno treasure. Structurally it differed from my dungeons in that there was an overall mystery to explore and a potential goal to work on.

As time went by the commercial side of RPGs became more adventure story based, with a plot for the players to follow and a goal that would signal the end of the campaign when reached.

In 2012 Justin Alexander began writing his articles on Game Structures and they clarified and sharpened a lot of my campaign and adventure designing process.

The articles begin with:

One of the most overlooked aspects in the design and play of traditional roleplaying games is the underlying game structure. Or, to put it another way, there are two questions which every game designer and GM must ask themselves:

(1) What do the characters do?

(2) How do the players do it?

These questions might seem deceptively simple, but the answers are complex. And getting the right answers is absolutely critical to having a successful gaming session.

Early on he looks at Dungeon Crawls, like my first efforts in ’74. These turn out to be very simple game structures indeed, allowing beginner game masters to get a feel for how to do it without making mistakes very often. You have a series of locations, each sufficient to itself with monsters, treasures, tricks and traps “firewalled” off from everything else. What happens in dungeon room 101 stays in dungeon room 101.

The players have a default goal when nothing else presents itself: fight monsters, get treasure.

The players also have a default action in a location when nothing else presents itself: choose an exit and take it.

That’s pretty simple stuff and it applies with modification to a hex crawl in the wilderness as well.

And then we get to cities.

And its suddenly not so simple.

In Thinking About Urban Crawls Justin acknowledges this and explores different approaches which I found quite exciting, and worked on a bit myself, but I ended up going a different way and discovering a “2^nd level” game structure that applies in all of “crawl space”, be it D&D, sci fi planet crawling like Traveller, weirdness like Paranoia or even a complex mystery setting like Call of Cthulhu.

What do we Want?

Before I look at that, let’s discuss why Justin wanted to work out a particular urban crawl structure.

If I can already use node-based structures to run urban-based scenarios, why am I interested in figuring out this “urbancrawl” thing?

Open game tables.

My current open table campaign started with dungeoncrawling. It later expanded to include a surrounding hexcrawl. In both cases, however, I had vestigial cities hanging out as “home bases” for the PCs: They were safe havens and places where they could resupply, but active adventuring wasn’t taking place there.

And it wasn’t taking place despite the fact that I had specifically prepped them the way I normally prep cities: With interesting NPCs and scenario hooks hanging all over the place. In non-open campaigns all of those hooks would get developed using node-based structures as the players explored them. But node-based structures are generally interdependent, specific, and non-holographic: When every week sees a different group of PCs sitting at my table, the node-based structures don’t work. They fall apart.

But if I could develop an urbancrawl scenario structure that works the same way the other ‘crawl structures work, then I would be able to prep effective material and the players would know how to engage it.

And I am in the same space here. I have been running a successful D&D Public Open Table since January this year. After 22 adventures the Kenmore Role Playing Society has 1 game world, 3 active DMs, 2 in training and 20 players. Thanks Justin, the ideas are great and they work! They work even better however when I saw my improvement to the crawl structure, which is closely related to, but not identical with, node based adventures.

So what is the game structure of which we speak? Here is Justin’s description:

What I’m looking for right now, though, is an urbancrawl. A scenario structure that would use the same fundamental principles that dungeoncrawls and hexcrawls use.

Let’s take a moment to review the characteristics of a ‘crawl (based on our analysis of the dungeoncrawl and hexcrawl):

1. It uses a map with keyed locations. (This provides a straight-forward prep structure.)
2. Characters transition between keyed locations through simple, geographic movement. (This provides a default action and makes it easy to prep robust scenarios.)
3. There’s an exploration-based default goal. (This motivates player engagement with the material and also synchronizes with the geographic-based navigation through the scenario.)
4. Characters can engage, disengage, and re-engage with the scenario. (You can go into a dungeon, fight stuff for awhile, leave, and when you come back the dungeon will still be there.)

This fourth property appears to exist because:

(A) Material within the ‘crawl structure is firewalled. (In general, area 20 of a dungeon isn’t dependent on area 5.)

(B) The default goal is holographic. (You can explore some of the wilderness or get some of the treasure and still feel like you’ve accomplished something.)

(C) The default goal is non-specific. (You can get a bunch of treasure from Dungeon A then get more treasure from Dungeon B and still be accomplishing your goal of Getting Lots of Treasure.)

(D) The default goal isn’t interdependent. (You can clear the first half of a dungeon and somebody else can clear the second half. By contrast, you can’t solve the second half of a mystery unless you’ve got the clues from the first half.)

The dungeoncrawl structure provides these features for location-based adventures. The hexcrawl structure provides these features for wilderness-based adventures. A fully functional urbancrawl structure would theoretically provide these features for city-based adventures.

It’s point A that I have decided to alter ( Material within the ‘crawl structure is firewalled.) and, I think, for sound reasons. And I think that doing so does not harm the essential “crawl” nature of the game structure.

To understand why I think we need to drop the firewalls we need to look at the problem of…

The Quantum Ogre

Back in 2011 the article On How an Illusion Can Rob Your Game of Fun was published on the Hack and Slash blog. It talked about a scenario where players could choose to go to woods grove A, or woods grove B. The game master has an ogre she wants them to encounter so whichever grove they choose, hey presto that’s where the Ogre is. The article coins the phrase “The Quantum Ogre” for this scenario.

Now there is a lot than can be said about Quantum Ogres, but I’ll leave most of that for some future article. What concerns us here is this:

Player choice has been thwarted, because the players were presented with a meaningless choice.

Now that is interesting. In a firewalled dungeon the default action is “pick a direction and go there”. But if the dungeon is firewalled, as specified, the players are making meaningless choices in the Quantum Ogre sense. They have no information with which to assign value to one exit or another.

In Justin’s article on Art of Rulings – Part 5: Skill and Difficulty he mentions meaningful choice as being important, and it seems to me vitally important when players are choosing a direction to go in.

What this means is that for any location with multiple exits the players should be given some information that would allow them to differentiate one choice from another. This means poking a hole in the “firewall” of at least one location somewhere, so that players know they want to get to that location without necessarily having been there, and that “a way to get to it” applies to at least one of the exits from the current location. That’s a minimum requirement. Locations that provide no choice value among multiple exits I am calling “dead choice locations”.

The location with a leaky firewall is a form of node in accordance with Justin Alexander’s node based scenario design concept. However “nodes” are essentially a special case, in that they are a meaningful part of an extended adventure. My leaky firewall location may be a node, or it may be an incidental place the players have an interest in, say somewhere to have a beer. So I am calling these locations “Attractions”, like the rides at a fun fair. A node is a subclass of attraction. A node is an essential part of an adventure, while attractions in general are any location that players may be attracted to, and that gives movement in a particular direction meaning.

A City is not a Dungeon

So lets look at cities in contrast to the classic dungeon crawl. The first thing of note is that a city is a service structure for its inhabitants. It provides a number of services, food, shelter, supplies, work, social interaction, information, security and so on. When players enter a city, town or village they know that some of these services will be available in some form and, knowing nothing else about the location, will want to use at least one service or pass through.

Now, unlike a dungeon, a city is also generally information rich about its basic services and how to get serviced. Each service is an attraction. You want some plate mail? Go to Bill the Smithy in Griffon street, you can get there from here by going down witchcraft lane about two blocks. Griffon street is on your right. You need a hospital? Over in the Derm district, near the space elevator pylon you can see over there.

What else is different in a city? Well, a city is full of private spaces where only a restricted set of persons are authorised to enter. The city will have systems in place, to varying degrees, to protect that privacy. This is a severe obstacle to geographic crawling. You can’t just traipse through people’s houses or the administrative offices of the intergalactic bank as part of your crawl.

So, what if you have keyed an advanced cybernetic organism in the top floor of Intergalactic Bank inc that is holding the secret passwords to 50 independent treasure vaults scattered about the planet? How do the players get to crawl about and bump into this encounter? The answer is that you must change it from a firewalled location encounter into an attraction. The easy way to do that is to place an adventure hook into the public space that players will pass through on their way to the forum. Once an adventure hook gives them a reason to enter a particular part of the city’s private space then they will come up with ways to deal with that obstacle in order to get to the main attraction.

Working through public spaces to get to public attractions gets the players to encounter entry points for attractions in the private space. And as you move into private space you may encounter hooks to other private space attractions.

This differs from the undirected exploration concept. The default goal of “exploration for its own sake” is actually an artificial artefact of the classic dungeon. It’s a feature of board games like Talisman. It doesn’t really work in an extended play situation, even if its a sandbox crawlspace. In the dungeon the players know, without being told, that the dungeon is somehow full of monsters and treasure and that by wandering around you will bump into them and have some fun. To extend this to the wilderness, a city or interplanetary space doesn’t really work. Players need a reason to move from A to B. In a sense they are not so much “geography crawling” but “attraction crawling”.

A Revised Crawl Structure

So here is a revision of the crawl structure that Justin defined:

1. It uses a map with keyed locations. (This provides a straight-forward prep structure.)
2. Characters transition between keyed locations through simple, geographic movement. (This provides a default action and makes it easy to prep robust scenarios.)
3. There’s a find-the-attraction-based default goal. (This motivates player engagement with the material and also synchronizes with the geographic-based navigation through the scenario.) There are perennial service attractions, and harder to find adventure attractions.
4. Characters can engage, disengage, and re-engage with the scenario. (You can go into a dungeon, fight stuff for awhile, leave, and when you come back the dungeon will still be there.)
5. The impact of play on GM prep for the next session is minimal

The fourth and fifth properties appear to exist because:

(A) Material within the ‘crawl structure is independent, in that an attraction and other keyed elements remain present even if some other attraction or element is removed from play. (In general, area 20 of a dungeon isn’t dependent on area 5.) (NB There is a question here of “dangling attractions”, where an attraction hook is no longer present. I’ll discuss this in a later article)

(B) The default goal is holographic. (You can explore some of the wilderness attractions or get some of the treasure attractions and still feel like you’ve accomplished something.)

(C) The default goal is non-specific. (You can get ammunition supplies from the service attraction in The City of Mystery, and there learn about and get a bunch of treasure from the Dungeon A attraction on Planet X, resupply your ship from the service attraction at Starport Alpha and then get more treasure from the Dungeon B attraction on Planet Z, and still be accomplishing your goal of Getting Lots of Treasure.)

(D) The default goal isn’t interdependent. (You can clear the first half of a city adventure attraction set, and somebody else can clear the second half. I’ll discuss how this can work with multiple play groups digging into a mystery structure in a later article)

In future articles I will discuss different styles of attractions, I replace the general idea of “adventure hook” with the broader concept of “Projections”, how multiple open table play groups can dig into an “attraction mystery” using “player footprints”, and will discuss procedural generators for projecting keyed attractions, and also for creating new keyed attractions from random projections.