If you haven’t heard, a lot of folks in the RPG space are taking part in the #Dungeon23 challenge. Recently started by Sean Mccoy on Twitter the idea is that if you key a single room a day starting on January 1st in a dungeon complex you’ll have a detailed 365 room dungeon at the end of 2023.

It provides a framework for easy, gradual, but constant progress where you have a huge complete thing at the end of the year! There’s tons of other resources and variations on the idea folks have posted. There’s even a subreddit now! You should totally do it.

I’ve been laying the groundwork for my Cinderstrom volcano megadungeon and the surrounding wilderness sandbox of the Steaming Basin for around a year now and it’s got a lot left to go, mostly in specific keying of rooms. Doing Dungeon 23 should help with that.

I thought I’d share my own messy, slow, process so far in making a connected sandbox region/megadungeon for eventual publication at Silverarm and see if it’s helpful for anyone.

Joel’s Method for Megadungeon Sandbox Design So Far:

1. Pick Concept and Theme. Just write down the cool ideas that get you stoked in a not super organized way that you want in this megadungeon sandbox. This part is easy dopamine and the least work of any project for me but provides the loose core that the entire project will metastasize and pick up other influences over time.
2. (Read The Hobbit, Dying Earth, and Conan Red Nails story. Volcano megadungeon.  Andalusian Islamic architecture. Metro style podway and science fantasy. Abandoned dwarf hold sorta like Moria. Dragon with berserker cultists.) 
3.  Lightly key and name Wilderness Sandbox locations, deciding on scope (7×7 hexes with 49 locations of interest. Procedurally generate hex contents to speed the process up using Hex Describe, replacing bad ones with completly different ideas and tweaking and twisting and stealing from good ones to create new version of contents that fits the setting.
4. Rough level layout of the megadungeon using Scapple to create flow chart style map. (~300 rooms divided into  levels ringing central lava chamber of volcano in three tiers connected vertically and horizontally metro style ancient podway system that also connects to some other dungeons in sandbox outside)
5. Come up with lots of great  room concepts rapidly using this brainstorming method from GFC D&D.
6. Detail and name other dungeons/settlements around the region and establish size, lightly key whichever you have energy for. (12 other dungeons, average of 10 rooms each)
7. Detail the factions that are already in head with a more detailed treatment, decide how everyone feels about each other, where they’re based, and what their goals are. 
8. Create Sub Regional Encounter Tables and other Random Tables (Campaign events, merchants, rival NPCs)
9. Jaquays your wilderness sandbox for interesting navigation and meaningful choices because the wilderness is a dungeon. Create a variety of specific detailed paths between locations. 
10. Start Actually Playing with players in an unfinished setting! Build out based on organic play needs helping you create and tweak setting to be better.
11. Have three different games in the setting sporadically end and start over a year period.
12. Read too many OSR RPG blogs and books for good design ideas to bookmark and steal ideas and procrastinate on the hard part of actually writing the specific detail of every keyed location slowly.
13. Transfer everything from a huge Google Doc to OneNote because organizing different areas and topics into folders is easier to parse.
14. Settle on a dungeon level and room key format that provides a high level of usability at the table and provides a clear framework for work ahead. Acknowledge this will likely change some when I find a format I think is even better.
15.  Nail down the loose history already in head and then simplify the huge historical timeline to no more than 50 short events. These must be referenced in actual things the players engage with for organic discovery in the world Dark Souls style. All history must connect to actual play and be load bearing lore the players can interact with.
16.  Create World Anchors or themes and content that reoccurs throughout the sandbox for an organic sense of congruence and cohesiveness outside of a railroaded narrative. Plan that every dungeon or hex must have some link to one of these ~1d10 anchors to build these themes up. 
17. Pass back through Step 4 again and make things more cohesive with a new understanding of the setting’s context and history .
18. Plant 1d4 hooks in each hex or dungeon leading to other hexes or dungeons to better interconnect sandboxes. (Still not done, takes a lot of work)
19. Use a version of the Basic D&D Dungeon Room Stocking procedure but adapting to wilderness hex versions to add more stuff in a good ratio of types, continuing to treat the wilderness as a single overworld dungeon with other dungeons fractally inside. 
20. Decide roughly how much total gold there is in the sandbox (Lots) based on how many levels that would get the average party that found all of it (From Level 1 to at least Level 10 for a party of four Fighters in OSE). Once there’s this budget of total treasure, distribute it around the setting with more of it being found in higher level places of danger. Figure out how many significant magic items will be in the sandbox (At least 46 in the Cinderstrom megadungeon, 23 elsewhere.) Distribute these across the various dungeons)
21. Get sucked deep into Metroidvania and Zelda style game design for the megadungeon and setting and creating a series of soft “locks” for accessing difficult and lucrative levels and areas  and discovered “keys” in other locations in dungeon or sandbox that would help bypass these (Keeping in mind player ingenuity allows for tons of other ways to sequence break through locks and limits of video game design applied to tactical infinity of the tabletop rpg sandbox). 
22. Tweak things more
23. Today. I have~420 total rooms to key in the megadungeon and surrounding sandbox. So far I’ve at least named and lightly keyed ~100. but only have a handful of detailed keys.

How I Plan on Using the Dungeon 23 Challenge:

I have my scope already established which helps me see that there is light coming at the end of the tunnel. Starting in January I’ll attempt to key at least a single room for every day of the week. However, knowing my brain and the way I work I think it’s most likely I’ll achieve these 7 rooms in a single day or two over the course of the week and do nothing towards it on most days. I read about someone starting an accountability Discord server for folks working on Dungeon 23 and I think that’s a great idea. It’s like having gym buddies to keep you making progress except with more hunching over a laptop looking at a screen.

The specific granular work of detailed keying and refining of every single room and wilderness location is never the fun part for me. It is, however, incredibly necessary to produce a megadungeon sandbox worth playing for others instead of a collection of cool ideas on a blog. Feeling inspired and producing neat ideas is awesome but it’s being willing to commit to the grind of implementing these things in a playable form instead of tweaking format and high level design that will let me publish a megadungeon in a yearish instead of have a series of cool blog posts about the design I would use in a hypothetical awesome sandbox that I never made.

Do you plan on participating in Dungeon 23?

When one tugs at a single thing in nature, he finds it attached to the rest of the world.

John Muir

In order for a sandbox RPG region to have a sense of verisimilitude it’s important for many of the locations in the area to be actually connected to each other instead of existing as a series of unrelated vignettes. The dynamism of considering how two areas affect each other creates more immersive and creative settings that increase conceptual density and packing more interesting ideas into the available space of your sandbox region.

One way to achieve this interconnection creates a web of hooks that pulls players from location to location and unites the patchwork quilt of locations into a unified whole. In game design terms all rumors, jobs, quests, and lore are hooks that exist to help create a call to adventure by telling the player something about a location or something at a location and provide context for the setting. They should be linked to specific locations instead of something abstract or unactionable. This location could be anywhere from a settlement, a dungeon, an area within a dungeon or settlement, or a desolate bluff in the desert wilderness.

At their core adventure hooks break down into some fairly simple formats.

For example:
I’ve heard there’s something valuable over there.
Here’s how you can avoid the dangerous thing over there.
Go there and get this thing and you’ll get something in return.
Bring this thing there and you’ll get something.
Go here and kill this person or monster.
Go here and rescue this person or monster.

This is a simple process. We can make these hooks more engaging by applying the therefore or but principle to the hooks and adding complications. The more of these hooks between locations we create, the more interconnected and dynamic our sandbox is.

One type of hook that’s worth discussing in more detail is the rumor. I prefer adding a good number of diegetic rumors spoken and known directly by specific NPC’s or found on specific documents rather than always rolling on a random rumor table and hearing the rumor mysteriously bestowed from a general entity “They say…”.

Multiple people might know the same rumor or have similar quest hooks, though their versions could be widely different. And if rumors are false or misleading, the PC’s know who to blame! Specific information can be a reward for seeking out and talking with specific people in the world.

I still think there’s still a big place for rolling on a Random Rumor Table. It would take way too much work and page space to specify what general information every single NPC knows about the region. Instead it’s better to roll to procedurally generate rumors on the fly that NPC may be likely to know, just once for establishing the background information a PC starts with, or if the PC’s hit the tavern and do some socializing.

I really like how The Halls of Arden Vul divides adventuring rumors (adventurers, people with practical experience with the locations, or people who talk to them) and historical rumors (sages, researchers, and ancient beings). The tables are broken down into Common, Uncommon, Rare, Very Rare, and Legendary rumors that have varying odds of the person knowing them. I’ll probably adopt something similar.

Getting back to the process of laying out the hook connections while planning a region my attempts to use a Google Sheets document to note all the location interconnections using a format similar to the Colossal Wastes of Zhaar crowdsourced hex crawl was difficult for me to wrap my head around. I’ve discovered I’m a bit more of a visual person and decided to use Scapple to take each of my hex locations from my map and create a diagram showing arrows designating hook linkages.

Each arrow represents a specific hook leading from one location to another.
Yellow locations are dungeon sites, Blue locations are settlements, and the brown one is the Cinderstrom volcano megadungeon.

I have lots to add and tweak but you can see some patterns emerging.

There are hubs that are both popular end points and starting points. Some hooks just lead directly to one location while other hooks lead to the next in a series of them. A couple desolate locations don’t have any hooks leading to them and can only be investigated by stumbling upon them. I haven’t finalized all of the quests or come with an equation for connection over just eyeballing it.

Every settlement has at least 3 hooks leading to other places with up to 7. Cinderstrom has 8 different hooks leading to it and in turn leads to 3 different locations outside of it. These numbers should increase for all locations as I build up higher interconnectivity by adding more hooks. Here’s a quick example of a segment of these hooks from the Bandit Ruined Villa.

Following this process of hook connection building can take a sandbox from a weightless series of random encounters and strange locales to a more cohesive overarching adventure.

With a video game like Fallout New Vegas or the Witcher III, side quest hooks are included to add variety and interest to the expansive world instead of demanding the player follows only one main plot bearing quest. While in a tabletop RPG sandbox region everything is a side quest!

However in a region with a tentpole megadungeon there’s a center of gravity to the sandbox and a nexus around which a hurricane of potential energy and plots swirl. Just as the megadungeon volcano Cinderstrom looms over the basin, its shadow and influence should touch a good number of the other hooks in the region. I’m looking to accomplish this by having all the major factions care about the megadungeon, having a lot of hooks that lead there seeded throughout the region map (8 so far), and putting the most valuable, weird, and powerful stuff inside of the megadungeon.

Nothing stops me from flipping the script and placing hooks and important items or people within the megadungeon itself that connect outwards istead. This replicates the organic feeling of interconnectedness found between rooms within a good dungeon and supports adventures outside of the megadungeon that remain linked.

One unformatted example of this in action:

I probably should add at least one more connection hook here from elsewhere to increase player’s odds of figuring out how to access the Thricelocked Door of Thrones in keeping with The Alexandrian’s Three Clue Rule This type of content gating behind specific access requirements shouldn’t be required to access the majority of locations. Instead these hooks are for an optional hidden bonus and danger for those who venture to investigate off the beaten path.

The thicker woven these web of hooks are the more likely your PC’s are to get stuck fast and cause impacts that ripple throughout the region while the organic story that’s generated through their episodic roving and delving develops overarching themes, reoccurring nemeses, and familiar territory that help tie past actions to future consequences.

There’s been heaps of great writing has been about the value of specific procedures in old school D&D that atrophied away and were forgotten as editions marched on. The reaction table, morale rolls, and a system of location exploration turns linked to diminishing light and increasing odds of random encounter are all valuable to support a specific style of gameplay that has time and resource pressures and doesn’t default to all things invariably trying to kill you.

Another element of specific procedure for sandbox games that I think could use a renaissance are Campaign Events.

The original AD&D Oriental Adventures comes with mountains of well deserved baggage but it did have one area that provided a really useful table for inspiration-the Campaign Events section. A DM could use the tables to roll up the overarching events of an entire year in advance.

There was also a daily events table but I feel that territory has been pretty covered by using location Wandering Encounters table stocked with events as well as potentially hostile foes for whatever city, blasted wasteland, or hole in the ground the PC’s find themselves in.

A neat element of the system is that the table rolled for monthly events are affected by whichever yearly event was rolled while the seasonal event is happening. If there’s a 4 month War for the seasonal event on then Major Battles and Bandit Activity are likely while there’s little or no chance of marriage events for the months that the War lasts. Curiously, seasonal events of natural disasters also make every other natural disaster likely while they last, if there was a Famine then a Minor Earthquake is suddenly more likely during this time. I think this non-scientific take can work well for a lot of fantasy settings as the natural world is often portrayed as magically interconnected and if there’s a period of natural disaster and instability it makes sense that it affects the rest of the natural order.

These tables are interesting because they introduce a broader plot that doesn’t involve the players while using the oracular power of dice to surprise the Gamemaster and inspire them through producing seeds, pushing the game in a direction they wouldn’t have taken it if they had just created the timeline of future events from their own plotting and creating an impression of the wider world going on and changing without them giving a great sense of verisimilitude to the setting. These events don’t have to be written in stone but offer inspiration and provide a living world that changes without the PC’s touching it.

When I’m running a campaign I’ve found that it’s easy to have everything just remain in a state of delicate stasis instead until the PC’s come and screw it all up. In my opinion, Hot Springs Island is the best sandbox RPG adventure published at this point. They refer to the setting of the island as a “keg of black powder” with volatile elements and complex factions in a precarious equilibrium waiting for the PC’s to come in and light the match that blows it all to hell with their shenanigans. One downside of this approach is that that large changes to the setting are only precipitated in a rippling butterfly effect as a result of direct or indirect interaction with the PC’s and centers most events of broad reaching disruptive change around them.

This is my default way to run sandboxes. If the PC’s haven’t interacted with it, it exists frozen in the same state until they do, or their actions impact it. This works for my brain the same reasons most video games don’t render every single level and simulate every entity in the world while the players aren’t there- I’m saving precious processing power. Still, there’s a lot of gameplay value in simulating a world that feels alive and changes in big ways in the background without any PC input.

Before getting to the tweaked campaign events table there’s a couple other ways I’ve seen modules and systems introduce change that exists independent of PC’s. The sandbox Mothership module Pound of Flesh uses an event meter that ticks up in one direction every session along a series of crises from a dockworkers strike to a cybernetic plague until they get to their inevitable cataclysmic conclusion unless PC’s act to prevent this. I used a more limited potential timeline of future events if the PC’s do nothing in my space western module, Desert Moon of Karth. (I took the future timeline idea from Deep Carbon Observatory.) These are great and focused but only provide one or three potential timelines of changes instead of a huge procedural range for a long term campaign in the same area.

Faction procedures offer another way of introducing more setting dynamism without needing the PC’s finger in all the pies. I ran a real enjoyable star hopping campaign with friends using Stars Without Number for 14 sessions. The detailed faction procedure was a lot of fun to run six factions through for an hour each week with an automated spread sheet of their assets but the solo RTS I was playing didn’t seem to deliver the same value for the players. I think something like Mausritter’s faction procedures would have been more useable for my purposes. In that procedure, factions each have several goals that require a certain number of progress marks depending on how difficult the goal is. Between sessions a die is rolled to see how many progress marks are added to their goal adjusted by how many unique resources they have at their disposal. If the goal is targeted at harming a rival, they lose one targeted resource when the goal is completed. If PC’s help or harm the goals of a faction they can add or remove a number of progress marks towards the goal depending on the scale of their interference. I think a faction system like this could pair nicely with an overarching campaign events table representing more unexpected outlier events and acts of nature than the steady progress of factions towards their goals.

In making my own version of a Campaign Event Table, I made a number of changes from the Oriental Adventures events table to make it more user friendly, and customized to my swords and sorcery campaign purposes and preferences.

First, I changed the types of disruptive overarching events to the categories Faction Chaos and Natural Chaos which sums up the origin of most types of large changes that affect the area. I also reduced the timeline scale from Yearly Events and Monthly Events to Seasonal Events (12 weeks) and Weekly Events. I did this because I tend to operate campaigns on a shorter in game time scale than I think they were assuming in 1985. If you run a game with regular downtime in between or that operates on the same time frame as reality in between sessions and dungeon expeditions it’s easy to have an event come up every session, which is my ideal pace. You could easily adjust what increment of time the tables represent to your liking.

I think these are best rolled in advance to create a timeline of outside happenings if the PC’s don’t interfere and synthesized into a more detailed and partially interconnected series of events. Also credit to John Arendt for the Campaign Events table in their Black City setting inspiring several of the entries.

How It Works

1. Roll on the Seasonal Event table. Roll for how many weeks of Faction or Nature Chaos caused if needed.
2. Roll 1d12 to see what week of the 12 week season the Season Event starts on. 
3. Roll a Weekly Event for each week, rolling on the Faction Chaos or Nature Chaos table if the week occurs during a period of either from the Season event. 
4. Repeat the process for more Seasons if more time in your game’s calendar is needed. Though probably best to just wait until you get right before the next season to roll again.
5. Look at the series of events and weave them into an interconnected timeline of what will happen in the region that season without direct interference by the PC’s.

Weekly Events

Blank Season Calendar of Events

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

An Example Season:

I rolled for the Seasonal Event of the year and got a 93, Red Comet. They then rolled a 1d12 to see what week the Seasonal Event started and got Week 7. Reading in the description of the Green Comet that it lasts for 1d6 weeks of Nature Chaos they roll to see how long it will last and get 2 weeks. The first 6 weekly events are rolled using the During Normal Times table, and weeks 7 and 8 are rolled on the During Nature Chaos table, and the last 4 week’s events are rolled on the During Normal Times table again for the following 12 week Season results.

1. New Dungeon Discovery
2. Accident
3. Birth
4. New Dungeon Discovery
5. Mistaken Identity
6. Merchant Caravan
7. Red Comet; Death
8. Red Comet; Death
9. Supply Issues
10. Skystone Impact
11. Haunting
12.  Outsider Visit

  Looking at this timeline of events I decide that the Season starts with a rival adventuring party discovering the Steaming Grotto’s hidden entrance to the volcano megadungeon of Cinderstrom and bragging about it a bit too loudly in a settlement’s drinking hall. Next week one of the members falls into a mud pot and is nearly fatally scalded on a subsequent expedition requiring bandages over their entire body. The third week, Anasa the herb merchant gives birth and there’s much celebrating and festivities in the village of Nabtaar. In the fourth week, the same rival adventuring party decides to venture away from Cinderstrom and stumble upon the Mound of the Horselord and prepare to loot it and brag copiously again. This angers the plainsfolk who hear about it and a rowdy mob is assembled the next week to punish the party defiling the tomb of their ancestor. Unfortunately their info is bad and they believe the PC’s are the ones that entered the mound while the rival party skipped town.

On the sixth week a merchant caravan passes along the trade road and brings merchants bearing strange goods that are available for the short week the caravan stops in the Cinderstrom Basin, the rival party flees the region with them. On the seventh week the Red Comet is visible and magic is empowered and strange things stir as it’s visible for two weeks, in the same week dealer of the black pomegranates in the Janeer Bastion is wracked by convulsions while gazing up at the Red Comet and dies. The eighth week see’s the strange death of the recent mother Anasa the herb merchant in the village of Nabtaar under similar circumstances. After that the comet passes and the natural world reverts to a less weird state. However, the Red Comet’s emanations polluted all of the magical potions brewed in the region and none are available for that ninth week until more are brewed. On the tenth week a blaze is seen in the sky as a late chunk of the Red Comet flies from the sky and lands on the Tallgrass Plain  with an ear shattering boom, killing a herd of mastodons and emanating magical radiation. The eleventh week the tortured and angry souls of the mastodons rise as specters and create a new hazard for any travelers across the Tallgrass Plain. Finally in the last week of the season a robed stranger with a huge ax comes to seek out the Mound of the Horselord that they heard about from the rival adventuring party fleeing north with the merchant caravan 6 weeks ago. They claim they are the heirs to the Horselord and wish to claim their birthright within the Mound.

This took me about 10 minutes to come up with and write down. The events rolled aren’t set in stone but fertile seeds for inspiration for a timeline of what will happen without player activity that actively prevents the events from occurring. In order to get information of these out to the players it’s great to add to drop evidence of the event into local rumors or NPC conversation if the player’s don’t directly witness the event.

Let me know what you think! Have you ever used a campaign events table? How did it work?