A few months back I ran a heavily modified portion of the Dungeon of the Mad Mage mega-dungeon for a group of friends. I made a bunch of changes to the dungeon, some of which worked well, and some of which did not. The one I think worked the best, however, was my reworking of Level 7, Maddgoth’s Castle, and specifically the Caverns that lead up to the titular castle.
A quick note before I begin; I won’t be including any specific room details or maps from the official book in this post, as I’d rather avoid any potential intellectual property issues. That said, this style of maze/puzzle would, I think, work well in a great many settings: marks left by the thieves guild in tunnels under the city, scratchings on the walls of a kobold warren, etc.
The Problem
The map of that floor consists of a series of winding tunnels that interconnect with each other and a series of about 10-15 more open chambers, including one large central chamber that contains Maddgoth’s Castle.
The party I was running for didn’t seem particularly interested in exploring every nook and cranny, and without some pretty heavy modifications or improvising, would lack the information required to make any meaningful decision while navigating. Worse, the party tended towards a left-hand or right-hand navigation style, which would leave them with a pretty good chance of skipping Maddgoth’s Castle entirely, which would be a real pity.
The Goal
As a result, I decided to rework the tunnels leading to the castle, but leave the castle itself pretty much as written. My goals in re-working the tunnels were:
- Adjust the path of the tunnels so the players were either forced or at least very likely to come across the castle, and have reason to explore it.
- Keep the tone of getting lost in a maze, and having to figure out how to navigate it.
- Add enough information that the players could make meaningful decisions when navigating the maze
- Create a bit of a puzzle around discovering the means of navigation.
The Solution
The first thing I did was sort out 12 fairly unique rooms – not all of them had an encounter, trap, or treasure, but each had some unique features that would allow the party to recognize them if they came back through again. Some of these were drawn from the source book, some were entirely my own invention, most were a mix of the two.
From there I grouped the rooms into sets of 3-4, and for each one drew a simple graph of the rooms in the group all connected by tunnels. Then I started adding small offshoots for each tunnel; some of these were dead ends, some led to the same room as the main tunnel, so to others, and a few back to the starting point. I ended up with a bunch of graphs that looked something like this:
From there, I connected the groups of rooms, occasionally with new tunnels, and occasionally with offshoots from existing tunnels. This allowed me to set up choke-points that ensured the party passed through either a specific room, or one of a pair.
On its own, this would have been a nightmare to navigate, and if I’d had a group of players big into drawing their own maps and leaving marks to keep track of their location that might have been fine. However, my group wasn’t interested in that style of play, but were big fans of puzzles. So the next thing I did was assign a letter to each room:
Then determined that, over each door, there would be a string of characters that indicated what was down that tunnel and its offshoots. There were a couple catches here:
- The symbols would be in the order the party would encounter the tunnels that led them there.
- The symbol for the room they were currently in would be capitalized/larger (but only for the instance in the string of character that lined up with that tunnel).
- I used “Dwarven Runes” instead of English letters, so it wasn’t immediately obvious that the strings of characters were not words. I won’t do that in this post though, for the sake of clarity.
Here’s an example of the key for one of the rooms in the graph:
For each room I prepped an index card with a list of the exits and the associated strings of dwarven characters. Each time the party entered a room and examined the markings I could hand them the card for that room. When they left the room, unless they copied the markings down, I took the card back.
Finally, one of the themes of the dungeon was a curse that caused any creature who took a long rest to lose the last 24 hours of their memories. As such, I prepped a second set of index cards made using a different mapping of rooms to symbols, so that if the party took a long rest I could take all the cards back, use slightly different descriptions for the rooms, and they wouldn’t be able to rely on their existing knowledge of which symbols meant what any more than their characters could.
How it Went
The players mostly ignored the lettering for the first couple rooms; they assumed they were names of the rooms they were going to, and that since none of them knew Dwarvish they’d just have to figure out what the names meant through trial and error. Their initial strategy was to copy down the runes of the tunnel they were heading down, then make a note of what was there when they got there.
Pretty quickly, however, they ended up back in a room they’d already seen, by following a tunnel with a different label. After that the wizard started collecting rubbings of the symbols over each door (I just let them keep the cards from that point forward). After one more time around the loop they realized one room had a path with a symbol that wasn’t present in any other room, so they tried that route.
I had ensured that the tunnel connecting the first group to the second didn’t have too many tricky offshoots, so they were able to reach the second area without too much trouble. From here they followed a similar tactic, noticing eventually that the first and last characters for most tunnels indicated a start and end room. Not always, but they missed those cases for the time being.
After thoroughly exploring the second group of rooms, they discovered that the only place a new symbol was showing up was in the middle of a “word” over one of the tunnels. At this point, one of the players acted on a hunch, and told the rogue who was scouting to try the 3rd side-tunnel he passed, as it lined up with the position of the character in the string.
This let them reach the third and final group, which they were able to navigate without any looping back or ending up somewhere unexpected. They ended up discovering the memory loss curse before it could impact the entire party, and were able to continue on to the Castle itself.
Things to Change
Overall the maze worked quite well. The players didn’t immediately find out what to do, but didn’t end up stuck on anything so long they got frustrated. There were a few things I would have liked to try changing though.
First, it might have been interesting to have the memory loss keyed to progressing to within a certain distance from the castle. I think having the symbols change, but the positions of the larger symbols staying the same might have served as an effective indication of the meaning of the size difference. As it is, that didn’t become apparent until fairly late in the process, and the party consistently had trouble with exits from rooms that were side-tunnels. Alternatively, putting enough dangerous combat in early to ensure they take a long rest would have made sure they ran into the curse.
Second, given how likely it was that the party would be revisiting rooms multiple times, I should have put some thought into what might change in the room between visits. There wasn’t likely a great deal of time, but having something roaming the tunnels leaving behind signs of its passage would have been a great way to foreshadow an interesting encounter.
Finally, I should have come up with a way to indicate a direction of travel for the tunnels, and which ones were side tunnels. I ended up describing the side tunnels as slightly smaller, and trying to indicate directions by describing the intersections with smaller tunnels as being more like a highway ramp than a stop light, but neither was particularly helpful.
I don’t know that I could pull this puzzle out again for the same group of players, and I suspect it would have quickly lost its luster if I’d tried to drag it out for a couple more groups of rooms. That said, it worked very well for my group, and hopefully will for some others as well.
