EDIT: I honed in on that tree. I saw nothing but it and I heard nothing but the clacking of my keyboard as I typed up this post. Upon stepping back… Alas! I saw point crawls.
Dungeon crawling has been the standard game structure for many roleplaying games for a long time. During this time people have presented all manners of preparing them from a series of randomly generated encounter tables for every room or corridor to entire algorithms that generate the dungeons and its mundane contents. Some game masters run them with the notes written near each room, others have a separate sheet of paper and a key to match descriptions to rooms, and some game masters are mad and ad hoc the whole thing. I have tried each of these methods with varying degrees of success, but I was never entirely satisfied with how they played out. I recently learned of tanglegrams which are like mindmaps that emphasise the relationship between people and things – you can read more about them in my original post here – and I believe they would work very well for helping your dungeons feel more interesting.
I believe the greatest impact on my dissatisfaction was my misunderstanding of how a dungeon crawl runs and using a system that did not actually supply rules to facilitate them, so I defaulted to the absolute basics of the dungeon crawl structure outlined in the article Game Structures – Part 3: Dungeoncrawl by Justin Alexander. This structure works well, and a session run this way can prove to be enjoyable – after all, a narrative can still be spun, but it is lacking in defining how these rooms are connected in the dungeon. In more traditional games, rules are supplied for exploring a dungeon and it can often involve turns with random encounters or resource management that is influenced by corridor length. Corridors being the connectors of the rooms in a dungeon. Without those rules, the dungeon can fall flat and lose that sense of exploration unless a game master particularly accounts for the corridors. Tanglegrams can provide an easy way to prepare these corridors between rooms because the corridor between rooms, in a way, is how these rooms relate to each other.
How does this work?
Take the dungeon above that was generated using the donjon random dungeon generator. It features four rooms, three dead ends, an entryway, and stairs leading down. The first step is to convert this into a tanglegram that has each room and a line connecting them to represent the corridor – the relationship. You can see my interpretation of this below.
Currently this tanglegram form of the dungeon is less interesting as it has less twists and turns, however this leads to a cleaner interface for the game master to work with. The advantage of this cleaner interface is now more detail can be added to those corridors as there is more space to work with. In Figure 2 I have included the dead ends as dotted lines and attempted to maintain the relative positions and size of the four rooms which are all connected with curved lines. The lines with arrows signify entrances and exits respective of the arrow direction. I suppose you could go into as much detail as necessary for this step but then you may as well map the dungeon in a more traditional way. The next step involves populating the dungeon. I am going to gloss over the actual rooms and begin to add in details for how they are connected, however when this is done the rooms should be considered as this allows the game master to foreshadow things from the rooms that are connected. You can see this below.
The tanglegram dungeon now looks a little more interesting with the included substance and more importantly: the relationships that connect the rooms and intersections. I suppose one could include more detail and even highlight encounters and the like in these relationships, but I elected to maintain the purity of the relationships between rooms. That first corridor connecting room two and three allows a game master to describe the feel of that connection and add a little bit of information for the players to consider – the splashing. This is further used in the connection between three and one and then one and four where the growing sound of voices becomes clearer. At a glance, it is easy to see where the party is heading and what hints you can provide them which should support them in their agency and decision-making. However, the greatest strength of mapping a dungeon in this manner is that flexibility to provides in what these connections can look like, for example I made the connection between one and four a shaft leading upwards instead of a corridor. Because this method does not care for geographical accuracy, the flexibility for connections increases. All a game master has to do when the party enters a room is check the connections it has and describe it as such alongside whatever else they had planned to occur in that room.
This method is not without limitations. Dead ends can still behave strangely with this approach as they are not necessarily rooms but still exist as an entity that is not just a connection. To get around this one might elect to treat them as such or use them as flavouring along the connection to break it up such as the raised platform in Figure 3. Lastly, this approach does not elicit the same feeling of exploration as the more traditional dungeon crawling, instead it provides flavour and more pseudo-choice for the players.
Overall, I think this is an approach that is worth exploring if you lack the rules for traditional dungeon crawling, you find that style of play not to your taste, or you want to adapt it to bolster the way you run corridors. Though this approach is limited with certain aspects of dungeons like dead ends and only working for theatre-of-the-mind games, it does provide an excellent strength in the form of flexible mapping in which a game master can include all sorts of strange and different types of corridors that connect rooms such as the ladder shaft above.