How do you run a maze?

[Shinizak]

Maze-like dungeons are great in theory, but I find that in practice they’re hard to pull off elegantly. Between the constant question of “where am I” and “whats going on?” It makes it hard to run a dungeon any larger than 6 rooms. How the hell do you pull off 20+ room dungeons? Or worse, mega dungeons?

[OldTrees1]

Are you asking about mazes (where solving the maze is part of the challenge) or about dungeons with complex interconnecting rooms?

When running a complex dungeon / megadungeon with many interconnections, I default to some way of giving the players the map. I tailor the method of map delivery to meet the needs of the megadungeon.

For example with Castle Raveloft I showed them a map of each room as they explored. When we were in person I also had flashcards with the room maps on them and connections labeled. I then handed those map cards to the players as they traveled. That way they could assemble the larger megadungeon map based on what their characters had uncovered.

Currently our group is online and using the virtual tabletop's fog of war to hide map information until the players uncover it. As we move away from an explored area, the explored area gets fog that covers NPCs but does not cover terrain/walls.

[NichG]

Not sure about mazes in particular, but larger dungeons...

- Provide the map and have a way to uncover/draw bits of it from record as its explored. Don't ask the players to do the mapping based on descriptions. (Yes, you can do this; yes, its a valid form of play; but given the difficulties described, this is an easy way to reduce them significantly)

- When you have some section, ask 'what is detail here doing?' and abstract things in play where you don't have a good answer. So e.g. you can have spiral winding passages on the map for visual effect, but don't make people move their minis or whatever through each kink of the passage unless there's actually going to be a fight there.

- Don't put things which would require ambiguity about what level of abstraction should be used - e.g. if you're going through entire winding passages in one 'move', don't put a trap or hidden switch on a particular square that would require tile-by-tile detail to interact with reasonably.

As far as mazes, the conclusion from points 2 and 3 leans towards '(usually) don't do literal mazes'.

An exception that I did run once was something where the entire area had a time limit that was small (rounds), so 'how fast can you get through the maze' actually mattered, and everything was just done round by round. But that was still more of a specific setpiece than a megadungeon sort of situation.

[Saintheart]

Maze-like dungeons are great in theory, but I find that in practice they’re hard to pull off elegantly. Between the constant question of “where am I” and “whats going on?” It makes it hard to run a dungeon any larger than 6 rooms. How the hell do you pull off 20+ room dungeons? Or worse, mega dungeons?

Short answer: you don't. Massive dungeons belong to the earlier editions of D&D, when attention spans were longer than that of a Dory. I just doubt they really work anymore other than for a certain kind of group.

Which is to say the problem could be your players, which more politely is to say, your expectations might not be the same as theirs. If people are constantly asking "Where am I?" and "What's going on?", then odds are on your players are not paying attention and certainly not mapping out where they're going. Which could be any number of reasons (aforesaid attention span of a Dory or allowing phones to be consulted at the table) but it comes down to the fact that if that's the crowd you're dealing with, then you might just have to admit you're not going to be able to run the sort of adventure that you want to run as DM. They're not involved in the game, they've come for some other form of enjoyment than filling in the empty sections of the map.

Not your players' fault, they're not a bunch of casuals who just want to get together for beer and pretzels gaming, they're actually taking notes and mapping things? Okay, then the problem lies with you or possibly the megadungeon.

The problem with the megadungeon could be that it offers too much choice too soon. Megadungeons rarely allow you to go anywhere you want, they're usually compartmentalised into smaller areas which have to be unlocked one area at a time. This is how videogames do it, and it's not just because they have to load different assets between areas. Indeed that's where the very concept of "level X" spells or monsters came from: because the original dungeons were designed as underground structures where the opposition got tougher as you dropped down level by level into the structure. Too much choice is the same as no choice.

The problem, as said, could also be that there are insufficient memory cues to help the players orient themselves. Have you got any landmarks in your dungeon? Any big "crossroads" rooms which have a big statue of something ugly pointing to the right? Sure, people have short attention spans, but unless you have a dedicated mapper in the group we tend to orient either by simple direction-based heuristics ("go right at each crossroads", "tie a piece of string behind us to see where we went") or by landmarks or distinctive features.

The DM's job is to give enough information for people to make choices. That doesn't mean they get everything, but they should be getting enough to orient themselves. Again, if that's the kind of players you've got.

Other issue being: what's the maze-like dungeon being used for, what's the party doing in there? Is their job to clear the whole place out, or look for a random McGuffin they've been told is somewhere in there? Is the maze its own defence system, i.e. to contain a minotaur at the centre or something? Because making players map out areas accurately without a particular reason isn't terribly satisfying gameplay either.

[oxybe]

If the players go into a megadungeon and they are aware of it's size, but are not actively mapping it... don't help them. Let them get lost and keep track of time

Now, I'm coming from 2nd ed I've found it best to break things down in scale depending on what the players are doing.

Combat Turn 1 minute
Dungeon Turn 10 minutes
Exploration Turn 1 hour
Travel Turn 8 hours
Day 24 hours

A combat turn is exactly what you think: the active action of swinging your sword, moving about, taking a defensive stance, etc... largely quick actions on a micro level.

Dungeon turns are basically the GM asking "so what are you doing?". carefully picking a lock, taking a rough sketch, doing a once over of a room, etc...

Exploration turns are dungeon turns on a larger scale: taking a breather to recoup, prepare your camp or a solid defensible position, doing a thorough look over a room, copying a larger volume of texts, examining and studying glyphs/or whatnot.

Travel turns are mostly used during... well travel. basically used to determine if there's something that happens mid travel on longer distances and, for what it's worth, the time it takes to cross one hex on the map.

Day is a day. I'm pretty sure you can imagine stuff that takes a day, but one day is the time it takes to explore a hex, not thoroughly mind you, but to get a feel for the area and possible points of interest

for the most part these aren't strict adherences, but it helps to break things down in these manageable chunks of scope when trying to parse how long the party has been doing things or to see how long they have until their torch dies and much light a new one or rely on their allies to traverse the dark, the actions they can take while spells are active, etc...

thus depending on how big your dungeon is, you can apply pressure on players. if a torch lasts an hour, or 6 dungeon turns/1 exploration turn, you know that they've been through a combat (10 min ish), gave the room a good once over (10 min) and faffed about in the hallway looking for traps that don't exist (10min) their torch is about half used up. you can also use this as a metric on how frequently to roll random encounters, if any. In a densely populated dungeon you may only gave an hour or two before bumping into a patrol doing their rounds. If they're on high alert, it may well be even less. do note that a random encounter isn't necessary a combat encounter. You might just bump into a kobold non-combatant out foraging for mushrooms, roots and grubs, for example, in your mega dungeon.

Beyond that remember that a dungeon, and especially a megadungeon, should be something special. you're cut off of civilization and all the perks that come with it. you lose the lifelines of hospitals/priestly curing, food and water must be searched for once your stores run out, consumable items like torches are difficult to replace, etc... so make these things count. this isn't a school field trip: you're either behind enemy lines and having to watch your back, or traversing an unexplored terrain where danger could be lurking around the next bend.

It may sound like a lot of bookkeeping, but it really isn't. like i said: break your sessions and expeditions in smaller and more manageable chunks of time and you can have the players feeling the pressure of dwindling sources of water and food, down to their last torch and unsure if they should camp for the next few hours or move forward to make sure the bugbears can't catch up with them.

[Vahnavoi]

Running a maze and running a large structure are two different issues.

For the former, you can save yourself a lot of time by buying a copy or two of Labyrinth board game. You can use the pieces from the game to handily generate various maze-like structures and they can be used in the manner other posters suggests, visually revealing the maze bit by bit, removing need of players to draw their own map. The cards of the game can be keyed to various random encounters to populate the maze with events, the pictures on the board pieces can be keyed to special rooms or fixed landmarks. The latter becomes especially important if you choose to use the moving function of Labyrinth to change the floorplan while players characters are within it.

Once you're familiar with how Labyrinth works, it's trivial to do the same on grid paper or by drawing your own board pieces on post-it notes.

For the latter, forget everything about "dungeons" for a second and go walk inside any real large structure - factories, museums, shopping malls, parking halls, campuses, what have you. Take some paper with you and try drawing a map while walking through. Pay attention to how you orient yourself and how much time it takes to move from place to place.

Why? Because if you don't have experience going through large structures, fat chance you will be any good describing such experience to other people.

So that's something on the game master's side, what about your players?

Have you actually taught them how to draw map symbols? Have you taught them to use the same symbols you use? Have you taught them to use a compass? A ruler? Unit conversions and scales for spatial measures? Cartesian coordinates?

Because these are skills that need to be developed if you want mapping to be functional part of a game. The good news is that games are a suitable medium for doing so, the bad news is that modern tabletop roleplaying games typically don't have the advice for it. You want to take a page or two out of a Scout handbook (etc.) and probably take a break from tabletop games and go play outside.

Once your players have some real experience of cartography and orienteering, they'll be much better at it in a game. More, it will make their questions of "where am I?" and "what's going on?" much more tolerable because you'll have better, shared understanding of how those questions ought to be answered.

[KorvinStarmast]

Besides what saintheart and Vahanavoi said, running a maze works better when you give the characters a blank sheet of paper, or the blank battle mat, and they have to map their progress through the dungeon. You describe what they see.


GM: "you enter the door and see a hallway, ten feet wide, going 40' north"
Mapper: draws that onto the battle map.
Players: describe what they do now that the see the hallway.
GM: the hallway goes another ten feet and ends in a T intersection, with hallways going east and west.
Players: sketch it in, describe what they are doing.

As the first player turns east, DM may call for a saving throw or an investigation check, since there is a concealed pit in the floor ...

And so on.

What you are doing here is called the exploration pillar. We did this for years and had great fun with it.

[OldTrees1]

GM: "you enter the door and see a hallway, ten feet wide, going 40' north"
Mapper: draws that onto the battle map.
Players: describe what they do now that the see the hallway.
GM: the hallway goes another ten feet and ends in a T intersection, with hallways going east and west.
Players: sketch it in, describe what they are doing.

Adding on to what KorvinStarmast said

From my experience this method works, but introduces random "off by 1 5ft" errors that accumulate. These errors are usually something the PC would not make but the Player did make due to slight miscommunication between GM and Player. As a GM I found correcting these errors was disruptive. Instead I suggest designing the map such that rounding errors don't cause problems.

[SimonMoon6]

You have to consider what the purpose of the maze is.

If the point is to frustrate anyone who enters the maze so that they will become lost and unable to find their way out, then it is perfectly acceptable for players to ask questions like "where am I?" and "how do I get out?" The answer is "You don't know" and "you don't".



But that may not be ideal for adventurers in a game.

If the point is just to take up extra time mapping, then fine, let them map and make sure every corner of every corridor is exactly 90 degrees, so that they can map and never get lost.

My preference would be for the PCs to be teleported to a random part of the maze (possibly while exploring it). Especially if they don't get to map things out. And all you do is to tell them when they see a left turn, a right turn, or a T-shaped intersection (etc). But then my players would always just choose one wall and stick to that (like always stay next to the left wall no matter what and *eventually* they'll get out).

Keep in mind, real world mazes found in tombs rarely have 10 foot wide passages with 90 degree angles. Things are twisty turny and tiny. So, maybe a 3 foot wide passage that turns at a 34 degree angle for 13 feet before turning to the left at an angle of 28 degrees. Of course, that's a nightmare to map and players are obsessed with mapping, so you'd just be creating a huge headache, but that's how you make a maze super-annoying.

The one maze I've seriously run involved a maze home to minotaurs that had special features. For example, some of the walls were illusionary walls, so the minotaurs could suddenly charge through a wall at unsuspecting PCs. The minotaurs could see through the walls but the PCs could not. This could also be used to shepherd the PCs down hallways filled with pit traps. And of course, other maze inhabitants take advantage of the constant turns in the maze by being mobile, using hit and run tactics like spring attack to make sure that the PCs have a hard time getting line of fire with spells, ranged weapons, etc. Again, this can sometimes force the PCs forward to chase these guys down, but then the PCs risk going into passages filled with pit traps (while the spring attacking svirfneblin (or whatever) simply dashed through an illusionary wall while out of sight).

[Psyren]

It's not that hard. Either you have a premade map they can explore, or you don't.

If you do, you don't have to do much to design the maze itself - just make sure there's interesting stuff/encounters in most of the rooms. (Note that "an empty place they can fortify and rest in" can qualify as interesting too, so long as there aren't too many.)

If you don't, you can procedurally generate one, or you can abstract it via the dice. Personally I would go with the latter, because I would find it nearly impossible to draw something engaging mid-session. ("Abstract it" in this context means treat it like any other exploration challenge, which means following the challenge design steps:)

Spoiler

Show

1) Is this challenge possible to overcome, but not trivial? If yes, dice should determine the outcome. If no, skip ahead narratively until you arrive at a challenge where the answer is yes.

2) Once you've decided the outcome, determine the difficulty / DC. Consider different difficulties depending on the method the PCs use. For example, deciphering a map of the dungeon might be easier

3) Define a meaningful consequence for success or failure. In particular, you should define how many successes or how large a success is needed to overcome the maze as a whole, vs. just making progress or finding goodies. Also, allow for multiple approaches to progressing the maze - maybe some PCs are skilled with architecture or the culture of the maze's creators, maybe some are skilled trackers/sniffers, and maybe some are relying on magical aid.

4) Determine the outcome of the roll. What does a failure mean? What do 2-3 consecutive failures or successes mean? Does a critical failure or success mean anything? Etc.

EDIT: You can also go for a hybrid approach. Rather than premake the whole map, or draw the whole thing on the fly, you can come equipped to the game with a stack of disjointed dungeon rooms, then abstract whatever happens in between. This way you can have more detailed encounters happen in the rooms (e.g. combat) but not have to be too fussed about making them all fit together in a larger space. This might be a great time to whip out all those maps you bought on Humble Bundle or Kickstarter

[Reversefigure4]

If you want a maze itself rather than a dungeon, you can always abstract it to skill checks. Rather than trying yourself to come up with the meticulous twists and turns, you can go to "Roll a Survival check. If you fail, you take a level of Fatigue as it takes extra time to make your way through the maze."

You can even combine them, so you take the Fatigue level then show up in the Hobgoblin Alchemy room.

If you're playing a more long-term dungeon, I'd suggest maps are more fun for players; but I'd recommend the skill checks if the maze is either a fairly background thing - you're checking how many hours or how much damage it takes the players rather than whether they can get through period - or something so complex it can't be mapped or the point is to confuse the players (the dungeon is three dimensional, the rooms move, there's teleportation that takes you to random places in the dungeon).

[LibraryOgre]

A fun, but abstract, way I played to do a maze was with playing cards. You laid out a grid of face-down cards (randomized). Players picked a card, and the results correlate to a chart.

Each pick of the cards took a certain amount of time (so, you might say 10 minutes per pull). Certain cards meant certain types of encounters, including battles with a minotaur, treasure, and traps.

And every few turns, the DM picked up all the remaining cards, shuffled them, and laid them down in a new, smaller, grid.

I think it was Fane of the Heresiarch for 4thcore.

[Psyren]

A fun, but abstract, way I played to do a maze was with playing cards. You laid out a grid of face-down cards (randomized). Players picked a card, and the results correlate to a chart.

Each pick of the cards took a certain amount of time (so, you might say 10 minutes per pull). Certain cards meant certain types of encounters, including battles with a minotaur, treasure, and traps.

And every few turns, the DM picked up all the remaining cards, shuffled them, and laid them down in a new, smaller, grid.

I think it was Fane of the Heresiarch for 4thcore.

Hand of Fate does this too

[Pauly]

I don’t run mazes.

I’ve been in underground mines, catacombs, the excavated ruins of various cities, large battleships and markets in undeveloped countries.
None of them are anything like the traditional D&D maze-like dungeon.

Some points.
1) The people who use(d) these structures needed to orient and move themselves. They did not make it deliberately confusing.
2) Excavating solid rock is hard work and takes time. Removing the rubble is hard work and takes time. Workers don’t like to do unnecessary hard work. The boss wants the project finished as quickly as possible. The layout should be as efficient as possible for the users of the structure to have made.
3) people don’t want to waste their time on unnecessarily long commutes. It’s much more logical to have ray-like structure with lots of small areas accessible from a main entry point than a snake like structure with side rooms along a long and winding corridor.

I’d suggest getting some floor plans of castles, catacombs, ruined cities, factories or even battleships and then use one of them for the basis of the dungeon.
If it’s a ruined structure you can arbitrarily make some passageways collapsed to make the party go the long way round to certain areas.

[Tawmis]

Maze-like dungeons are great in theory, but I find that in practice they’re hard to pull off elegantly. Between the constant question of “where am I” and “whats going on?” It makes it hard to run a dungeon any larger than 6 rooms. How the hell do you pull off 20+ room dungeons? Or worse, mega dungeons?

These days, it's with thanks to like Roll20.

Every time I run a cavern or dungeon, and it's got more than a few rooms - I will use Roll20.

Otherwise, it's theater of mine. I just find having a visual when you're doing a large cavern/dungeon to be way easier.

But this is also why large caverns / dungeons are rare.

Well, large dungeons are.

Large caverns can still easily be done with theater of mind, with unique areas.

[Chronicled]

Maze-like dungeons are great in theory, but I find that in practice they’re hard to pull off elegantly. Between the constant question of “where am I” and “whats going on?” It makes it hard to run a dungeon any larger than 6 rooms. How the hell do you pull off 20+ room dungeons? Or worse, mega dungeons?

I have found this to be the very best advice on the matter, especially if running a session over Virtual Table Top: https://graphiteprime.blogspot.com/2...tual-maze.html

[MrStabby]

The "how do you run a maze?" question usually has an answer that handwaves it away: "After nearly two days of painstaking measurment and meticulous navigation you finally find your way to the center of the maze. In front of you is a pair of large bronze doors..."

I figure that my job as DM is to deliver fun. Most of my players don't consider bookkeeping to be fun. We don't tend to track arrows, we don't tend to track 5ft increments people have walked up foggy hallways and so on. Each element of a gaming session should add fun, but more than that, it is competing with every other element that you skip to cover the time spent mapping these things.

Now a big location that isn't specifically designed to lose and confuse - thats a different matter. A virtual table top covers that. Still an imperfect soluton as I find that that it puts players in a combat mindset and this can cause them to suffer cave rage and attack things on sight.

On think I find is to really boost the exploration aspect of the location. Hallways that show portraits of long dead kings, libraries with certain specific topics removed wholesale, statues defaces with the same set of strange runes and so on. Build a lot of interesting detail to pull people in and then later make those details matter. Room after room of enemies quiely waiting in a queue to fight the party as they move through them quickly loses their appeal (well I say quickly, sometimes it does take a few sessions).

[Red Bear]

in my experience, it's impossible

[Thrudd]

The format in which you are running the game makes a big difference. Are you sitting at a table (or virtual table), with some sort of visual representation of everyone's position and the ability to place or draw/redraw terrain features to show where everything is? Do you have the ability to give players handouts?

The way this worked, in the old days, is that exploration was divided into turns (10 minutes long in AD&D), units of time during which the characters were able to move a certain distance. Being usually indoors and/or underground, light sources present determine the distance that can be seen. So, the DM describes what the players can see, the players say where they go and what they do. Then the DM says what they can see from their new position, having potentially moved the distance allowed in one dungeon turn, at the end of that turn. The players may choose to draw their own map to keep from getting lost and remember where things are and where they've been. Some DMs would be nice, and allow players to simply state that they "return the way we came": others might require you to describe your movement precisely, and allow you to get lost if you choose a wrong direction. After a certain interval of exploration turns, the DM is usually checking for random wandering monsters drawn from a table, and/or secretly noting the position of known creatures that might be in movement on their own map, as well as noting any other changes that are impacted by the passage of time (like how long light sources and spells last).

Obviously, this requires the DM to have a precisely drawn map of the dungeon, usually on graph paper. They keep this behind the screen. Everyone has probably seen maps like this. Sometimes, the DM may provide the players with a version of this map, an in-game map that their characters have procured somehow. Usually it should be incomplete, lacking many details, etc., but gives them an idea where they are and where they want to go. It is very easy to draw a map of a dungeon complex with more than six rooms.

If you have the physical table space and/or the money, and/or the time to design it all digitally, you might have the actual dungeon built, which you can slowly reveal with fog-of-war techniques, making it much easier for everyone to visualize. But this isn't necessary at all. You can do it simply with minis or markers for the PCs, to show their positions relative to one another on each turn, a dry-erase grid on the table and/or modular pieces of terrain you can put up. This could be anything really - you could have paper cut-outs in the shape of things like walls and doors, legos, pieces from other games. In this case, everything is described theater of the mind-style until combat starts, or there's an area with an important puzzle or traps that require precise positioning, then you put up the terrain on the table. You don't even need a grid at all- in the old days, most people were using rulers or tape measures, they way they still do in other miniature tabletop games. In the old days, it was often 1 inch= 10 feet. 3e changed the convention to 1 inch= 5 feet. You can do whatever is easiest for the amount of space you have and the size of the miniatures you are using.

AD&D also had specific rules for chases, when parties or monsters flee from encounters, which worked outside the normal combat and exploration movement rules - you didn't precisely track where the characters are when they are fleeing or chasing at full speed, their end position is determined partly randomly, once the chase has ended (either in escape or in a new encounter).
I'd recommend house ruling some version of chase rules, for what happens when characters are running through the maze/dungeon at full speed without taking their time to carefully note their surroundings, for example. Intelligence or wisdom checks should likely have some impact on this, helping smart characters to more easily chase or flee without getting lost.

Going entirely theater of the mind, this type of scenario obviously doesn't work as well. This is when you get people always confused about where they are. D&D was not originally designed to be played that way, even though it became very common anyway, given the inconvenience of always needing supplies and preparation and table space.

If you are entirely theater of the mind, it's best to just abstract it all - at best, you might make some random rolls to decide how long it takes PCs to navigate a maze before reaching one of the nodes of interest in the dungeon, as well as randomly determining which locations they find. For instance, if your dungeon has 10 major rooms with important stuff in them, and it is possible to reach them in any order depending on how the maze is navigated, you might roll a d10 to see which room they find first, and roll a d6 to see how long they wander in the maze before getting to the first room. The number on the d6 could tell you how many times you roll to see if a random encounter happens - there should be traps and monsters wandering around that they could run into, which you roll on a table to determine or choose from a list. If the PCs have some intel on the dungeon before entering, you might allow those with some foreknowledge to make intelligence checks to both navigate the maze sections more quickly and to reach desired locations rather than ending up at a random one. Once they have reached a location, you might reduce the difficulty of finding their way back to that location. Each time they visit the same room, improve the chance of finding their way back there and reduce the time it takes. Maybe the characters even have a map showing a chunk of the dungeon layout (which you don't need to actually physically have or draw) - in this case, they are able to successfully navigate to wherever they want to go with a fairly easy intelligence roll - but, someone needs to have the map in-hand to do this, which means it is vulnerable and whoever is carrying has their hands occupied. Unless they take an action to stow it away, the map might be vulnerable to being burned, torn, eaten by acid, etc, and then they need a much higher intelligence roll to remember what was on the map before it got destroyed...

[Jay R]

There is a simple method for finding your way out of a maze or labyrinth: Touch the wall or hedge with the hand nearest to it, left or right. Keep that same hand touching the wall and keep walking. This may take you on a horribly long route, but it will eventually get you out.

The "right-hand" method doesn't guarantee to get you out. It guarantees to get you back to where you started. If you were brought in unconscious, or teleported in, it might not work. Also, if you are ever moved against your will (teleport traps, trap doors, running away from monsters, etc.) it can no longer be counted on.

[MrStabby]

The "right-hand" method doesn't guarantee to get you out. It guarantees to get you back to where you started. If you were brought in unconscious, or teleported in, it might not work. Also, if you are ever moved against your will (teleport traps, trap doors, running away from monsters, etc.) it can no longer be counted on.

Also, it depends on the shape. To be certain it would work you would need the maze to be on a plain surface that doesn't intersect itself - so no bridges or tunnels.

[Rynjin]

Well I can tell you at least one way NOT to tun a maze. Don't make it part of a combat encounter involving a cadre of Blink Dog sorcerers who spend 8 hours of real time blinking through walls, casting Lightning Bolt, and then blinking out.

Just as an example.

[Telok]

You run a maze... like a rat

Also, it depends on the shape. To be certain it would work you would need the maze to be on a plain surface that doesn't intersect itself - so no bridges or tunnels.

I always throw in a couple stairs, a disconnected interior loop (tho that does mean there's more than one "correct" path), and don't put the exits at the edge. Shifting walls are always a nice classic thing.

But in general if you have players who despise taking notes or mapping anything you don't make a proper maze. Use a node based set of encounters or puzzles, or a flowchart, or just give them a solution like a scroll of find the path or similar (best if its an expendible resource they think they need later in the session). Then make the adventure about what's in the maze or how fast they go through it.

[Elkad]

Adding on to what KorvinStarmast said

From my experience this method works, but introduces random "off by 1 5ft" errors that accumulate. These errors are usually something the PC would not make but the Player did make due to slight miscommunication between GM and Player. As a GM I found correcting these errors was disruptive. Instead I suggest designing the map such that rounding errors don't cause problems.

The PCs WOULD make those errors. Unless they are getting out the tape measures and transit levels and Engineering+Cartography skill checks and spending far more time.

I never correct them. When they open a door into a new room and it won't fit on their map, I'll draw the room on a seperate map for the combat phase, and then let them try to figure out how to fix their own map. Did they measure incorrectly? Did the floor slope and they didn't notice? Is the map accurate and the room is an extra-dimensional space? I don't answer those questions unless they spend the time to investigate.

And I've taken their map away by various means and then chased them out of the dungeon too, making them rely on memory.

[OldTrees1]

The PCs WOULD make those errors. Unless they are getting out the tape measures and transit levels and Engineering+Cartography skill checks and spending far more time.

The PCs can see the distances. That lets them be a bit more accurate with relative distance compared to the players (who are blind relatively speaking). So while the PCs would make some +/- 5ft errors, they would make fewer meaningful errors than the Players would under that methodology.

So all those errors that the PCs would make? I was not talking about those. I was talking about the extra errors the PCs would not make but the Players do make as a result of GM to Player miscommunication about what the PC sees. Those out of character errors accumulate. Ideally either correct them or make them not matter (sounds like you took this method). Let the players focus on the errors their character made instead.

And I've taken their map away by various means and then chased them out of the dungeon too, making them rely on memory.

Oh that sounds fun. In your experience how complicated did the dungeon need to be for this experience to work well without being too complicated? Did short chases vs endurance chases change that?

[LibraryOgre]

So, Hackmaster has a solution to this: The Cartography skill, and levels of mastery, and some metagaming.

Hackmaster has some default levels of mastery... Unskilled means you haven't spent any points on it at all, Novice means you've spent some points and have less than a 25, average is less than 50, etc.

Unskilled people can create a line map
Novice can create a rough, fairly simple, map
Average can draw a fairly accurate map

BUT. If your PC has Novice Mastery, you cannot use graph paper... you have to freehand. If you have an Average Mastery, you can use graph paper, but the GM will only give dimensions once.

[Tanarii]

Give them a piece of graph paper and a pencil.
(This assumes any of the characters have appropriate tools and are mapping in universe.)

Going entirely theater of the mind, this type of scenario obviously doesn't work as well. This is when you get people always confused about where they are. D&D was not originally designed to be played that way, even though it became very common anyway, given the inconvenience of always needing supplies and preparation and table space.

I disagree. The way it was originally designed to play was theatre of the mind for everything except the field of combat. The party was expected to have a designated mapper, and to work as a party to make sure they were accurately mapping what the DM was describing. Things like 10ft poles, 50ft of rope, and a dwarf ability to detect slopes were invaluable tools for properly measuring and calculating distance before committing it to paper. I didn't play that way at the time, but I've tried it since. And it opened up a whole different insight into what the game was intended to be.

Of course, Theatre of the Mind is really a concept put forward for an 'alternative' way to run combat. It's opposed to using a battlemat for combat.

[Algeh]

What I do when I'm trying to communicate with players about a space is that I'll sketch it out on a whiteboard, but sloppily, as they enter the room and glance around. (I don't make extra errors or non-straight lines on purpose, but I don't go out of my way to use a ruler for either straight lines or exact dimensions, just scrawl the thing up there freehand with my limited patience and limited fine motor skills while looking at my GM map.) I'll also note the location of any obvious features in the room with similar half-hearted markings as I describe them. If the players take time to ask me questions, or the characters use skills/abilities/whatever, I erase and refine the relevant parts of the sketch to be more accurate.

This works well for my groups to both represent the not-quite-a-skill-roll "looking around more" that characters might be doing if they decide not to just keep walking straight on through (as shown by the players asking questions like "does that look like a right angle in that corner, or is it really caved in a bit like the sketch?" (I'd also roll for appropriate passive skills either at the start or as players showed interest in a relevant area, depending on the skill and the game system)) and to give them a starting point if they want to map things properly and are using the in-character time and tools to do so. ("Ok, now that you've measured that wall, it's exactly 15 feet long" (erases and redraws sketch with better proportions).)