How To Run A Dungeon - Fixing The Lost Mine of Phandelver
By JimmiWazEre
Opinionated tabletop gaming chap, who’s having to write this on his wife’s laptop because his broke :(
TL;DR:
Lost Mine of Phandelver gives you dungeons but no guidance on how to run them well. Good dungeon play needs urgency, resource pressure, meaningful time tracking, and dynamic encounters. This post breaks down classic and modern dungeon crawl procedures; from Justin Alexander’s traditional dungeon turns, to The Angry GM’s Tension Pool, Goblin Punch’s Underclock, and Dungeon Masterpiece’s encounter tables — and then shows how to use them to make Phandelver’s dungeons tense, reactive, and actually fun to run.
Introduction
Are you trying to run Lost Mine of Phandelver? I ran it recently. Have you noticed how (despite being a ‘starter set’) it does absolutely nothing to teach you how to run a dungeon? Bummer right?! Literally - there’s arguably five dungeons in this module, and it doesn’t show you how to run them at all. In fact, the closest it comes is the final dungeon where it even acknowledges how boring it’s going to be, and weakly suggests rolling for random encounters on a d20 table as and when you feel it’s appropriate.
Do better, Lizards-Ate-My-Toast.
You see, if you don’t know what you’re doing with dungeons, they can very easily turn into this very boring, very samey experience, with your players meticulously checking every tile for traps as they move from room to room, occasionally interrupting monsters that have apparently been sat there for an eternity - waiting to meet the PCs! Meanwhile the PCs have been long resting every couple of encounters to make sure they’re at maximum power all the time. And my God, I’m bored just thinking about it!
The chief cause of this dry experience is that there's no urgency or risk management. So how do you get that, I hear you ask? Damned fine question if I may say so myself, pat yourself on the back. You my friend, should read on, because unless you’re particularly looking for a simple linear gauntlet of pre-defined encounters, you probably need a “Dungeon Crawl Procedure”.
What In The Name Of Sweet baby Jeebus Is A Dungeon Crawl?
Here’s the problem, if you grew up in the 90’s or later, and have only ever played 5e - it’s likely that your only detailed point of reference for what a dungeon experience is like comes from video games - maybe something like Zelda (Ocarina of Time - best game ever made. Fight me!) The issue here is that they teach the player that a dungeon is this linear place, to be solved in a set way, with battles in predefined places. It works in a videogame because of the spectacle and hand eye skill involved.
Unfortunately, this doesn’t translate well to TTRPGs I’m afraid and these games bear little resemblance to what a D&D dungeon is supposed to be.
The fact that people try to emulate these video game experiences is why they fall flat at the table. It’s why your players probably don’t like dungeons, and it’s why you probably don’t like running them.
So, What’s missing?
To run a better dungeon, I advocate for the following components:
There should be no predefined method or route for ‘completing’ the dungeon. The player’s motivations and methods should be their own, and you should expect them to shift as they learn new things and as the situation inside the dungeon develops.
The dungeon should be punishing, and it should be a place that drains resources which cannot be easily replenished whilst the characters remain inside. This could be HP, or light, or spell slots, or rations, or more likely - some combination thereof.
The dungeon should be dynamic. It should move and breathe, and be both proactive and reactive in response the player character’s trespass. There should be opportunities within for all the major pillars of play - combat, social, and exploration.
Time should matter, it should be tracked carefully. Time affects your resources, and the position of dungeon inhabitants, and wasteful players should feel all these factors as keenly as a pin in their arm.
Which brings us nicely onto the “how” part of this post. Well my dudes, you have options. You see, it’s been a hot minute since the 1970’s and quite a few people have stepped up to the mark and developed processes for running dungeons. Here’s a handful of them:
Dungeon Crawl Processes
Justin Alexander - So You Want To Be A Game Master
In his book, Alexander explains a very traditional style. It’s a method that’s as old as the hobby itself, and it sets the fundamentals of most of the methods to follow - tracking time, resources, generating improvised encounters, and the concept of a Dungeon Crawl as a minigame within D&D as legitimate as combat.
Marching Order
Getting players to declare a set marching order for the party up front solves a lot of hassles later on. As GM, now you know which characters are likely to trigger/spot traps, and which are likely to be picked off from the rear.
Dungeon Actions
After Marching Order, the next thing to define are the Dungeon Actions, these are not dissimilar in concept to the actions you can take in combat. Don’t read these as an absolute list, but rather as some common actions, which you can improvise upon as required.
This list includes, but is not limited to the following:
Move carefully (a snail’s pace. default movement speed to reflect the extreme caution of PCs moving through this pitch black, dangerous, scary environment).
Move fast (for when the PCs throw caution to the wind out of absolute necessity, or if they’re backtracking over a recently explored space).
Unlock a door.
Disarm a trap.
Investigate an area (getting more detail about some room feature than has been vaguely called out in the room overview).
Look for secret architecture (hidden doors, traps, pits).
Keeping watch (reducing/removing chance of being taken by surprise).
Casting a ritual spell.
Something else (talking to an NPC, loading your pockets with treasure, helping another PC, lighting a torch etc).
The idea here is that each player does one of these things per Dungeon Turn. Sometimes your players might want to do something so insignificant that you rule that they can have another action. This is fine. Trust your gut.
Dungeon Turn
This is an intentionally loosely defined amount of time - usually ten minutes (this is because ten plays nicely with many timed effects in D&D which are usually roundly divisible by this figure). Don’t sweat the precise granularity of it vs the actions taken in the Dungeon Turn, it requires not overthinking it in order to be effective.
Once the Dungeon Turn ends, the GM performs a bit of bookkeeping on their Dungeon Running Sheet. Justin Alexander features one on his website for you to print out and use:
The idea is that you record each ongoing item or spell effects duration per row, and then for each Dungeon Turn mark a tick in each row (all rows should read the same number of ticks in Alexander’s version, other designs may vary). When a given duration is met, that effect ends.
Whilst Alexander advocates doing this behind the scenes and making it feel less clunky and mechanically like a boardgame, other D&D scholars disagree. Dadi on Mystic Arts describes this process, but instead leans into the idea that the players should be aware of what’s going on behind the scenes to better inform their decisions, and thus debates.
You do you.
Regardless - there is one final step the GM takes as part of their bookkeeping.
Random Encounters
These are so important, and so many GMs are terrified of them in case they “ruin their story”. I say; it’s time to cowboy the chuff up, get comfortable with improv, and embrace the dice my friend! Seriously, stop worrying so much about game balance, and just trust your players to make the right choices to get out of whatever peril the dice serve up for them :)
Random encounters are valuable because they make your dungeon feel alive and dynamic. Rather than the players feeling safe that they can only encounter creatures as they travel to a new area, now, the creatures can come to them. Alexander recommends the following method:
Once you’ve done all your Dungeon Turn bookkeeping, roll a d8. A result of one means that an encounter will occur.
Roll on your pre-prepared Dungeon Encounter table to decide which encounter it will be.
Determine how far away the encounter is by rolling 2d6 x 10 feet.
Unless it’s obvious, make a 2d6 reaction check to determine the attitude of the creatures you’ve encountered. Not all encounters have to be fights.
Determine if one group is surprised by the other, usually Stealth vs Perception checks. Any players taking the ‘keep watch’ Dungeon Action infer a better probability of success here.
The Angry GM - The Tension Pool
One problem with the traditional method described by Alexander is that the only sense of increasing dread comes from resource drain. The odds of an actual encounter remain static, and this effects the psychology of the group if you’re trying to foster a sense of ever creeping doom!
Responding to this shortfall, the Angry GM has a nifty little replacement for traditional random encounter checks called the Tension Pool and it all starts with a glass bowl.
Each Dungeon turn, toss a d6 into the glass bowl (AKA the Tension Pool). This should be visible to all the players. Then:
Each time during a dungeon turn that a PC does something risky or noisy, you roll any dice in the Tension Pool. Results of six prompt a roll on your pre-prepared encounter table.
Each time the Tension Pool fills up with 6d6, roll all the dice in the Pool as above to check for encounters, and then reset the pool to zero.
This is pretty cool, two things are happening:
We’re tracking the passage of time, with a dice in the pool representing a dungeon turn.
The likelihood of a random encounter is visibly impacted by the passage of time and the actions of your players.
That said, not everyone agrees that this is enough, I reckon that with a small modification, you could also use this to abstractly track effects using different colour dice of different denominations. For example, if someone lights a torch, toss in a d8 (or whatever seems right - I’ve not play tested this). The d8 will never trigger an encounter check like a d6 does, but each time the pool is rolled, that d8 might come up with an eight, and if it does - the torch is snuffed out (and the d8 is removed).
This way that’s less stuff to track on a piece of paper, and we’ve also now baked in variance on item and spell effect durations - if that torch goes out, maybe it was a gust of wind? If a spell effect ends, maybe the caster tripped on a flagstone and lost concentration?
Goblin Punch - The Underclock
Arnold K over at Goblin Punch has an entirely different method for using random encounters to build tension. He calls it the Underclock and it works like this:
Grab a d20, a nice big one. Or use a piece of paper, or a paper dial, or whatever you have that can track to 20. This is the Underclock, keep it out in the open so the players can see it.
Starting from 20, each Dungeon Turn the GM rolls a d6 and subtracts the result from the Underclock value.
Results of six on the d6 explode (this means you roll an additional d6).
When the Underclock hits less than zero it triggers an encounter on your random encounter table. At zero exactly, the clock resets to three instead.
If the Underclock ever reads three, a foreshadowing event occurs and the PCs learn a clue about the nature of the impending encounter (naturally, you’ll have to roll the random encounter at that point for your own reference).
The nice thing here is the players are more informed (but not perfectly so) about when an encounter is due. You can represent this as them hearing noises, or ‘spidey sense’, or whatever works for you.
This forewarning means that the players have another interesting decision to make - do they press on, do they try to hide, or do they prepare an ambush instead?
Dungeon Masterpiece - Random Encounter Tables
Baron de Ropp at Dungeon Masterpiece makes an excellent point regarding Random Encounter Tables.
Traditionally, they’re either single die table of possible encounters, or they’re a multi-dice table, which introduces a bell curve only the range of outcomes. Then on top of that you layer distance and reaction.
De Ropp highlights that this structure alone does not do anything to weave a larger narrative together, nor is it scalable, nor does it do much to help the GM to come up with a unique yarn to spin about the specific encounter.
To resolve this, he has a number of tricks:
If you have quests and rumours planned out - seed these into your random encounters. That pack of wolves you just defeated, maybe one of them had a golden arrow buried in its flank. Who made the arrow? Perhaps there’s someone in the woods that specialises in such trinkets?
If your table contains six entries, corresponding to a d6, why not add two more entries to it. Order the tables by difficulty, and then as your players advance in skill, add +1 or +2 to their dice result to weigh the results in the favour of more difficult encounters.
This one’s the real juice. De Ropp suggests adding two more columns to your random encounter table. Behaviour and Complication. You fill these in on a per row basis in a way that makes total sense for that given row. For example - Wolves. The behaviour might be “Hunting Prey” and their complication might be “Their pups are sick”. Here’s the clever bit - you roll three times on the random encounter table, generating a potentially different row per column. You might come up with “Goblins”, “Grifting for Cash”, “Their pups are sick”. This is your improv prompt for the scene, and by combining the elements from different rows, you’ll come up with some really unique encounters.
I’m a big fan of building Encounter Tables this way, and aside from the small amount of extra prep work they take - there’s not really much in the way of downside that I recognise.
Conclusion
So there you go. Whether it’s the Goblin Caves, Redbrand Hideout, or Wave Echo Cave - you’ve now got a detailed set of options for running these dungeons in a way that’s time tested and true. Let me know below the line if you have any other tips for people looking to improve the way that they run dungeons.
Hey, thanks for reading - you’re good people. If you’ve enjoyed this, it’d be great if you could share it on your socials - it really helps me out and costs you nothing! If you’re super into it and want to make sure you catch more of my content, subscribe to my free monthly Mailer of Many Things newsletter - it really makes a huge difference, and helps me keep this thing running! If you’ve still got some time to kill, Perhaps I can persuade you to click through below to another one of my other posts?
Catch you laters, alligators.
Here’s the problem, if you grew up in the 90’s or later, and have only ever played 5e - it’s likely that your only detailed point of reference for what a dungeon experience is like comes from video games - maybe something like Zelda (Ocarina of Time - best game ever made. Fight me!) The issue here is that they teach the player that a dungeon is this linear place, to be solved in a set way, with battles in predefined places.