Hoo boy, if there’s one thing the internet is real good at, it’s telling DMs how to DM, and how to DM better. Does civilization really need yet another pile of DM advice?
Beats me. Fortunately I’m not responsible for civilization, only an infinitesimal slice, mostly myself. And this is a compilation of the advice, principles, and hacks that I’ve found most useful. I’ve written this in the first person, so it sounds a little less bossy or prescriptive, but almost all of this advice originated elsewhere (mostly from Zak Smith, Matt Colville, Ginny Di, and Mike Shea). If it doesn’t work for you, in whole or in part, that’s fine too. Everyone’s game is different, and this post is in part a list for me to review regularly, and a thought archive for my future self.
A lot of this is general, some of it is D&D specific — take what works for you and ignore the rest. If something strikes you as unwise or even wrong, or you have additional gems to share, I’d love to hear about it in the comments.

Some sweet art from Matt Colville’s Flee, Mortals! book of monsters.
RUNNING THE GAME IN THE MOMENT
Desperation in my players is good. Gross and disturbing stuff is good. Players who are scared or grossed out — players who are feeling something — are invested.
I don’t worry too much about detailed monster stat blocks or about running monsters “correctly”. What I need: AC, HP, damage, any special abilities; everything else is fungible. Bonuses to damage can be average or minimum PC level (e.g., vs level 5 PCs, monsters have +5 to hit and weapons do [die]+5 damage).
I use HP ranges for bosses — but if I go high, I reward the PCs accordingly for beating an extra-tough boss, with extra XP or loot or both. Don’t worry that anyone else at the table is tracking the monster’s hit points; trust me, they’re not. It’s hard enough for the DM to keep track of it, and the DM has to.
Opponents might negotiate, they might surrender, and they might run away. Not every fight has to be to the finish.
I try not to stress myself out with planning or second-guessing. I give the PCs some bad guys to fight or negotiate with, find a map of a place for them to meet up, and that’s an encounter. Don’t forget to build in some kind of reward or payoff for the PCs, information or loot or both — this is a frequent blind spot of mine!
Raymond Chandler: “When in doubt, have a man come through a door with a gun in his hand.”
RPG version: “When in doubt, have some armed schmucks — or a monster — roll up and insert themselves into the situation.”
The “Hunter/Hunted” model straddles the line between session prep and in-game running. If the PCs don’t stop the bad guys, the bad guys enact a little more of their plan, the next fight gets a little harder, a little closer to home, or both.
For one-shots: if time’s running short and I don’t have time to get the PCs to the boss, I have the boss come to them. Or if the PCs are messing around not making progress. People get real focused when Darth Vader kicks down the door.
If a PC wants to do something not covered by the rules, I come up with fair odds, communicate the odds and the consequences, and let them roll for it. “I think that has about a 1 in 3 chance of working. So roll a d6, and if you get a 5 or 6, great, you did it. But if not, not only do you fall off the wall, but the guard is 100% going to hear you.” D10s are great for slicing probability into 10% increments (in fact I am very surprised that there haven’t been more D10-based systems in the history of RPGs [not counting full percentile systems]).
I let players narrate the result of a critical hit, a killing blow, or just any really impressive roll.
SESSION PREP
Once a campaign is up and running, I don’t need a lot of pre-cooked story (this paragraph is a little less relevant for one-shots and intros). I try not to plan or know in advance how the story is going to go in the session. As stories occur to me, I try to make them backstories: I invest them in the villains and their origins and their planned-but-not-yet-enacted plots.
Adversary goals and desires are good. Bad guys want X and will do Y to get it and will do Z to get away if necessary. Villain backstories and larger plans are great, because they tell me what the adversaries will do when things go off the rails — as they inevitably will.
When planning a session, I have a story idea that I hold loosely, or at least an inciting incident and some ideas of where things might go, but I try to think in game mechanics terms — what encounters will occur — what groups of NPCs and adversaries in what environments, and what information will I need at hand to run those encounters? Monster cards, spell cards, stat blocks, special abilities. Maps! Thinking in terms of game mechanics helps me prep.
At the table, in the moment, I think in story terms, not in game mechanics terms.
– Discover a story with the players, rolling and calling for rolls when necessary, as opposed to using the rolls as wheels for advancing a pre-planned story. Roll dice when the outcome is in doubt, or to focus player attention.
and
– Whatever my players choose to do, that’s the story. As long as everyone’s enjoying themselves, that’s the game. The goal is for everyone to have a good time, not to hit particular story beats.
The longer I do this, the less story I plan for each session, and the more I’m comfortable running long stretches with little to no rolling. On the other hand, players like to roll dice, and sometimes I call for lots of rolls just to keep everyone engaged and on their toes. Can be handy when things are dragging in that fifth hour of play but we’re not quite to a good stopping point.
Do plan for some loot: gold, weapons, armor, magic items. And then every few sessions, give the PCs an opportunity to spend some of that loot. Players love to go shopping.
Ginny Di’s DM says don’t write the story for the players, write the story that will happen if the players do nothing. I love that.
Recycle old PCs or book, movie, and video game characters as NPCs and adversaries (also works on the fly, in the middle of a session).
Finally, I think character generation should be ongoing, in that players should feel free to add bits of backstory or other refinements to their characters as the campaign goes on. And I actively try to dissolve the boundary between character generation and worldbuilding. In the campaign design document I created for the Saturday Night D&D group, I wrote:
If your character is trying to find the Goggles of Clairvoyance, escape the Sisterhood of Divine Poisoners, or ask a question of the 1000-year-old druid Dya Eterna, then those are things that now exist in the campaign world.
WORLDBUILDING
Anything that I or my players know about the world other than what has come up in play is approximate, possibly incomplete, and subject to ongoing revision.
Have ideally just one Central Weirdness, from which all other problems flow.
Central Weirdness -> subsidiary weirdnesses -> knock-on effects -> those jerks over there:
- is basically an org chart for the bad guys, and an ascending scale of adversaries for the PCs to encounter;
- all of which outside of the Central Weirdness are fungible/swappable/replaceable, which has two salutary effects: I can dangle high-level lieutenants as nemeses for the PCs, but if the PCs manage to kill one, another can take its place (ideally the new baddie is a little tougher, and weird in some unique way, so this isn’t just musical chairs);
- AND (nearly) all fights or investigations eventually lead back to the Central Weirdness, which gives the campaign coherence and direction.
Along those lines, noir-ish crime stories are good DMing inspiration:
- usually more than one, maybe multiple Weird Things turn up, with no obvious or even plausible connections.
- as each Weird Thing is investigated separately, links start to appear — these will generally be surprising when revealed to the audience (i.e., the players).
- eventually all threads lead to the one Central Weirdness that set everything in motion.
- if it’s dragging on or getting too complicated, I can simplify things behind the screen; maybe instead of a long chain of cause-and-effect, Minion X did Deed Y because the BBEG just told them to, for its own mysterious reasons. Cthulhu gonna Cthulhu.
Having a couple of historical events to refer back to, to shade current events and give shape to the world, is really useful. For my current campaign, there’s the Demon War centuries ago that brought down the old empire and basically broke civilization, which is a continent- or planet-spanning world-building event. Then in the Vale specifically there was the Orc War 30 years ago, that shaped a lot of how the Vale is now — low population, fortified settlements, and not many orcs or goblinoids of any kind around.
Why I picked the Nentir Vale (the default setting for D&D 4E) for the Saturday Night D&D campaign: defined geography and factions, not immediately familiar to people who had played 5E before, and I could just drop the PCs in somewhere, give them a few rumors to chase down, and let things start pinballing.

The cover of The Book of Concealment, my favorite DM notebook. Everyone should have one or two of these things.
AND FINALLY…
Write everything down. Probably 50% of my baddie plots occur to me on long walks with my dog. And there are things I planned before the first session of the current campaign that still have yet to happen, or be discovered. But they’re in my notebook or Google Doc, biding their time.
















































