I've run a heavily homebrewed Lost Mines of Phandelver campaign for years now — multiple line-ups of friends, session logs from 2023 and 2024 and prep notes going back to 2022, a binder of NPC dossiers and reworked locations thick enough to be embarrassing. Somewhere in there I started writing the why down, because that's what I do with everything I love: find where the common wisdom is wrong, build the better version, write the method down. The result is a pile of notes I call my "Golden Rules," plus a series of design essays titled, with no imagination whatsoever, "Better X in D&D."
This is me pulling the best of it into one place — not a beginner's guide, but the answer to a narrower question I keep circling: when a session is genuinely, electrically good — when a whole table forgets their phones exist for three hours — what is actually happening? It's almost never the thing people think. Not the lore, the dice, or the voices, though I do all three. The peak is structural: a few rules about making a world that's already alive before anyone rolls initiative.
A confession up front: I'm a better designer than I am a finisher — my favorite teaching one-shot is abandoned at problem two. So read this as the part I'm confident about, from someone still learning to finish his own ideas.
Action is boring; drama is not
Here's the heresy everything else hangs off of:
“Action is boring. Good action is actually good drama and story revolving around the stakes of the action.”— my Golden Rules
Most D&D combat is two columns of numbers taking turns shrinking: the orc hits you, you hit the orc, somebody tracks hit points, and forty-five minutes later the orc is dead and nothing about the story has moved. I even critique the published module I'm running for being "very combat heavy and boring" — which feels rude to say about an official adventure until you've watched a grind fight send the whole table reaching for their phones.
So I run combat as the rare, expensive, peak thing it should be:
“Every combat should secretly be a puzzle to solve, with the players abilities providing possible solution… Combat should be rare and peak.”— my Golden Rules
The reframe is small and changes everything: a fight isn't a thing the players do, it's a threat that builds tension and comes crashing down when it unfolds. The puzzle's solution is hidden in the players' own abilities — so the best combat is designed around your specific party. The monster does something weird they have to solve while it's trying to kill them; the question isn't "how much damage do I roll" but "what is this thing and how do I beat it." The goal is comprehension.
I learned a lot of this watching anime critically. Some Demon Slayer and Jujutsu Kaisen fights have gorgeous animation and go completely flat, and you can feel exactly when: the stakes weren't set up and the choreography is all there is. The fix is to make the enemy known but not fully known before the fight — a silhouette, a rumor, one terrifying demonstration. And the corollary nobody wants to hear: be willing to delete combat. The instant the fun part is solved, stop rolling and move on. As my note puts it, if slogging, let 1 turn count for 3 or 5 and decide the encounter. Which generalizes into the rule I run my whole table by.
Skip the boring parts (but keep the weight)
You are not obligated to play out anything boring — not the slog of a random encounter, not the puzzle that died on contact, not the conversation where the players will never ask the right question. You're the DM. You can just skip:
Just roll a check and skip thru it — you beat them taking some damage, or you realized the solution is this. Conversation doesn't have to linger if the players can't get the right info — just say yada yada, you chat for a bit and the key thing you get out of it is xyz.
The catch people get wrong is that the skip has to stay consistent with consequences, or you quietly teach the table that nothing matters. In my own words: a boring but high risk fight should deal significant damage even if skipped. You're compressing the runtime, not refunding the stakes. Skip the tedium, never the consequence — because consistency between choices and outcomes is what makes the world feel real, and that realness is the whole product.
This is also where the painter's metaphor lives — the rule I'd tattoo on a DM if I could give them only one:
“like a good painting, imply detail, worldbuilding, intrigue interest, dont waste too much time showing it.”— my Golden Rules
You can go slow and lush on one room of the haunted house and skim every other in a sentence, and that's not just allowed, it's better. A good painting doesn't render every leaf; it renders three so convincingly your brain fills the tree. Worldbuilding works identically — let one vivid detail stand in for fifty you never wrote.
The world moves whether they do or not
The biggest upgrade to my DMing was deciding the story should always progress — but never along a track I'd pre-laid:
There exists a set with villains and goals and factions, and if the players dont intervene the story will keep on chugging a certain direction. Whether the players make decisions that intervene with it or not, they should always be progressing their own story.
This is the difference between a railroad and a living world. I'm not running a fixed plot — I'm running a situation: factions with goals and momentum that grind forward on their own. The players don't follow a story; they perturb one already in motion, and never sit in a world that freezes whenever they leave the room.
The trick is one rule I steal for every NPC, more important than any backstory I'll ever write: each one needs something specific they are doing right now, today, that they'd keep doing if the players never showed up. Not a history — a present-tense action with momentum. Backstory is what you reach for when you can't think of what a character wants this week, and the wanting is what makes them feel alive.
A motto on my campaign design doc captures the whole philosophy:
“Everything should always be going wrong in interesting ways even when it goes right.”— my campaign-design note
In practice this produces things no one prepped. A band of Redbrand bandits the party recruited into a makeshift town guard quietly grew into a thirty-person militia that outgrew their control — never on any sheet, just a consequence compounding into a problem. The fun isn't in the plan; it's in letting it collide with players who do things you'd never script, and treating every collision as canon.
The narrative engine behind all this — the villains, the tragedy structures, the worlds themselves — lives over on Worlds & Stories. This essay is about the machine that makes those worlds move at the table.
The Three-Faction Rule: Good, Bad, and a neutral Ugly
This is the framework I'm proudest of, distilled from a Mystic Arts video by a creator who goes by Di, then beaten into a shape I run. The problem it solves: a political situation is easy to build but feels impossibly complex to unravel — exactly the asymmetry you want, because a little prep buys you hours of player paranoia.
Every conflict gets built from the same hidden template — three factions:
- Good — the players' side. It starts out losing.
- Bad — the clear threat, naturally at war with Good.
- Ugly — the neutral third. Messy, self-interested, complicated. It doesn't fit good or bad, and it's the tipping point.
The asymmetry is the engine:
“Bad beats Good unless Good can win over — or neutralize — the Ugly faction. If the two big sides start evenly matched, you're really telling players the adventure hasn't begun yet.”— Better Factions and Politics in DnD
That last line took me embarrassingly long to internalize. Evenly matched sides aren't a story — they're a stalemate, with the players as spectators. Good has to start on the back foot, because that's the only thing that makes the players themselves matter: they're the ones who can tip a losing fight by swinging, or sidelining, the Ugly faction. From there the drama writes itself. The Ugly faction can decide the war, the Good side is too proud to court it, the Bad side will ally with anyone. Will the duke swallow his pride to win the wizards over? Different NPCs disagree, loudly — and that disagreement is the politics.
Building one takes six quick steps. Steal a simple conflict (monster X wants the land, civilization wants to keep it). Name three factions and add the neutral third. Cast each with faces, deliberately giving the players' side weak leaders so the heroes are genuinely needed — a senile old duke, his useless heir, a brilliant ruthless enemy general. Give each one simple worldview, then define its methods, because how a faction acts is where the friction lives: the slow-deliberating duchy, the fast-and-bloody warband, the wizards who'll summon a power that fights for whoever recruits it. Then twist it.
Three factions is a lens, not a limit
Here's where it gets genuinely fun. The good/bad/ugly split isn't a cast of three — it's a pattern you can drop anywhere.
“Three factions is a lens, not a limit.”— Better Factions and Politics in DnD
Drop the same split inside each faction. The players' kingdom has its own good/bad/ugly: the warmongering heir (bad), the clear-eyed daughter who holds no power and so funds the party in secret (good), and the brilliant, senile duke who takes no stance (ugly). Then split those — the army gets a reckless commander, a cautious one, and a general stuck between them; the council fractures the same way; the enemy nests too. You spiral it out anywhere your attention lands, from one simple rule.
The payoff is the best trade in all of DMing: it's trivial for you to track — all the same three-beat pattern — yet impossibly intricate for the players to map. Around the third nesting they stop seeing the template and start treating your politics as a real, tangled, unknowable thing — which it now effectively is.
One variation I lean on: for morally-grey play without the misery, make every faction bad but seed a good subfaction inside each — so the whole world is corrupt yet there's always something worth fighting for. I avoid pure grimdark on purpose; players need hope, or they disengage.
I've come to think of it as the five-room dungeon of political play: one reusable template, infinite arrangements. The mechanical, systems-design half of my table — the 2d6 bell-curve random tables, the house rules — lives over on Games & the Table.
Stakes the players carry in their own chests
Frameworks move the world, but the world only matters to a table with a reason to care, built into the characters before session one. So every PC gets short-, medium-, and long-term motivations, whether the character is aware of them or not. My favorite example, smuggled into my notes with the most honest parenthetical I've written:
a character who cares about being strong might subconsciously work towards the goal of fighting the strongest thing to prove their superiority, even if they wouldn't admit this as a goal — me irl.
These goals are the first thing I design, before race or class or stats, plus a shared goal on top of the personal ones — a real reason they're all here together. I have a hard rule against loner characters and solo players, and it's not taste, it's structural: the entire point of this hobby, for me, is to tie the in-real-life group together, and a character who works against the party has no reason to be with it. Early tensions are great, but they have to be an arc that resolves, not a permanent wall.
And early — ideally in session zero — I plant a premonition. Something terrible is coming: the destruction of the whole town, say. The players don't know how or when, but their characters have reason to want to stop it. That single seed does an absurd amount of work, because now every faction, rumor, and dragon-shaped shadow carries the same low hum of dread. Is it the goblins who'll burn the town? The corrupt order? The thing the old wizard found? The curiosity never resolves, so the tension never drains.
The last piece is performance — cheapest to describe, hardest to do. Talk slowly. Take it seriously when it's serious: let the table joke and laugh, then ground back into consequence with slow, deliberate, attention-demanding talk. Watch how Brennan Lee Mulligan or Matthew Mercer can be cracking up with the players one second and snap back into a deadly-serious tone the next — it pulls everyone back into what's happening. Your seriousness, at the right moment, tells the table this fake world is real enough to matter. That performance side of me — the voices, the timing, running improv for six people at once — is really just the Bard's engine pointed at a table.
Why I keep doing this
Zoomed all the way out, every rule here says the same thing: make a world that's already alive, then get out of its way. The DM's job isn't to tell a story; it's to build a machine that generates one, in real time, with the players' fingerprints all over it.
But here's the thing the frameworks don't capture, and the reason I'll keep running these games long after I stop finding new rules to write down. D&D, for me, is the one place every interest I have shows up at the same table at once — worldbuilding, probability, anime-fight analysis, the painter's eye, performance, teaching. And underneath all of it, the part that's ever really mattered: it gets people I love into one room, forgetting their phones exist, building something none of us could have built alone. The factions and the combat puzzles are scaffolding. What they hold up is connection.
If you run games too — especially if you'd argue with half of this — that's exactly the conversation I want. Come say hi, or wander into the worlds these rules hold up.