New GM resources | Game design

Advice for any new Dungeon or Games Master, sourced from dozens of campaigns with hundreds of players over thousands of hours and still making new mistakes. This wisdom to new GMs is broadly split into three categories – preparing to run a campaign, writing an adventure, and running the session.

Some mechanical examples are specific to D100 RPGs like Dark Heresy, Rogue Trader or Imperium Maledictum, but the principles still work for other systems like DnD or Starfinder.

Fail quickly, keep the game moving, and have fun!





Keep your resources handy but manageable. Too many notes is as bad as not enough notes.

Session Zero and safety tools

There’s dozens of video essays that explain it better, but it’s vital for any group (even people you’ve known for a long time). Expectations have to be managed before the first dice are rolled – what do we feel about character death? How explicit do you like your body horror? How will your group handle intimate or sexual encounters? How is your edgy lone wolf rogue sniper expecting to get along in a campaign about political intrigue?

Ask (or bribe) a trusted player to be your Assistant GM

Depending on your group, you could rotate this responsibility between your players or ask for volunteers. They can look up rules while you keep the game running or keep track of secondary effects in combat such as damage conditions, spell durations, environmental modifiers, or which cultist goon had their hand blown off by a grenade.

Memorise an average stat

You’ll never need to write down a friendly NPC statline ever again. In D100 systems like Dark Heresy or Imperium Maledictum, an average value for a skill is around 30. If you ever need to call upon a random citizen’s stats, you know they’ll be testing on around 30 for something they’re competent at. An experienced NPC might have +10 or +20 to this.

You don’t need to know every skill, talent, trait, or piece of wargear a black market dealer might have, only that their Barter skill is 40.

Don’t plan in too much detail

Bullet points are more effective for scan-reading than whole sentences, and much of what you have planned will be missed or skipped over by players. Keep it in the back pocket for another adventure.

Work backwards with the time you have

How do you want to end your next session? If you’re in the middle of an investigation, give yourself a few convenient stopping points between key exposition scenes. If you’re about to kick down the mob boss’ door, make sure you’ve got an encounter that can expand or contract to fill the time you have.

You’re like a TV Producer putting together a live show on the fly – be prepared to cut certain segments or have a few ad breaks (such as random encounters or recurring NPCs) so you can keep to schedule. 

Have a handful of generic names and taverns on standby

Everyone’s first stop for clues is the pub, and everyone wants to know the name of the trader with the weird accent you made up on the spot. If you have a particularly dangerous dive bar in the back pocket, you could end up in an encounter that helps push the story in a productive direction (see Chekhov’s Goblin below)

Make a Battle Post-It

This could be any easily-accessible piece of paper that lists the page numbers for anything that might come up in a combat encounter that you can quickly refer to – critical hit charts, weapon stats, environmental effects, conditions, and so on. You can ask your assistant GM to look up the rule while you get on with running the next part of the encounter.

Keep it secret, keep it safe

Keep your notes in a folder that you can close when not in use, or have a dummy map or piece of card handy to cover your notes. You don’t need to use a massive GM screen to conceal everything you do, but you should try to keep your surprises safe from wandering eyes.


No plot survives contact with the player characters. Bend your story so it doesn’t break. 

Start with a bang

The classic star wars opening, or in medias res, gives your players a chance to flex their new characters in a safe environment. Create a contained space, populate it with subordinate goons of one of your villains, and give them a compelling or strange motivation that will encourage your players to investigate once the fight is over.

Perhaps you’re resting overnight in a secluded watering hole and some masked strangers kick down the door demanding to know where Fingers McGee is, or your train is heisted by some hoverbike bandits, or you’ve rounded up some thugs on a routine patrol when one of them opens a portal to the beyond with a strange artifact and unimaginable horrors pour through.

Using a transitory environment means you can quickly move on to introducing the plot via key NPCs, patch the players up, and set them on their mission proper. If you’re worried about how to balance combat in your game, this gives you a safe opportunity to test how combat capable the party is, knowing you can patch them up immediately afterwards, and you can adjust future encounters accordingly.

Playing an unfinished adventure is always better than not playing an adventure because it’s unfinished

You’ll do yourself a mischief trying to work out all the details before sitting down to play – it’s much better to let the finishing touches to the adventure be collaborative. Your players will treat you to great ideas and unique experiences as dice are rolled around the table, and it’ll be much easier to slot those ideas into an adventure whose ending isn’t fixed.

Do not set your plot in stone

Keep your story in mind, plan your set pieces, breathe life into your villains, but don’t concern yourself with how or when they’ll crop up. Let your players meander between the leads you’ve given them and you can pick up on the leads they create themselves, dropping in your set pieces when appropriate. This creates a collaborative, dynamic encounter that is rewarding to GM and player alike.

For example, they could be investigating mutilated livestock, but rather than going to the farm (where you have cultists waiting to perform a ritual), the players decide to go to the records office to work out who bought and sold the livestock. You don’t have anything prepared for a records office, but you can sprinkle in some clues that the farmers haven’t filed their taxes in a while, and the cultists you had planned for the farm instead turn up to seize and burn any records of their activities. Your encounter still gets used, your plot points still get picked up, and your players get a big payoff for using their wrinkly problem-solving brains. 

Chekhov’s Goblin

You want to build a sand funnel rather than a sand box, and combat encounters are at their best when they nudge players in the right direction. Fights can take a lot of real time to resolve and be a large drain on resources or player morale if it goes badly. Random encounters can exhaust players, or even send them in an opposite direction to the adventure (“Those mutants in the road must be related to the corrupt mayor somehow!”). If you don’t have a reason for that combat encounter, your players will fill in the gaps.

You can’t predict exactly what players will end up in a fight with, so keep a few loose plot threads handy (see above, do not set your plot in stone) that you can weave into the encounter. Perhaps the merchant guard who tried to grapple the light-fingered rogue also bears the tattoo of the cult they’ve been following – he could be interrogated for clues! If, that is, he doesn’t whistle for backup from his mates in the sandwich shop over the road…

Have a conduit character

These are sometimes referred to as ‘GM Tools’ because they allow you to talk directly to your player characters, and are strongest when used as an intermediary between players and their patron or quest-giver. They can provide missed leads or vital information that would otherwise be unavailable to the party without their boss constantly butting in and telling them how to do their job.

They are best as non-combat characters – this avoids players trying to coax them into a fight and highlights the importance of knowledge-based skills their own characters might not have. Keep them in the background, but don’t be afraid to kill them off during a dramatic home base raid or stab your players in the back as a double-agent.


Rulings, not rules, keep the game moving.

Fail quickly

Things will go wrong and you’ll need to make a call whether to stick with the mistake or switch back to the correct version. There’s no right or wrong answer, the only imperative is do it quickly and keep the game moving.

Sticking with a mistake can be an exciting improvisation opportunity, especially if your players don’t realise you’ve made a mistake. Perhaps an NPC is somewhere they weren’t supposed to be, or a player had been accidentally using a skill they didn’t really have – could it have been a doppelganger, or an elaborate part of the evil plot?

Likewise, switching away from a mistake helps players understand they’re in an environment that it’s safe to make mistakes. Perhaps you got a key rule wrong during combat or gave someone an unfair advantage – stick with the mistake for the session and say you’ll discuss it after the game to see what everyone’s opinion on it should be. Make your ruling to keep the dice rolling.

Don’t let the dice control the story

You should only call for a dice roll when there is a failure state for an action. Holding back an important plot point behind a failed dice roll really sucks unless there are other opportunities to find that plot point (see above: don’t set your plot in stone). In the same vein, if someone wants to perform an action their character is very capable of doing in a zero-pressure environment, just let them do it. They might want to pick a lock in an abandoned church – either allow them to pick the lock with no roll, or before they make the roll give them an idea of what might happen if they fail: “There’s no guards about so you could take your time but the locks are old and brutal, designed for hard iron keys – if you fail badly, you could break your delicate lockpicks”

It’s better to finish a fight in the same session you start it

If you split a fight in two, you’ll lose a lot of the momentum you’ve created throughout the session, it will be an unsatisfactory end to the evening, and you have to set the board up twice (as well as making painstaking notes of where everyone was standing and who was on fire). If you can sense a fight coming up towards the end of a session, try to fill the remainder of the time with non-combat activities. Encourage players to scope out the battle zone, acquire gear, lay traps, use stealth, or interrogate witnesses. It’s always fun to open a session with “Roll for Initiative!”.

It’s not Players vs GM, it’s Players vs Problems

You are not an all-powerful deity that the player characters are trying to defeat, and you shouldn’t punish players for doing something you don’t like. There are many other articles on the internet about the importance of communication between players and GM, and much of it can be tackled in session zero before the game even begins.

Your principle task is to get players to inhabit the world you’ve created and deal with the conflicts in a collaborative way that makes a compelling story. Don’t feel shy about rolling your dice out in the open – twists of fate or tragic demises are easier to stomach when players can see exactly what lead to them. 

If you have a bad poker face, make it a melodramatic one

Sometimes characters will lie to players, and you don’t want to give the game away. Use the classic lie detector test busting technique – pretend you’re lying when you’re telling the truth. Try sprinkling in some evil grins when delivering regular information to throw players off your scent and them focussed on the characters, not on your acting abilities. Remind them about in-game perception or intuition tests to scrutinise NPCs for signs of falsehoods, rather than your own.

And finally, the most important rule:

Have fun

You’re a player too!