I was introduced to Dungeon World about 8 months ago, and I’ve been involved ever since.

I was introduced to Dungeon World about 8 months ago, and I’ve been involved ever since.

I was introduced to Dungeon World about 8 months ago, and I’ve been involved ever since. I was a GM for the first time in my life the other day (I even developed a large detailed map), and everyone says it was great. I really really want to create a class now, but I’m not sure where to start or what the steps are that need to be taken. I’m thinking of a specialised druid (like a werewolf or a werebear). I have many other ideas but for now I want to keep things simple. Any help on what to do next would be appreciated.

19 thoughts on “I was introduced to Dungeon World about 8 months ago, and I’ve been involved ever since.”

  1. This is excellent advice by Alex Norris :

    Always start from the fiction. You need a concept that: 

    * isn’t covered by something else; 

    * is wide enough to generate 30 moves (4 starting, 10/10 advanced, 3 racial, 3 alignment); 

    * is narrow enough to have a clearly-defined thematic identity. 

    In practice, this means you need 2-3 distinct concepts mashed together to cover enough conceptual ground for a full base class. Trying to recreate a single book/game/film character isn’t going to leave you with enough moves; that’s what compendium classes are for. 

    Every move you write has to reinforce the class’ theme. Don’t write moves that don’t plug in to the class’ 2-3 thematic identities. 

    Don’t steal moves from other playbooks. If your class concept needs a move from another playbook to be fully realised, you either need to rethink your concept or you can just give your playbook multiclass moves. 

    On that note, don’t use Multiclass Dabbler/Initiat2e, they’re pants. If you’re going to have MC moves, do them in the same style as Inverse World’s MC moves. 

    As mentioned, don’t make moves that are just better versions of basic moves and replace them. If you want to make something that is similar to an existing basic move, make it a move that modifies how a basic move works instead. 

    Don’t build a class around a single starting move, with multiple advanced moves that improve it. Write your advanced moves so the class expands horizontally, not vertically (conceptually wider, not mechanically more powerful). 

    Don’t fall into the trap of trying to straight copy over ideas from other systems, especially D&D. Dungeon World moves have their own tone and their own mechanics. 

    Remember that fiction comes first, and that a move that gives a purely fictional bonus is much more interesting than a move that gives a purely mechanical bonus. 

    Keep in mind what other classes can do when writing your own: you don’t want to make a class that is better at talking than the Bard, or better at fighting than the Fighter. 

    Remember niche protection. Making a class that is better or as good as the above at both fighting and talking is even worse. 

    Keep it simple – don’t include hold mechanics where you don’t need them, for example. If you write a move with options, don’t go above four options to choose from (and three is ideal in most cases). 

    Remember that each type of move (no roll, result, choice or hold) conveys a different tone (no roll means you can always do it, roll for result is for simple actions with same-beat resolution, choice is for complex actions with same-beat resolution, and hold is for moves where you power up and discharge that power in separate beats). 

    On a similar level, know that “choose X bad things that don’t happen” and “choose X good things that happen” are different, not just in terms of tone but in terms of the result too: in the former case, every option you didn’t pick happens. 

    Also: time units. DW doesn’t have rounds; instead, it has “a few instants,” “a moment,” “a few moments,” “a short time,” “some time,” “a while,” etc. 

    Remember that a single move should be a single unit of rules, and should strive to be short and simple. You generally want: 

    * one of the starting moves to be a signature ability (the Psion’s Expanded Consciousness, the Druid’s Shapeshifter) 

    * one of them to be some kind of move the character can use in combat if the signature ability isn’t a combat move (the Bard’s Arcane Art, the Shaman’s Help from Beyond); 

    * one of them to be a short utility move that plugs in to the class’ theme and provides flavour (the Thief’s Flexible Morals, the Barbarian’s Musclebound); and 

    * one of them to be a move about interacting with the world in a unique way – either a social move or one about perceiving things (the Paladin’s I Am the Law, the Ranger’s Hunt and Track). 

    You might also want to spin off the scene-setting part of one of your starting moves into a separate thing to keep each individual move simple – for example, the Cleric’s Deity or the Druid’s Born of the Soil. 

    It’s generally best to think of two stats that your class will use. Try not to mix stats in the starting moves, or if you do, do it as a balancing concern – MAD is a real thing.  

    On the other hand, the fiction comes first, and it doesn’t make huge amounts of sense to have, say, a move about suplexing opponents rely on Wis just because you’ve got a Wis-based class. If that happens, either rethink the move’s fiction so it makes sense, or change it to another move entirely. 

    The second stat comes into play for advanced moves, and lets you offer players the ability to go for a different “build” than default. You only really want a few moves that rely on this second stat (maybe two per tier).  

  2. Thank you so very much this is exactly what I needed. I’ll post my concept character sheet soon. Any idea which themes are popular at the moment?

  3. Thomas du Plessis 

    That advice applies to making a base class, but a “specialised druid” (or specialised anything) is specifically the kind of thing that you’d handle with a Compendium Class instead. It’s a really bad candidate for a full class.

    For CCs, it’s a lot easier:

    First, think of what place the skills/powers the CC deals with have in the setting: are they a secret taught only by one organisation? A boon from a specific god? A curse bestowed by a monster or magic item? etc.

    This will let you determine the entry requirement for your class (“when you are accepted into the Order of the Emerald Howl and inducted into their mysteries, you may take the following…”).

    You then come up with 4-5 moves.

    The first move is going to be the entry move, the one that everyone who wants that CC has to take, so make some effort to make it something that a player would actually want to take – it shouldn’t be something that has a very limited application or is just plain weak, but you don’t want to make it complicated either. I tend to like making these moves that modify existing moves, because that’s a convenient way of reinforcing that having the CC makes you different from others. You can get away with a boring mechanical bonus, but that’s boring.

    You then have 3-4 extra moves. These should be a mix of moves that outright give you new, cool abilities; moves that build on the entry move; and moves that further modify starting/class moves to drive the CC’s theme.

    As an example, here’s my Assassin CC: https://docs.google.com/file/d/0B30fzv28XdrYcUhET21iWnVEcDg/edit

    The entry move is just a mechanically formalised way of getting an assassination target – you could RP it every time, but then it wouldn’t be something that represents “just a job.” It’s a new ability, but it’s not one that’s hugely complicated, and it’s the thing that really makes an assassin an assassin (killing for a living), so it makes sense for every person who takes this CC to have it. It’s also not something where two PCs are going to step on each other’s toes if they both have this move, which is something else to keep in mind.

    Stakeout is a move that modifies a basic move (Discern Realities) to represent your assassin training: you can go on stakeouts and observe stuff from a distance, which lets you Discern Realities with a different stat (because this is more about putting together target movements etc. than intuition, and also because the Assassin happens to use Int as its main stat) and without having to interact with whatever you’re looking at (so you can do it from a roof next door). As an added bonus, you have a new Discern Realities question.

    Requiescat in Pace is a new ability that’s thematically related to being an assassin: you know how to kill people, so you know how to make it look like someone’s dead. It’s a neat twist on the theme and lets you play an assassin with a heart of gold who gets in trouble by faking a target’s death because they were innocent, for example.

    Finally, The Professional expands what you can do with Contract Killer, and lets you use it as a negotiation tool if you can engineer things right – it represents the perks of the PC choosing to be a committed, career assassin.

  4. Ok thank you so much, I think I’m on the right track then. I’ll post my Compendium Class on the group in the same format as yours if you don’t mind.

  5. I think regadles of whether you’re making a new base class or a compendium class, you need to start with a strong idea of what the class is. As with any class-based game, it can be difficult to find “unused” fictional/design space between the existing Dungeon World classes.

    I’m not saying don’t try to make a new base class. But if you continue, you’ll have to find something that isn’t already covered by the other classes. That’s why it is much easier (and perhps a better place to start) to make a compendium class or to make an alternate version of an existing class.

  6. OK, these are my own ramblings. As much as I believe truth is objective and absolute 🙂 , that simply does not apply for creative processes! So these points are how I (me, personally) approach some of the problems, and not intended to cause controversy or debate. I have immense respect for all the people and their opinions above. And what I present here is just that. My personal opinion. 

    1) Steal moves from other playbooks, especially if those playbooks are in the core rules. Because DW is open source, using existing moves will help solidify the code. Think of it as helping to write a computer program. If there already is an existing function that does a specific calculation in the code, just call that function. Writing new code for stuff that already exists is simply bad coding practice. That does not mean you may not modify moves to suit your character’s theme. But if you want your character to do something that already has a move you had better have a damn good reason to write a new move for the same thing. 

    2) Make at least some moves that are mechanical upgrades of starting or low level advanced moves. The reason for this firstly fictional: As the character progresses he gets better at what he does. He could do +1d4 damage if he first spat in his opponents face? Well, after doing it a 100 times he gets better at it, and will eventually do +1d6 damage for the same move. Secondly many players WANT to get better at what they do – this is an important psychological phenomenon that makes people want to play RPG’s. You will read it in all the books on game design, and I see it on a weekly basis at our Pathfinder Society group, where people will sit for hours with their rulebooks trying to get the maximum mechanical bonuses for their characters. Being stronger makes characters cooler. And that is why we play RPG’s: To be cool. The third reason is that purely fictional moves  are sometimes superfluous: There are a lot of purely fictional moves out there that simply tell the player how he may play the character – with no mechanical advantage to the character. So why write the move in the first place when the player could just say “I do this or that”?

    3) Forget about niche protection. The fighter or thief in the core rules are not the only possible fighters or thieves in all DW universes. Somebody made a whole set of alternative playbooks and they are brilliant. (He has a Templar instead of a Paladin etc.) Someone made a “Thief. The Dark Project” thief with fire, water and moss arrows, and it is absolutely brilliant. If your fighter has a different flavour than the core fighter, go ahead and build him. DW can only be richer for it. (But do it with my point 1 in mind: USE THE EXISTING CODE!) It is up to the group to decide which version of a character they want to play in their game. The big question of Compendium Classes pops up here. If your new class is not sufficiently different, then just make a CC for it. 

    4) The place of hold mechanics: In other RPG’s there are rounds. An effect may last “x rounds” for instance. In DW world an effect lasts long enough to “spend x hold”. That means you get to do so many things for the duration of the effect after which the effect expires. A good example is the Druid’s shapeshifting ability. She changes into a bird long enough to accomplish three, or one thing according to her roll. So if you think of “hold” as a means of measuring time, you cannot go wrong. 

    OK, I’m done rambling. And remember I am not right. Find your own truth in this matter, my young padawan. 


  7. The City Thief is explicitly a variant of the Thief and doesn’t claim to be an entirely new, entirely different playbook. It’s just a variant that drops poison moves in exchange for some more thiefy, urban stuff (mostly because I didn’t like the poison moves and I happened to be writing a class based on Dishonored at the same time as I was pondering what to replace them with; Thief was a pretty obvious source to crib from).


    All base classes essentially fit into one to three of these three categories. In general, you should be aware of what categories the playbook you’re writing fits in, because that will influence how you balance it against other playbooks:

    1) it’s meant to be used side-by-side with other existing playbooks.

    Playbooks that fall into this category need to respect niche protection – they shouldn’t be a better Fighter than the Fighter, or a better Skydancer than the Skydancer, or a better Channeler than the Channeler, etc.

    This is what the vast majority of people aim for when writing their own base class. It’s a hard category to be in and it keeps getting harder the more people release full classes. Ideally, you’d want to be aware of every other playbook out there, but that’s obviously not possible, so just making sure you don’t overlap with the core classes is a pretty good start.

    2) it’s meant to replace an existing class in the line-up of playbooks you let your players pick from.

    If it’s meant to replace the Cleric entirely, it’s totally okay for it to be a better Cleric than the Cleric, but it needs to own this and cover at the minimum all the bases that the Cleric would normally cover in terms of party role. It’s also not okay for it to be a better Cleric than the Cleric and a better Fighter than the Fighter, because that’s violating niche protection.

    Jacob Randolph’s Templar, Mage and Priest are all replacement classes. They’re not meant to be used alongside the Paladin, Wizard or Cleric, they’re meant to replace them completely.

    3) it’s a variant class.

    You changed part of an existing class to make it slightly different from its original incarnation, but it fills the same core roles because it’s mostly the same class (i.e. the City Thief vs. the Thief). This is really more of a specialised case of #2, so the same caveats apply.

    The Templar, Mage and Priest are also variant classes, because they reuse a big chunk of what they’re variants of. The Mage and Priest are variants that essentially throw out the engine and replace it with an entirely different engine that works approximately the same, though, so they’re a bit beyond the scope of your average variant class.


    Basically, the thing to remember is that a standalone class is going to be “competing” against other classes in play, in terms of mechanical power, fictional power (spotlight) and fictional space. It will exist alongside a number of other classes, and it shouldn’t be trying to be better at what they do than those other classes are.

    Early on, I wanted the Shaman to speak to natural spirits as well as ghosts, but that’s a Druid thing – so I ended up making the Shaman purely about communicating with the afterlife, because that way it wouldn’t overlap with the Druid.

    By contrast, the City Thief isn’t meant to be played alongside a Thief, so it’s okay if the City Thief is slightly better at stealing things. You’re not going to have a case where the City Thief and Thief players are both trying to do the same thing and the City Thief is plain better at it so the Thief player ends up feeling disappointed and frustrated.

  8. Sorry, just one more thing:

    Decide on your character’s role: Tank. (Fighter, Paladin) Secondary fighter with other skills. (Ranger, Thief, Cleric) Support Character (Bard), Magic user: (Wizard.) Then balance your characters fighting and other skills according to its role. If your character is a tank: He has to do d10 damage and have significant armor. The core fighter has the option to choose moves that can give him a potential 2d8 +1d6 + 2 damage with one blow. (If I remember correctly) So make sure you do not make a pansy tank. He has to be able to compete with the core rules fighter. The same goes for other class roles: Balance your character against its role model in the core rules. 

  9. Alex Norris Hehe! Its like you read my mind while I typed my second post. I have to agree with everything you say here! (We posted almost simultaneously.)

  10. Wynand Louw: dealing damage totally isn’t part of being a tank – damage is its own separate thing!

    The Fighter is meant to be the king of damage, so you really shouldn’t be making a class that can match his damage output unless you’re writing a class intended to replace the Fighter. That d10 damage and all those +damage moves are essentially the Fighter’s niche. Leave it alone.

    I’m also firmly of the opinion that no one else should have a d10 base damage, especially not the Paladin, but that’s another matter.

    Also, “magic user” really isn’t a role. Part of the problem with certain other RPGs is that they let spellcasters do everything under the justification that it’s magic.

  11. Alex Norris As a group should normally have only one tank, the character that is the tank should hold his own as the tank. That is the whole point of roles. There should be a choice of tank character playbooks to choose from. I assumed thats what you meant with your post on alternate and variant classes… which I totally agree with.

  12. Tim Franzke Of course a game with 4 fighters can work, but that is not normally recommended. The classes are supposed to support each other. But you are absolutely right. 🙂

  13. The first thing I’d change (and one of the only) in a DW 2nd Edition is to change “THE Fighter” into, simply, “Fighter.” Unique classes work in AW, but aren’t necessary in DW.

Comments are closed.