As the Dungeon World Turns

Dungeon World (Sage Kobold Productions 2011, 2012) is probably the best-known descendant of Apocalypse World, the family of games “Powered by the Apocalypse” (PbtA for short.) It owes its popularity not only to its use of the most popular trappings in the history of role-playing (the Gygaxian dungeon crawl and character roles) but also to the clarity of the writing and some simple but effective choices of mechanics.

One of its significant modifications to the AW model is the way characters gain experience and advance. In AW (including the second edition previews released so far) and several of the PbtA games based on it, one player and the GM each highlight a stat on your character sheet at the beginning of the session; your character gains experience by using (i.e., rolling dice based on) the highlighted stats. I really hate this; typically, the stats highlighted are not the ones that interest me as player. In addition, a few specific character moves may instruct you to mark experience under limited circumstances.

In Dungeon World, there are no highlighted stats; instead, you mark experience every time you roll a failure (6 or less on two six-sided dice.) In addition, the End of Session move rewards the behaviours typical to the Gygaxian inspiration: playing your character’s alignment, learning something new and important about the world, overcoming a notable monster or enemy, and looting a memorable treasure. Between these, you can expect characters to earn two to six experience points per session.

But there is one more thing that can earn you experience, and it’s very unexpected in light of the old-school model DW emulates: relationships between characters. Here is how it works:

• Each character template (e.g. Fighter, Cleric, Thief, Wizard, etc.) includes a choice of four bonds framed as mad-libs, such as “________ has my back when things go wrong,” or “________ knows incriminating details about me.”
• At character creation, bonds should be discussed among the group; every character must take at least one bond, and can have up to four. You can take multiple bonds with the same character.
• As part of the End of Session move, choose one of your bonds that you feel is resolved. If the player of the character with whom you have the bond agrees, mark experience and write a new bond with whomever you wish.

There has been a good deal of discussion about this, and in particular how it plays out in one-off sessions where players have less invested in the group of characters. One criticism is that some of these bonds are explicitly written as antagonistic and this may result in dickish behaviour with players you don’t know, for example at conventions. Ryan Macklin has written his feelings on the matter and I agree with his solution⁽¹⁾ — use the mad-libs as inspiration, not prescription. Rob Donohue suggests the use of flags instead of bonds for one-off games. Others prefer randomly generated bonds from a general list.

Relationship map after Episode 3

But I’ll tell you, for a continuing campaign with trusted friends and skilled players like Sean, Karen and Adrienne in our Land of Ten Thousand Gods saga, these bonds are pure gold. For one thing, Edmund (the gamemaster) did not hold us to the mad-libs, so some were used as-is and other bonds were created from scratch based on extensive discussion.

And because the two sources of experience points in a player’s control are the alignment and the bonds, we hit those hard. But while the alignment doesn’t change except under unusual circumstances, every time we resolve a bond for experience, that bond changes. If we want the new bond to come into play, it has to be relevant (to be used in the Aid or Interfere move) and have a chance to be resolved (for experience.) This is also a chance to direct where we hope the story will go, to define the conflicts we want to see played out.

The bonds are being used like the Beliefs, Instincts and Goals in the Burning Wheel games (including Mouse Guard, Torchbearer, etc.): a mechanically-rewarded way of investing in the story and characters. But not all bonds are constantly changing; because we can have up to four bonds and there are only four characters in the party, we’ve had the leisure of creating bonds that evolve more slowly or provide more of a flag for role-play than a source of experience. For example, my character has a bond with her goddess that I don’t foresee changing except in case of major story development.

As a result of the focus on bonds, we’ve had amazing sessions where the GM could hardly get a random encounter in edgewise; the action and drama revolved around the character relationships. And it’s so much fun! It’s exactly what I needed to make dungeoneering interesting. We still fight the monsters and, given the chance, take the treasure (or at least Merit does), but we have powerful motivations.

You can read Edmund’s introduction to the setting, and Sean’s episode recaps.

Relationship map after Episode 9

---

Notes
⁽¹⁾ But not his reasoning, relying on the “’alpha players can get away with shit’ ethos.” No, no, no. If I’m running this game and you ask me whether you can have a name that’s not on the playbook list, or phrase a bond differently from the list, I’ll say yes. Fuck alpha player ethos. Return