I’m Raid Designer Byron Miller, here to talk to you about the Hall of Chains raid instance. In the raid, we meet Desmina, and we help her defeat Dhuum, the fallen god of death who was sealed away by players in the original Guild Wars®. With the gods having left Tyria, he’s making a play for power.

Building a raid boss is an exciting mix of “What does this boss do?” and “How do I make this boss do the thing that it does?” Usually this breaks down into two main phases of development: planning and implementation.

Planning Phase

The planning part of our process involves Microsoft Paint, magnetic pen-and-paper tabletop mats, and whatever figurines I can steal from Crystal’s desk on the way to the meeting room. We usually spend a few weeks in the tabletop and design-document phase while we iron out the narrative, the core gameplay aesthetics, the mechanics that will eventually make up the fights, and the rewards we hope to deliver. We pull in representatives from all disciplines to these meetings so that, from early in the process, we get people thinking about what they can add to make our boss fights a memorable cinematic experience. With bosses like Dhuum who have a large body of established lore and aesthetics, it’s an exercise in bringing the boss into the brave new world of Guild Wars 2. Other bosses, like Desmina, are more of a blank slate where the narrative and visual aesthetic can be generated fresh. Desmina and Dhuum’s (and Grenth’s) relationship was in some ways born of this difference.

A boss can change quickly and dramatically during the planning phase because we haven’t started implementing the encounter. If, for example, Ben Arnold comes in and says, “Hey folks, I just made it so you can walk on the ceiling, and it’s super easy to implement,” there’s a reasonably high chance that we’ll want to explore the possibilities of an encounter that would utilize this new tech. (Note: We cannot do this, but Ben is a wizard, and I’m fairly sure the power of the Mists flows directly through him.)

Because we’re not actively building the encounter at this point, it’s easier to pivot and walk down a different path than the one we started on.

Dhuum specifically went through a ton of changes during the planning phase of development. Early paper designs had a heavy influence on actively chaining Dhuum—in part by banishing players to the corners of the room and shackling them directly to him.

As we worked through other aspects of the design, we realized that preventing nearly half of the raid from actively engaging with the boss was going lead to an encounter that disempowered some players and put a lot of undue pressure on the remaining players to make up for the missing jobs and roles. However, we loved the idea of players participating directly in the chaining of Dhuum, and we didn’t want players focused solely on Dhuum for the entirety of the fight, so we took those concepts and worked them into other parts of the encounter. To keep players engaged in the story, we designed mechanics that would put pressure on Grenth’s reapers, which in turn gave us compelling narrative and mechanical reasons to pay attention to their health. During the final phase, we wanted players to assume the roles of the slain reapers, letting them complete the ritual and activate the magical chains that make Dhuum vulnerable and subject to containment.

Implementation Phase

Once we’ve determined what a boss is going to do, we have to actually make good on our scheming and build the skills we’ve designed. But just because we’ve got a design on paper doesn’t mean that the encounter will feel right when implemented. Sometimes mechanics that seem to synergize on paper end up feeling terrible when layered on top of each other. We can look at some early skills for Dhuum that needed to be reworked after implementation.

Translated from paper to game, Dhuum’s Cull skill was too similar to others we’d designed, and it felt underwhelming compared to skills like Dhuum’s Harvest or Cataclysmic Cycle. In short, it lacked suitable impact. With help from Ben and an assist from the progenitors of linear algebra (thanks, J. Willard Gibbs and Oliver Heaviside!), we were able to build something new and wild.

The Cull you see in the live game is a series of skills that takes custom variable inputs and calculates a randomly growing and splitting series of malevolent fissures in the ground, out of which seep deadly magic.

With a few number changes, we go from this:

To this:

About a third of the way through building the original prototype of this skill, I was struck with an important thought: I should have paid more attention in my high school mathematics classes; the practical application of math in game design is deeper than I could have imagined at the time.

Stay in school.