In the recent War for the Atlas expansion we introduced 9 new gems, including four Skill gems that focused on necromancy, and five support gems. Each of these were designed to augment as many skills as possible in new and interesting ways. Senior Game Designer, Rory, has taken a moment to give us an insight into the design and development process for some of these gems.

The Corpse Skills

The first step of skill development was to decide what our intent should be when adding the new skills. Sometimes the goal of a new skill is to fill out a small gap in a build's progression, so that a skill theme has a variety of options to try out while progressing towards a final build. Sometimes the purpose of a new skill is to introduce an entirely unique playstyle to the game, like the addition of Ancestor Totems to give melee characters a different way to play. Sometimes the purpose is to expand a single skill out into a full playstyle. This was the direction we chose for this set of skills.

We looked at the possibility of both expanding on the core style of Detonate Dead and improving the Detonate Dead skill. We designed a selection of skills that deal pure offensive damage by using corpses as a resource. This was something we'd talked about over multiple design sessions in years past, so we soon narrowed down the options to a few designs we were happy with, each with a different way of interacting with corpses.

We went through a few options for other skills we could add that didn't interact with corpses or interacted in other ways. Some skills that were created for the 3.1.0 patch are still in development, including some skills that introduce entirely new build styles. The time pressure of releasing a league and an expansion together meant we had no choice but to delay them to focus on the cohesive set of skills that would work well when released together and free up the team to put more effort into the new bosses.

Fire

One of the early doubts during discussions with other developers was the plan for everything to deal Fire damage. Originally, Unearth was also planned as a fire skill.

We have a few skill sets that have a diverse set of elements, like traps and elemental attacks. These do have the advantage of naturally creating a more diverse variety of builds within the skill set but often restrict players and their choices when the passive tree or item choices they take end up restricting their benefits to only one of the skills in the skill set. For example, building Lightning Trap could mean taking bonuses to Projectile effects and Lightning Damage, making it a painful process if the player later wants to try out Ice Trap.

Though we did change Unearth to a Physical Damage skill for thematic reasons, we elected to build the new corpse skills as fire skills in the style of Detonate Dead to let players mix and match the skills during the leveling process. The nature of the skills happened to result in all of the skills benefiting from both Area and Fire Damage bonuses, which further enhanced the ability to mix and match the skills. The result is a fluid leveling progression that gives players the opportunity to experiment.

The Failed Prototype

We had also prototyped a skill codenamed "Infernal Sweep" which was a fiery area of effect sweep attack that exploded nearby corpses. Early testing showed it felt bad to use, having to swing twice to first kill some enemies then again to detonate their bodies for larger area damage. The skill was visually spectacular, but didn't play well unless it was boosted in both starting area and damage to the point where it both invalidated the roles of Sweep and Infernal Blow. This skill went back to the drawing board and we'll likely see it again in future with additional mechanics or without the reliance on corpses.

The Support Gems

Early in the development of the skills, around the time when we were first starting to get skill effects, it was decided that we would add a series of new mechanic-altering support gems. This list started large and was trimmed down as the programmers investigated the planned support gems, however Spell Cascade was clearly one that was going to influence the balance of the new corpse skills.

We decided fairly early on that corpse-targeted skills should count as area-targeted skills for the purpose of Spell Cascade. This would mean the potential for overlapping damage would be quite high, unless we implemented major restrictions on the area that corpses could be detonated. We decided against the restriction, instead opting to balance around the potential for overlapping damage purely based on how enjoyable it could be.

The nature of our support system does mean that the damage would feel significantly lower without the Spell Cascade support in most setups, but the additional corpse cost did mean that characters would at least have to work harder to maintain a full complement of corpses, or be restricted by skill cooldowns.

Iteration

All of the skills received changes during development, some to fix balance problems, some to solve technical problems, and some to improve how the skill felt to use.

Detonate Dead had its damage boosted massively, to make it powerful at lower levels at the point corpses had very little life. Its area was also boosted, as its new effects let the effect scale to area of effect, unlike the old corpse blood splatters.

Volatile Dead had improvements to its AI system to make the orbs act more intelligently, focussing on unique and rare monsters while spreading out to take out weaker individual monsters. The speed the orbs detonated was greatly increased, so monsters didn't have a chance to flee before the orbs detonated.

Bodyswap had a new effects technology added so we could have blood drip off the character after bodyswapping. Very important. We also made improvements to how the skill worked with totems to enable the cool-looking build, though it became clear that further totem AI improvements would have to be made to let the skill work as well as it should, and we didn't have the time available to do that.

A lot of effort went into reducing the framerate drop of Cremation without compromising the epic visuals. On the very last last day before release we solved an issue that meant mines could result in hundreds of cremation geysers active at once, which was also rather bad for framerate.

We're excited to see what players do with the corpse interaction skills during the league, as they've proved far more popular than first anticipated. We hope you all enjoyed this glimpse into the skill development process!