The latest issue of Gamasutra sister publication Game Developer magazine, available for subscribers and for digital purchase now, includes a postmortem of Raven Software's Singularity, written by a group of the game's key staff.
Singularity is a first person shooter featuring a sci-fi narrative that guides players through time between the present and the 1950's, and includes weapons that can temporally manipulate objects in the game world.
Announced in 2008, Activision published the title for Xbox 360, PlayStation 3, and PC in Europe on June 29, 2010, with a North American release on June 29.
These excerpts -- written collectively by Raven's Rob Gee, Brian Raffel, Steve Raffel, Jon Zuk, Dan Vondrak, and Gustavo Rasche, and extracted from the September 2010 issue of Game Developer magazine reveal various "What Went Right" and "What Went Wrong" highlights from throughout the creation of the game.
Along the way, the team detail how the game evolved from its initial prototype, and the challenges Raven experienced when developing a new IP with a large team and budget.
Short Term Goals With Purpose
The first playable build is generally where everything is figured out—all the main features of the game are functional, and a vertical slice of the game can be shown in a polished state. From a solid first playable build, everyone should know to “just do more of that” and build the rest of the game.
One issue with getting to that milestone is the distance from the prototype phase. When a project is longer than two years, six months can pass before the game needs to be shown again. To keep the team moving forward, we needed to set some short term goals.
The first thing we did was sit down and hash out a plot for the rest of the game and nail down what levels we wanted. The team was forced to wait this out for a week or two but it was an important foundation. From there, we split the team into two separate groups and gave each of them a level with a goal to have it planned and blocked out in two weeks, refined in another two, and ready for polish in a few after that.
We thought that this would flow gracefully into the next set of levels. The plan from there was to choose the best section of the game, add a little polish, and show off our vertical slice. Although production had a long term plan for the whole project, we kept the team thinking more about the next few weeks ahead of them.
As we worked through this plan for the next few months, an opportunity arose to possibly get a major magazine cover. This was great because it would let people know what the game was about, build buzz, and generally just get the game into people’s consciousness. Because we had the team operating on short term goals, it was easy to adjust and insert this new goal into the pipeline.
We decided to continue working on normal production but peeled some people off to add polish to the Freighter level for our presentation to the magazine. We worked for a solid four or five weeks to tighten up game mechanics, optimize the level, work on the HUD, and more. At the end of that time period, we presented the game to the magazine and found out a few weeks later that we would get the cover. Having that short term goal with a solid purpose (get magazine cover) really spurred the team into action.
Not Enough Focus On Story Or Additional Gameplay Beyond Prototype
Since the prototype was primarily intended as a proof of concept of the TMD, not as much time as we had hoped was put into developing the story and supporting fiction beyond some general points. We knew what type of gameplay we wanted and that the game would involve a storyline that jumped between two time periods, but we were too short on specifics such as characters, creatures, key locations, etc. This made for a difficult transition into pre-production as people came off other projects and Singularity ramped up rather quickly.
Also, when we had the main character Renko’s voice in the game, it was really difficult to find the right voice actor, and to nail the dialogue. We tried a number of variations; Renko always seemed to come across in a negative way—too funny, too cocky, too serious, or too whiny.
Once we removed his voice, we had the new challenge of rewriting the dialogue for other characters, since they had no one to play off. But overall, dropping his voice really helped the player experience and made the moments in the game more impactful to the player.
Too Much Stuff, Too Little Time
Singularity was an ambitious project from the start, and going into that final phase of production, we bit off quite a lot, ultimately more than we should have given the time we had. For example, at one point, we re-modeled and re-textured all the weapons.
We also changed functionality for many of them, and of course tweaked out all the weapon data, such as damage, aim friction, and turn speed. We expanded TMD and weapon upgrades, layering on a light RPG system to provide more choice and customization (which was a good thing, but still added considerable work).
Lots of other work went into the final phase—we added more enemy types, reworked the UI, added the all-new multiplayer with our third-person action versus FPS environment, and then of course had to polish everything that was already in the game.
This really left us short-handed at the end. We did cut features to help ourselves out, though. We had a time teleport system, for example, which allowed players to retry levels and would ultimately allow them to go back and solve puzzles with their upgraded TMD that couldn’t be solved the first time through the level.
But we had to drop it because there was no way we could have gotten it tested for all the possibilities players could come up with. There is internal debate to this day regarding which features we never should have started, or should have cut earlier.
As Singularity’s development was winding down, we already started talking about building more polish time into the next project—and doing a little more pre-production early to test out the ideas we have—so we can cut the bad or overly-complex ideas earlier.
The full postmortem of Singularity explores more of "What Went Right" and "What Went Wrong" during the course of the game's development, and is now available in the September 2010 issue of Game Developer magazine.
The issue also includes a feature on the technical side of massively multiplayer first person combat, a piece on using metrics to aid playtesting, as well as our monthly columns on design, art, music, programming, and humor.