Mass Effect: Andromeda's development
In March of 2017, the fourth entry in the beloved Mass Effect series, titled Mass Effect: Andromeda was released to middling reception, averaging from 65 to 75 on Metacritic. Most of the game's criticisms were centred on the weak story, flat characters, awful dialogue and bad voice work. Many fans who were excited for the game only to be disappointed were left asking one major question: what on Earth happened? However, in June 2017, two and a half months later, Jason Schreier posted an article that goes into great depth discussing many of the issues with the game's five year development cycle after a three month investigation.
NOTE: Everything you about to read is based on what was uncovered by Jason Schreier's investigative report, so he is the one to thank (even if Kotaku as a whole is not to be trusted).
2012 (Early development)
The main purpose of Mass Effect: Andromeda was originally to act as soft reboot for the series, hence events and characters aren't referenced in this entry. BioWare was split into two teams: the Edmonton team was working on a new IP (codenamed Dylan, later Anthem), while Mass Effect: Andromeda was handed to the Montreal team. The team decided that this fourth game would take place after the original trilogy after getting feedback from fans, focus groups, and internal members. The goal of this title was to zero in on exploration, which is a major aspect of the first Mass Effect entry that was fully explored at the time.
2013 (Pre-production, and where the trouble began)
The game was originally proposed to feature over 100 procedurally generated worlds, similar to what No Man's Sky later attempted in 2016. The game was also planned to include actual space travel, a ship to pilot, and seamless transitioning from space flight to landing on planets, of which there are working prototypes. Transitioning from the ship to a ground vehicle and back again was all planned to be seamless as possible to create the feeling of immersive space travel with no interruptions. Montreal even brought in Gérard Lehiany to direct the game's narrative. However, the initial game design caused a lot of trouble.
The most glaring problem is that the team tried to build the story around the idea of procedurally generated worlds without taking into account how it would work. Most developers would design their game worlds around the story, which in turn would give them a more solid development framework. The team simply had no idea how the story would work around the proposed idea of procedurally generated worlds. Not helping the Montreal team is that the team was understaffed, there were technical hiccups, and a lack of knowledge of the tools needed for such game design, since only the design team within the Montreal studio had an understanding of how to create procedurally generated worlds using an engine called "World Machine" that would create some impressive looking mountains and simulate the effects of erosion. The other teams though couldn't get to grips with the toolkit, and would have to touch them up by hand just to get it to look great.
On top of the issues with World Machine, the studio was stuck with the Frostbite Engine, which is of course EA's proprietary engine. EA had mandated the use of Frostbite for many of their games, which caused many headaches for BioWare especially, since the engine was originally intended for first-person shooters like Battlefield, not for action role-playing games like Mass Effect. Although BioWare managed to get the engine to function as intended with Dragon Age: Inquisition, it still had issues as a lot of assets had to be made from scratch. One developer told Schreier that a lot of the development troubles that Inquisition had, carried over to Andromeda. One such issue is that Frostbite did not feature an animation rig right out the box, causing the animation team to animate everything by hand. Manveer Heir decried the engine for its impracticality and essentially describing it as a pain in the ass to work with, stating that what takes 2 days build in Frostbite would take 2 hours in a more competent engine. Other source compared the engine to a Formula One car, saying that it does certain things really well, but when doesn't, it's a disaster. Even simple stuff like a saving/streaming system and increasing the map size had to be altered from scratch.
2014 (Conflicts and departures)
Some BioWare developers described 2014 as the worst year of their professional lives for several reasons, among them being internal conflicts and in-office politics. The Montreal studio frequently clashed with the Edmonton studio, often accusing the latter of sabotage and stealing ideas and staff for their own projects. By the end of 2014, over a dozen staff had left BioWare, and a lot of open job positions, particularly in the animation team, were never filled. This left the animation team woefully understaffed.
BioWare suffered a major setback in August 2014 when the Mass Effect Trilogy's creative lead, Casey Hudson, left the development studio altogether, as did the game's director, Gérard Lehiany shortly afterwards. The latter was soon replaced by Mac Walters, who had a different vision for the project, which of course led to re-scoping. Technical difficulties continue to plagued the developers; combat and vehicles were coming along very nicely, but the space travel and prodecurally generated worlds were lagging behind, and while prototypes were built, the team found driving around in these worlds to not be that fun. The game left pre-production too early, as some of the design framework was not properly finished, nor had any key decisions like refining what they already had. Hell, they didn't even have a proof of concept, or "Vertical Slice" as it has become known as in recent years.
2015 (Full production, and scrapped ideas)
In the summer of 2015, the reality of the game's situation was slowly starting the rear itself: some of the original ideas the team had for Mass Effect: Andromeda were simply not feasible. Because the team was having so much difficulty implementing their initial ideas, they decided it was better to cut their losses and just scrap space travel and procedurally generated worlds entirely and re-scope. The 100 procedurally generated planets was soon reduced to just 30 planets with procedurally generated terrain instead, only to be further reduced to just 7. At this point, the decision to re-scope this late on in development was slowly backfiring, as the more important aspects like story, level design, and cinematics were falling far behind. It does show why games are usually designed around the story and setting, not the other way around.
Not helping matters is that a lot of the development team had staff stationed all over the place, with staff in Edmonton, Montreal, and Austin causing all sorts of communication problems, since all three Providences were in different time zones, which resulted in even the simplest of tasks taking much longer than necessary. To matters worse, the team even had to outsource to locations in South Africa and Vancouver, which only exacerbated the communication issues. In fact, this lack of communication and co-ordination are a big reason why the game only started coming together 3 years into the game development, with a massive rush to finish in the final 2 years, resulting in "Crunch time".
2016 (Regression)
Normally, when a game reaches its final months of development, it gets polished and fine tuned; everything starts to really come together as a cohesive package at this stage. In Mass Effect: Andromeda's case, however, everything just started to regress because the game was so badly rushed; anything that was already made gets bug tested and approved by higher-ups. only to just fall apart because of the shoddy framework. All of the fundamental gameplay elements were nearly complete, but story, animation, and level design took a backseat during the past year. As a result, the "Downstream Team", consisting of the effects, cinematics, and audio divisions had a ton of work piled on them at once, as they are typically the last teams to polish the game.
At this point, BioWare Montreal was feeling hopeless, and described Andromeda as the hardest project they had ever undertaken, and ended up spent the last few months development just trying to keep the game from collapsing instead of polishing it. The aforementioned indecision also took its toll on the animation team, which lead to one of the game's biggest sore spots when it would eventually launch the following year: the animation.
The biggest contributing factor for why the animation ended up looking so bad is, once again, indecision, but also a lack of leadership. For one, the team remained understaffed at this point, but also could not decide on which kind of animation tools they should use; some staff wanted to use "Snappers" an Egyptian made rig that could render some impressive facial animations, but the team were uncertain about its compatibility with Frostbite, and how it would scale across the whole game. Other staff wanted to use EA's in-house face rendering program called "FaceWare", but some argued it was inferior. The team used a common program called "Face FX" for lip-synching, which is typically used to convert sound into lip movement. It is mostly automated and refined by hand, but due to time constraints, BioWare Montreal had little time to touch up the lip-synching by hand, so had to settle for the automated system.
The team also lost a ton of time and work in pre-production after they decided to switch animation programs. All of these factors resulted in BioWare Montreal having to outsource to other BioWare studios when it came to cinematics, as far out as China, just to get the animations done. This only further hampered the team's efforts due to the radically different time zones causing communication delays.
2017 (Release and fallout)
Just before the game released, video clips and screenshots from the game surfaced showcasing some bugs and glitches, really poor looking animations and characters models, specifically the faces. The game was quickly mocked by a large portion of the gaming community, and became something of a meme overnight, with gifs of the game's worst animation quirks spreading across the internet like wildfire. The team sent out playable builds to mock reviews so that a Metacritic score can be deduced. In spite of everything that went wrong during the development process, things were looking up slightly for Montreal, as mock review score came back with a mock score range of 80-85/100.
When the game finally launched on March 28th 2017 in North America, the game was met with mixed reviews from critics, but outright ridicule from gamers. Scoring a 70% on Metacritic for the PS4 version, it was BioWare's lowest rated game as of 2017 (even lower than Sonic Chronicles: The Dark Brotherhood). This was a fatal blow to BioWare Montreal, as after the reviews poured in, the studio was scaled down, and the Mass Effect series was shelved. Eventually, the Montreal branch of BioWare was shut down, and EA's support for Mass Effect: Andromeda ended less than a year after the game launched.
The end result of the whole thing is a game that was far too ambitious for the resources the team had, which wasn't that many. It also probably didn't help the game that EA, the publisher, did very little to help the game by delaying it to a later date, which is glaring since the infamous publisher has a history of letting games get released in an unfinished state, and is more than happy to release a rushed product. The hope is that future projects from BioWare, EA, and even other developers and publishers would give their studios the time they need to allow a game to come together, and the importance of proper management and leadership.
Unfortunately though, BioWare Edmonton did not learn from Mass Effect: Andromeda's mistakes, and went through an equally turbulent development cycle with their newest project, Anthem.
References