#i fully plan for this to be my new dao. my new skyrim. the game i sink thousands of hours intl over 10+ years

[pinacoladamatata]

One thing I'm actually genuinely excited for is see everyone play BG3 and all of us piecing together what choices trigger certain outcomes and watching how wildly playthroughs can diverge! Because there's no way I'm playing every race/class combo to the end, I'm just so excited to see all the different scenes people might experience and share!

[felassan]

Insights into DAI’s development from Blood, Sweat, and Pixels

The book is by game industry journalist Jason Schreier (it’s an interesting read and well-written, I recommend it). This is the cliff notes version of the DAI chapter. This info isn’t new as the book is from 2017 (I finally got around to buying it). Some insight into DAO, DA2 and cancelled DA projects is also given. Cut for length.

BW hoped that DA would become the LotR of video games. DAO’s development was “a hellish seven-year slog”

The DAI team are compared to a chaotic “pirate ship”, which is what they called themselves internally. “It’ll get where it needs to go, but it’s going to go all over the place. Sail over here. Drink some rum. Go over here. Do something else. That’s how Mark Darrah likes to run his team.” An alternative take from someone else who worked on the game: “It was compared to a pirate ship because it was chaotic and the loudest voice in the room usually set the direction. I think they smartly adopted the name and morphed it into something better.”

A game about the Inquisition and the large-scale political conflicts it solves across Thedas, where the PC was the Inquisitor, was originally the vision for ‘DA2′. Plans had to change when SW:TOR’s development kept stalling and slipping. Frustrated EA execs wanted a new product from BW to bolster quarterly sales targets, and decided that DA would have to fill the gap. BW agreed to deliver DA2 within 16 months. “Basically, DA2 exists to fill that hole. That was the inception. It was always intended to be a game made to fit in that”

BW wanted to call it DA: Exodus, but EA’s marketing execs insisted on DA2, no matter what that name implied

DAO’s scope (Origin stories, that amount of big areas, variables, reactivity) was just not doable in a year, even if everyone worked overtime. To solve this problem, BW shelved the Inquisition idea and made a risky call: DA2 would be set in one city over time, allowing locations to be recycled and months to be shaved off dev time. They also axed DAO features like customizing party members’ equipment. These were the best calls they were able to make on a tight line

Many at BW are still proud of DA2. Those that worked on it grew closer from all being in it together

In certain dark accounting corners of EA, despite fan response to DA2 and its lower sales compared to DAO, DA2 is considered a wild success

By summer 2011 BW decided to cancel DA2′s expansion Exalted March in favor of a totally new game. They needed to get away from the stigma of DA2, reboot the franchise and show they could make triple-A quality good games. 

DAI was going to be the most ambitious game BW had ever made and had a lot to prove (that BW could return to form, that EA wasn’t crippling the studio, that BW could make an ‘open-world’ RPG with big environments). There was a bit of a tone around the industry that there were essentially 2 tiers of BW, the ME team and then everyone else, and the DA team had a scrappy desire to fight back against that

DAI was behind schedule early on due to unfamiliar new technology; the new engine Frostbite was very technically challenging and required more work than anyone had expected. Even before finishing DA2 BW were looking for a new engine for the next game. Eclipse was creaky, obsolete, not fully-featured, graphically lacking. The ME team used Unreal, which made inter-team collab difficult. “Our tech strategy was just a mess. Every time we’d start a new game, people would say, ‘Oh, we should just pick a new engine’.”

After meeting with an EA exec BW decided on Frostbite. Nobody had ever used it to make an RPG, but EA owned FB dev studio DICE, and the engine was powerful and had good graphic capabilities & visual effects. If BW started making all its games on FB, it could share tech with sister studios and borrow tools when they learned cool new tricks. 

For a while they worked on a prototype called Blackfoot, to get a feel for FB and to make a free-to-play DA MP game. It fizzled as the team was too small, which doesn’t lend itself well to working with FB, and was cancelled

BW resurfaced the old Inquisition idea. What might a DA3 look like on FB? Their plan by 2012 was to make an open-world RPG heavily inspired by Skyrim that hit all the beats DA2 couldn’t. “My secret mission was to shock and awe the players with the massive amounts of content.” People complained there wasn’t enough in DA2. “At the end of DAI, I actually want people to go, ‘Oh god, not [another] level’.”

It was originally called Dragon Age 3: Inquisition

BW wanted to launch on next-gen consoles only but EA’s profit forecasters were caught up in the rise of iPad and iPhone gaming and were worried the next-gen consoles wouldn’t sell well. As a safeguard EA insist it also ship on current-gen. Most games at that time followed this strategy. Shipping on 5 platforms at once would be a first for BW

Ambitions were piling up. This was to be BW’s first 3D open-world game, and their first game on Frostbite, an engine that had never been used to make RPGs. It needed to be made in roughly two years, it needed to ship on 5 platforms, and, oh yeah, it needed to restore the reputation of a studio that had been beaten up pretty badly. “Basically we had to do new consoles, a new engine, new gameplay, build the hugest game that we’ve ever made, and build it to a higher standard than we ever did. With tools that don’t exist.”

FB didn’t have RPG stats, a visible PC, spells, save systems, a party of 4 people, the same kind of cutscenes etc and couldn’t create any of those things. BW had to create these on top of it. BW initially underestimated how much work this would be. BW were the FB guinea pigs. Early on in DAI’s development, even the most basic tasks were excruciating, and this impacted even fundamental aspects of game design and dev. When FB’s tools did function they were finicky and difficult. DICE’s team supported them but had limited resources and were 8 hours ahead. Since creating new content in FB was so difficult, trying to evaluate its quality became impossible. FB engine updates made things even more challenging. After every one, BW had to manually merge and test it; this was debilitating, and there were times when the build didn’t work for a month or was really unstable.

Meanwhile the art department were having a blast. FB was great for big beautiful environments. For months they made as much as possible, taking educated guesses when they didn’t know yet what the designers needed. “For a long time there was a joke on the project that we’d made a fantastic-looking screenshot generator, because you could walk around these levels with nothing to do. You could take great pictures.”

The concept of DAI as open-world was stymying the story/writers and gameplay/designers teams. What were players going to do in these big landscapes? How could BW ensure exploring remained fun after many hours? Their teams didn’t have time for system designers to envision, iterate and test a good “core gameplay loop” (quests, encounters, activities etc). FB wouldn’t allow it. Designers couldn’t test new ideas or answer questions because basic features were missing or didn’t exist yet. 

EA’s CEO told BW they should have the ability to ride dragons and that this would make DAI sell 10 million copies. BW didn’t take this idea very seriously

BW had an abstract idea that the player would roam the world solving problems and building up power or influence they could use. But how would that look/work like in-game? This could have used refinement and testing but instead they decided to build some levels and hope they could figure it out as they went.

One day in late 2012, after a year of strained development on DAI, Mark Darrah asked Mike Laidlaw to go to lunch. “We’re walking out to his car,” Laidlaw said, “and I think he might have had a bit of a script in his head. [Darrah] said, ‘All right, I don’t actually know how to approach this, so I’m just going to say it. On a scale of one to apocalyptic... how upset would you be if I said [the player] could be, I dunno, a Qunari Inquisitor?’” 

Laidlaw was baffled. They’d decided that the player could be only a human in DAI. Adding other playable races like Darrah was asking for would mean they’d need to quadruple their budget for animation, voice acting, and scripting.

“I went, ‘I think we could make that work’,” Laidlaw said, asking Darrah if he could have more budget for dialogue. 

Darrah answered that if Laidlaw could make playable races happen, he couldn’t just have more dialogue. He could have an entire year of production.

Laidlaw was thrilled. “Fuck yeah, OK,” he recalled saying.

MD had actually already realized at this point it’d be impossible to finish DAI in 2013. They needed at least a year’s delay and adding the other playable races was part of a plan/planned pitch to secure this. He was in the process of putting together a pitch to EA: let BW delay the game, and in exchange it’d be bigger and better that anyone at EA had envisioned. These new marketing points included playable races, mounts and a new tactical camera. If EA wouldn’t let them delay, they would have had to cut things. Going into that BW were confident but nervous, especially in the wake of EA’s recent turmoil where they’d just parted ways with their CEO and had recruited a new board member while they hunted for a new one. They didn’t know how the new board member would react, and the delay would affect EA’s projections for that fiscal year. Maybe it was the convincing pitch, or the exec turmoil, or the specter of DA2, or maybe EA didn’t like being called “The Worst Company in America”. Winning that award 2 years in a row had had a tangible impact on the execs and led to feisty internal meetings on how to repair EA’s image. Whatever the reasons, EA greenlit the delay.

The PAX Crestwood demo was beautiful but almost entirely fake. By fall 2013, BW had implemented many of FB’s ‘parts’, but still didn’t know what kind of ‘car’ they were making. ML and team scripted the PAX demo by hand, entirely based on what BW thought would be in the game. The level & art assets were real but the gameplay wasn’t. “Part of what we had to do is go out early and try to be transparent because of DA2. And just say, ‘Look, here, it’s the game, it’s running live, it’s at PAX.’ Because we wanted to make that statement that we’re here for fans.”

DA2 hung on the team like a shadow. There was insecurity, uncertainty, they had trouble sticking to one vision. Which DA2 things were due to the short dev time and which were bad calls? What stuff should they reinvent? There were debates over combat (DAO-style vs DA2-style) and arguments over how to populate the wilderness.

In the months after that demo, BW cut much of what they’d shown in it. Even small features went through many permutations. DAI had no proper preproduction phase (important for testing and discarding things), so leads were stretched thin and had to make impulsive decisions.

By the end of 2013, DAI had 200+ people working on it, and dozens of additional outsourced artists in Russia and China. Coordinating all the work across various departments was challenging and a full-time job for several people. At this sheer scale of game dev, there are many complexities and inter-dependencies. Work finally became significantly less tedious and more doable when BW and DICE added more features to FB. Time was running out though, and another delay was a no.

The team spent many hours in November and December piecing together a “narrative playable” version of the game to be the holiday period’s game build for BW staff to test that year. Feedback on the demo was bad. There were big complaints on story, that it didn’t make sense and was illogical. Originally the PC became Inquisitor and sealed the breach in the prologue, which removed a sense of urgency. In response the writers embarked on Operation Sledgehammer (breaking a bone to set it right), radically revising the entire first act.

The other big piece of negative feedback was that battles weren’t fun. Daniel Kading, who had recently joined BW and brought with him a rigorous new method for testing combat in games, went to BW leadership with a proposal: give him authority to open his own little lab with the other designers and call up the entire team for mandatory play sessions for test purposes. They agreed and he used this experiment to get test feedback and specifically pinpoint where problems were. Morale took a turn for the better that week, DK’s team made several tweaks, and through these sessions feedback ratings went from 1.2 to 8.8 four weeks later.

Many on the team wished they didn’t have to ship for old consoles (clunky, less powerful). BW leadership decided not to add features to the next-gen versions that wouldn’t be possible on the older ones, so that both versions of the game played the same. This limited things and meant the team had to find creative solutions. “I probably should’ve tried harder to kill [the last-gen] version of the game”, said Aaryn Flynn. In the end the next-gen consoles sold very well and only 10% of DAI sales were on last-gen.

“A lot of what we do is well-intentioned fakery,” said Patrick Weekes, pointing to a late quest called “Here Lies The Abyss”. “When you assault the fortress, you have a big cut scene that has a lot of Inquisition soldiers and a lot of Grey Wardens on the walls. And then anyone paying attention or looking for it as you’re fighting through the fortress will go, ‘Wow, I’m only actually fighting three to four guys at a time.’ Because in order for that to work [on old gen], you couldn’t have too many different character types on screen.”

Parts of DAI were still way behind schedule because it was so big and complex, and because some tools hadn’t started functioning until late on. Some basic features weren’t able to be implemented til the last minute (they were 8 months from ship before they could get all party members in the squad. At one point PW was playtesting to check if Iron Bull’s banter was firing, and realized there was no way to even recruit IB) and some flaws couldn’t be identified til the last few months. Trying to determine flow and pacing was rough.

They couldn’t disappoint fans again. They needed to take the time to revise and polish every aspect of DAI. “I think DAI is a direct response to DA2,” said Cameron Lee. “DAI was bigger than it needed to be. It had everything but the kitchen sink in it, to the point that we went too far... I think that having to deal with DA2 and the negative feedback we got on some parts of that was driving the team to want to put everything in and try to address every little problem or perceived problem.”

At this point they had 2 options: settle for an incomplete game, which would disappoint fans especially post-DA2, or crunch. They opted to crunch. It was the worst period of extended overtime in DAI’s development yet and was really rough: late nights, weekends, lost family time, 12-14 hour days, stress, mental health impacts.

During 2014′s crunch, they finally finished off features they wished they’d nailed down in year 1. They completed the Power (influence) system and added side quests, hidden treasures and puzzles. Things that weren’t working like destructible environments were promptly removed. The writers rewrote the prologue at least 6 times, but didn’t have enough time to pay such attention to the ending. Just a few months before launch pivotal features like jumping were added.

By summer BW had bumped back release by another 6 weeks for polish. DAI had about 99,000 bugs in it (qualitative and quantitative; things like “I was bored here” are a bug). “The number of bugs on an open-world game, I’ve never seen anything like it. But they’re all so easy to fix, so keep filing these bugs and we’ll keep fixing them.” For BW it was harder to discover them, and the QA team had to do creative experimentation and spend endless late nights testing things. PW would take builds home to let their 9 year old son play around. Their son was obsessed with mounting and dismounting the horse and accidentally discovered a bug where if you dismounted in the wrong place, all your companions’ gear would vanish. “It was because my son liked the horse so much more than anyone else ever had or will ever like the horse.”

MD had a knack for prioritizing which bugs should be fixed, like the one where you could get to inaccessible areas by jumping on Varric’s head. “Muscle memory is incredibly influential at this point. Through the hellfire which is game development, we’re forged into a unit, in that we know what everyone’s thinking and we understand everyone’s expectations.”

At launch they still didn’t have all their tools working, they only had their tools working enough.

DAI became the best-selling DA game, beating EA’s sales expectations in just a few weeks. If you look closely you can see the lingering remnants of its chaotic development, like the “garbage quests” in the Hinterlands. Some players didn’t realize they could leave the area and others got caught in a “weird, compulsive gratification loop”. Internet commentators rushed to blame “those damn lazy devs” but really, these were the natural consequences of DAI’s struggles. Maybe things would have been different if they’d miraculously received another year of dev time, or if they’d had years before starting development to build FB’s tools first.

“The challenge of the Hinterlands and what it represented to the opening 10 hours of DAI is exactly the struggle of learning to build open-world gameplay and mechanisms when you are a linear narrative story studio,” said Aaryn Flynn.

“DA2 was the product of a remarkable time-line challenge,” said Mike Laidlaw, “DAI was the product of a remarkable technical challenge. But it had enough time to cook, and as a result it was a much better game.”

Read the chapter for full details of course!