News - Classic Postmortem: XCOM: Enemy Unknown, which turns 5 today

News - Classic Postmortem: XCOM: Enemy Unknown, which turns 5 today - Printable Version

+- Sick Gaming (https://www.sickgaming.net)
+-- Forum: Sick Gaming Community (https://www.sickgaming.net/forum-1.html)
+--- Forum: Lounge (https://www.sickgaming.net/forum-8.html)
+--- Thread: News - Classic Postmortem: XCOM: Enemy Unknown, which turns 5 today (/thread-83506.html)

News - Classic Postmortem: XCOM: Enemy Unknown, which turns 5 today - xSicKxBot - 10-12-2017

Classic Postmortem: XCOM: Enemy Unknown, which turns 5 today

<div style="margin: 5px 5% 10px 5%;"><img src="https://www.sickgamedev.win/wp-content/uploads/2017/10/classic-postmortem-xcom-enemy-unknown-which-turns-5-today.jpg" width="646" height="363" title="" alt="" /></div><div>In honor of the 5th anniversary of the release of the brilliant <a href="http://store.steampowered.com/agecheck/app/200510/">XCOM: Enemy Unkown</a>, we present this classic postmortem, which first appeared in the January 2013 issue of Game Developer magazine.
The game was a “reimagining” by <a href="http://www.firaxis.com">Firaxis</a> of the classic 1994 strategy title UFO: Enemy Unkown, as well as a reboot of the XCOM series. It was a smashing success, earning numerous awards and GOTY accolades.
This in-depth look at what went right and what went wrong during development was written by <a href="https://twitter.com/gsdeangelis">Garth DeAngelis</a>, who was the lead producer and a level designer on XCOM: Enemy Unknown.
There may have been wounds, but somehow, the XCOM: Enemy Unknown development team evaded permanent death.
In 1994, Microprose released a special PC game called UFO: Enemy Unknown. The turn-based strategy title accumulated a devoted fanbase for its unique take on high-level management against an alien invasion blended with boots-on-the-ground, intimate combat controlling individual soldiers. Fast-forward almost two decades, and Firaxis Games has released XCOM: Enemy Unknown, a reimagining of Julian Gollop’s original design.
The road to completing XCOM was an arduous one. We made many of the same mistakes that other devs have made (and documented in previous Game Developer postmortems): feature creep, communication shortfalls, not enough time, not enough people… If you’re a game developer, you know the story. But through it all, the entire team remained resolute, and in October 2012, against all odds, the development team managed to <del>save Earth</del> ship XCOM on PC, Xbox 360, and PlayStation 3.
<img alt="" src="https://www.sickgamedev.win/wp-content/uploads/2017/10/classic-postmortem-xcom-enemy-unknown-which-turns-5-today.jpg" />
Lead designer Jake Solomon planted a small seed within the creative walls of Firaxis Games in 2004. As the self-proclaimed “biggest fan of the original X-COM” and as a designer/programmer working directly with Sid Meier, it made sense to spark discussion regarding resuscitating one of the greatest strategy titles of all time. After shipping Civilization Revolution, the stars began to align, and the rebuilding of XCOM became a reality.
Before any code was written or any art completed, Jake defined core pillars that would act as the foundation for pillars that would act as the foundation for designing the rest of the game. All facets of game development had clearly evolved since 1994, from game and narrative design to user interface, but Jake remained adamant that certain high-level elements from the original remain holy. These inspirations included maintaining a turn-based combat system, which seemed risky in the modern age of frenetic first-person shooters and real-time action games.
The game would also preserve the symbiotic relationship between the micro combat layer and a macro, grand strategic mode, where the player runs and builds their own secret headquarters to counter the simulated alien invasion. Fans of the original game understood the appeal of these interdependent systems, but to those unfamiliar with the XCOM franchise, this was a foreign game structure, and that equaled potential big risk.
Other smaller pillars included: updating systems such as the fog of war and how visibility would be communicated to the player in a 3D environment; fully destructible environments, a satisfying staple from UFO: Enemy Unknown, but a potential challenge in a game of this scope; reintroducing players to permanent death, and the fact that when it comes to XCOM soldiers, there are no such things as extra lives; and recreating the tense atmosphere from the original.
It was critical to place the player in a world they recognized—in settings they may recall from their own neighborhood or city—and then introduce a menagerie of new and classic aliens into these usually safe environments. This led to an unnerving despair that XCOM fans are all too familiar with. Collectively, these pillars provided the foundation for a challenging experience that presented true consequences, just like UFO: Enemy Unknown.
<img alt="" src="https://www.sickgamedev.win/wp-content/uploads/2017/10/classic-postmortem-xcom-enemy-unknown-which-turns-5-today-1.jpg" />
Firaxis underwent a pitch process that previously had not been attempted with past projects, but then again, we were pitching what was essentially a new IP as a big-budget endeavor, so we needed to make a major splash.
Getting the green light to remake an antiquated turn-based strategy game would be a challenge, and simply verbalizing what made XCOM so special and ripe for a rebirth wouldn’t be enough. The team needed something that would make everyone understand why we were so passionate about undertaking this project—something beyond a static presentation.
Jake began working with project art director Greg Foertsch and a small band of artists to create a gameplay previsualization. Over the course of multiple months, Greg and his team ate, breathed, and slept UFO: Enemy Unknown, ultimately planning a storyboarded sequence that would illustrate how Firaxis’s take on the XCOM universe would not only look, but also how it would play.
On Fridays, Jake would sit in a room with the team while they immersed themselves in the original game, becoming familiar with its nuances and big concepts alike. This collaboration led to a compelling previsualization that not only got the development team on the same page, but also communicated to nondevelopers the potential for a classic turn-based experience to be reborn as something cutting edge, distinctive, and thrilling in the modern age of video games.
<img alt="" src="https://www.sickgamedev.win/wp-content/uploads/2017/10/classic-postmortem-xcom-enemy-unknown-which-turns-5-today-2.jpg" />
Even before beginning work on XCOM , we heard it all before: Games had become too easy. The development (or marketing) buzzword “accessible” translated to “dumbing down,” the idea that developers would take an otherwise deep, rich, and satisfying game and distill its intricacies to its barest form so the entirety of the world could understand, buy, and play said game.
It sounds hyperbolic, but I’ve seen games with easy modes that literally played themselves, making failure impossible, so this stigma against accessibility wasn’t without merit! Making a game “for the masses” could be the ultimate transgression, especially for a complex game with a hardcore past, and we anticipated that XCOM fans would be skeptical that our work would hold up to those who fell in love with the original.
While UFO: Enemy Unknown may have been magnificent, it was also a unique beast when it came to beginning a new game. We often joked that the diehards who mastered the game independently belonged in an elite club, because by today’s standards the learning curve was like climbing Mt. Everest.
As soon as you fire up the original, you’re placed in a Geoscape with the Earth silently looming, and various options to explore within your base—including reading (unexplained) financial reports, approving manufacturing requests (without any context as to what those would mean later on), and examining a blueprint (which hinted at the possibility for base expansion), for example—the player is given no direction.
Even going on your first combat mission can be a bit of a mystery (and when you do first step off the Skyranger, the game will kill off a few of your soldiers before you even see your first alien—welcome to XCOM!). While many fans on the team found this learning curve to be a part of the game’s charm and wore it as a badge of honor, we ultimately knew that, in 2012, we needed to enable gamers to experience the truly fun elements without overly testing their patience. But neither could we bear to dumb XCOM down.
<img alt="" src="https://www.sickgamedev.win/wp-content/uploads/2017/10/classic-postmortem-xcom-enemy-unknown-which-turns-5-today-3.jpg" />
We were on a mission to flip the perception on streamlining, to remove the stigma that accessibility equaled a dirty word. We wanted anyone to be able to give XCOM a whirl without expecting them to become fluent in the game’s many systems on their own accord. At the same time, we needed to preserve all of the richness, depth, and challenge ingrained in the core pillars. If someone wanted to walk away from the experience due to the game’s challenge, we were okay with that; but we didn’t want to alienate anyone simply due to a lack of information.
To accomplish this, we built an optional, integrated tutorial that peeled off the components of XCOM one layer at a time. It was important to keep this hour-and-a-half experience optional, as experienced players could save Earth again without the tutorial force-fed to them (and we also knew some players, even in 2012, would want that old-school badge of honor by skipping the tutorial altogether, which is somewhat appropriate for certain types of X-COM fans).
The introduction to the game wasn’t the only area we redesigned. Jake and the design team refined low-level mechanics from the original, such as removing Time Units and capping the squad loadout at six. Both of these changes were the result of internal playtesting over the course of many months, with the development team finding a combat “sweet spot” with respect to approximate time spent on a map and number of decisions made per turn (we found, depending on map size, battles should average 20 minutes, not to exceed 50 minutes on the absolute longest missions). Six units also made every decision vitally important, promoting group tactics with no moves feeling like unnecessary filler.
This “new era of accessible” mindset also helped the design and user interface teams build a platform-agnostic experience. This is an element that could have gone horribly wrong (and did have its inherent challenges, detailed later), but the team did an admirable job of crafting a historically PC experience for consoles as well. We knew games like XCOM weren’t traditionally available on Xbox 360 and PlayStation 3, but we’re extremely happy we could provide the same experience (without compromising features or “dumbing down” the console versions) across all platforms.
<img alt="" src="https://www.sickgamedev.win/wp-content/uploads/2017/10/classic-postmortem-xcom-enemy-unknown-which-turns-5-today-4.jpg" />
The developers at Firaxis are extremely professional, with each discipline playing the hero role at some point and overcoming monumental obstacles throughout development. From the audio group to the animation and narrative team, they were continually course-adjusting due to dependencies, yet still producing incredible content to polish the game.
On the engineering front, months of changing design had to be technically supported in many complex situations. Systems were built, iterated upon, and some were even discarded after determining a new direction was needed. For example, over the course of a few milestones in midproduction, design asked for sightlines to be drawn from every game unit, soldier, and alien alike, so it was clear what each unit could see.
Our graphics engineering team and artists diligently worked to make this system digestible, but unfortunately, it was tough for the player to determine what was going on amidst the plethora of multicolored lines. After months of trying to get sightlines to work, we eventually realized that the strongest solution was to remove them.
There were plenty of other challenging systems to decipher: the building visibility system underwent various ceiling, wall, and floor rule changes; destruction fidelity fluctuated through a shaky toughness system; and the fog of war was a full 3D cloud early in production, which proved to be a nightmare for both graphics engineering and performance.
Additionally, each gameplay layer (combat and strategy) received drastic overhauls after months of playtesting. In all of these cases, initial engineering efforts had to ultimately be thrown out. To the team’s credit, they understood the nature of iterative design and admirably continued to put in the time needed to make the game a better experience.
<img alt="" src="https://www.sickgamedev.win/wp-content/uploads/2017/10/classic-postmortem-xcom-enemy-unknown-which-turns-5-today-5.jpg" />
In addition, the engineers banded together in a Herculean effort to fix thousands of bugs in postproduction. XCOM is a large, system-driven game with many procedural elements. This meant that many bugs were not only difficult to reproduce, but challenging to even find! Together, engineering raised the bar of the final player experience by squashing these bugs feverishly. Obviously, we couldn’t find and fix every bug, but we’re proud of the effort given in the race to the finish line.
The art and content teams also worked minor miracles. A primary example was the game’s levels. We all love maps and levels, and want more of them; but they are a nexus of many different disciplines somehow crafting the same sculpture all together, and this requires tight coordination and lots of time. Firaxis had never created a level-driven game before (with a strategy system still on top of it, no less), so we had to learn how to build a pipeline that would let us efficiently design and build level assets.
This specifically required an inordinate amount of collaboration between level design and level art, weeks of gameplay testing and feedback per map, and an extreme amount of content creation (we needed to have approximately enough maps for two full playthroughs). In the end, our modestly sized level team ended up exceeding the original goal of 70 unique maps.
Beyond levels, there was still an entire headquarters to build on the strategy layer, with dozens of expandable rooms that could be hand-placed by the player. After making various isometric prototypes, we realized the base wasn’t nearly as gripping as we’d like; something was missing.
Lead technical and HQ artist Dave Black pitched the “ant farm,” a diorama-style side view that instantly connected with the entire team. This was an entirely new process as well, but Dave and the art team concurrently exceeded expectations on headquarters while finalizing all of the combat maps.
<img alt="" src="https://www.sickgamedev.win/wp-content/uploads/2017/10/classic-postmortem-xcom-enemy-unknown-which-turns-5-today-6.jpg" />
We’ve heard countless horror stories about publisher-developer relations, with publishers stifling creativity, dictating direction, or creating impossible deadlines—but our partnership with 2K Games was not one of those horror stories. While there was give-and-take from both sides (as in any relationship), we were overwhelmingly happy with 2K Games’s support—especially considering no major publishers have funded a large-scale, multiplatform, turn-based strategy game in recent memory.
2K believed in our vision and greenlit the project, something we’re not so sure would have happened elsewhere. The 2K Product Development group believed in the potential for a reimagined XCOM and also understood that taking risk was necessary. We were ecstatic to learn we would be given this opportunity.
Furthermore, 2K trusted in us as a studio to own the creative direction of the title. While they provided in-depth milestone feedback, every item was up for discussion, and they ultimately trusted in our design vision. 2K also provided us with additional resources to build an integrated tutorial, something that became critical late and ballooned beyond our initial resource estimations. This type of support proved invaluable to finish the game.
<img alt="" src="https://www.sickgamedev.win/wp-content/uploads/2017/10/classic-postmortem-xcom-enemy-unknown-which-turns-5-today-7.jpg" />
Also, 2K’s public relations team was instrumental in raising the awareness for XCOM. They took the time to understand the vision and value of the project, and allowed the team leads to directly and candidly communicate that vision to the player base. PR worked diligently to uncover many valuable opportunities for the game, including a cover reveal with Game Informer magazine and various demo presentations to targeted press.
These presentations planted the seed in our most passionate advocates—the press—to pass along what they liked (or disliked) about the game’s potential. There was also a strong working relationship between PR and the development team, leading to joint initiatives like the “Jake Solomon Undercover” video and exciting panel discussions like PAX’s “1000 Stupid Ideas on the Road to Glory.”
[embedded content]
XCOM required constant design iteration, with some features being implemented beyond Alpha. It may sound cliché, but Firaxis has always lived by the mantra “Find the Fun,” and the company takes that very seriously. Sometimes, fun can be a challenge to find, especially in a product that is unlike any other we’ve built before. XCOM boasts two interdependent systems that could almost be standalone games, and discovering that special synergy between the two was the key to unlocking the magic within the XCOM universe.
Trying to focus concurrently on both gameplay layers was challenging. We spent various milestones on certain features that didn’t progress as we’d hoped. By midproduction, the strategic layer was a turn-based card system for various months, and it stagnated while the team focused on improving combat. Ultimately, the strategy layer was molded into the version we’re satisfied with, but it was neglected for too long and required a late Half-Life-inspired Cabal process to get there.
We (myself, Solomon, and other members of the dev team as necessary) would meet every morning, every day, until each component of the strategy layer had a concrete game plan and a clear implementation schedule. Additionally, the tutorial and narrative, critical components of the game, couldn’t be pushed to final until the design was locked. And since the design tentpoles ran late, the narrative team (including animators, writers, and audio) came under immense pressure to finalize high-quality cinematics in an extremely short timeframe.
<img alt="" src="https://www.sickgamedev.win/wp-content/uploads/2017/10/classic-postmortem-xcom-enemy-unknown-which-turns-5-today-8.jpg" />
The extra design time helped make the game as good as it could possibly be from a gameplay perspective, but it’s worth asking whether we could have made tough calls on certain systems earlier in the schedule. This is one of game development’s largest challenges: Holding a game’s design to immovable deadlines can be stifling to the iterative and tricky-to-quantify creative process.
Shipping an unpolished combat game with a completely disconnected strategy layer would have spelled disaster for the future of XCOM, so we kept the process malleable much later into the schedule, allowing the team to find the answers through discovery and experimentation.
Practices like the design cabal helped the team focus on areas of the game that weren’t fun, but in a perfect world, we would have locked down as many high-risk systems as possible as preproduction wrapped up. We did ultimately cut content, but the bulk of our wishlist shipped in the final product, which was great for the game but taxing on the team.
<img alt="" src="https://www.sickgamedev.win/wp-content/uploads/2017/10/classic-postmortem-xcom-enemy-unknown-which-turns-5-today-9.jpg" />
By Alpha, only a few systems needed to be implemented, but a new challenge was looming around the corner: the bug database. Before Alpha, the team had a good sense of the state of the game and which systems were most playable, but it was difficult to quantify the true workload until QA began fully testing the game for a few weeks. The reported bug count rapidly multiplied like termites silently infesting the framework of an otherwise beautiful house.
Initially, we weren’t quite sure what to expect, but as the picture became clearer, we knew we were in for an inordinate effort to keep pace with the influx of bugs. There were concerns about the amount of work needed to fix the game relative to the engineers on the team. We had a ship date to hit and we wanted to get our dedicated engineers help.
But the mythical man-month is a very true concept. While our publisher was generous with additional resources to assist toward the end of production, we found that a flood of external helpers had undesired consequences. Knock-on bugs due to unfamiliarity with the codebase, content that needed to be fixed by internal artists, and communications inherent in outsourcing relationships all led to an extreme amount of overhead that ultimately fell onto the laps of internal team members who were already responsible for an aggressive workload.
Outsourcing challenges also hit the content-creation team during production. Communicating with an external cinematic team overseas led to a staggered communication channel. Since there were dozens of unanticipated clerical issues just to get their tools up to speed with ours (no fault of theirs), it was extremely challenging to troubleshoot any setbacks. Also, providing creative feedback to most external partners often led to significant delays due to the remote feedback loop and misinterpretations of feedback via email.
<img alt="" src="https://www.sickgamedev.win/wp-content/uploads/2017/10/classic-postmortem-xcom-enemy-unknown-which-turns-5-today-10.jpg" />
Once we were late into production, the leadership team wanted to maximize each developer’s working hours by being judicious regarding meeting requests, even amongst ourselves. Process-driven meetings were reduced along with costly, 20-plus-person large-scale meetings. We still maintained informal but intimate one-on-one reviews with each discipline’s lead, which was intended to be more focused and fruitful per developer. While the leadership team and some team members appreciated this, others were understandably yearning for additional official communication channels.
Also, team members wanted quicker information on the high-level changes to the design of the game, but with our lead designer doubling as a gameplay engineer, he would often be tied up with coding. Finally, cross-discipline groups (like level design and level art, and feature-specific teams) surely could have benefited from a more formalized stand-up process, which we implemented toward the end of production.
Moving forward, the leadership team knows it needs to strike an appropriate balance between optimal information flow and excessive meeting time, hedging toward more opportunities for formal communication.
<img alt="" src="https://www.sickgamedev.win/wp-content/uploads/2017/10/classic-postmortem-xcom-enemy-unknown-which-turns-5-today-11.jpg" />
Not only was the game structure of XCOM unlike anything the studio had built before, this was also the first time we’ve had to concurrently develop versions for three different platforms. It turned out managing all three was a massive amount of work.
Design-wise, the team knew there would be feature parity between PC and consoles; the only difference would be the control scheme. While the design and UI team did an admirable job on this front, there were continuous challenges throughout development to accommodate multiplatform user interface design, specifically tied to this genre. The team had to ensure all tactical commands were accessible via gamepad, and this involved quite a bit more than accommodating a point-and-shoot mechanic.
The movement system, mapping a system to support dozens of contextual abilities, and crafting a uniform Shot HUD were just a few areas that took time to master across the board. While this specific instance arguably didn’t go “wrong,” it is a small example of the multiplatform challenges faced daily.
<img alt="" src="https://www.sickgamedev.win/wp-content/uploads/2017/10/classic-postmortem-xcom-enemy-unknown-which-turns-5-today-12.jpg" />
The system-specific optimizations needed for each platform were significantly more difficult, particularly for the consoles. Understanding the console constraints for items like number of maps, audio files, texture budget, and animation sizes was a continual process between engineering and the specific disciplines. There were also severe, system-specific bugs, technical requirements, and crashes that ate up much of our senior engineers’ time.
Our systems engineering team was a very talented duo, but they didn’t have a dedicated platform engineer, which meant that they had to partner on all of these complex issues across the board. While they worked together effectively, they simply had too much work on their plates: universal systemic issues, art optimization requests, and other general and technical requirement bugs, just to name a few major workloads.
Our lead engineer assisted on the most difficult issues when he was free from putting out other fires, and another internal systems engineer joined the cause late in the project to own the Xbox 360 technical certification requirements, but these were solutions that emerged late in development.
<img alt="" src="https://www.sickgamedev.win/wp-content/uploads/2017/10/classic-postmortem-xcom-enemy-unknown-which-turns-5-today-13.jpg" />
We’re not proud about the fact that we had to crunch to finish XCOM. We have a dedicated and passionate team, and all team members put in serious extra hours at some point for the good of the project. For many, the malleable structure of the game led to frustration as we were knee-deep in the trenches. Certain dependencies were continually pushed (especially impacting audio, effects, cinematics, and user interface) and the lack of testing on late gameplay systems led to a heavy bug load for the engineers.
On the art side, the content creators had production crunches to finish all maps. As said before, this was the first time we created a game of this structure, and the first time we had to iterate so much on the process itself. While we improved certain inefficiencies throughout production, we simply couldn’t accurately predict how much time we’d need to make the game the way we wanted to make it.
<img alt="" src="https://www.sickgamedev.win/wp-content/uploads/2017/10/classic-postmortem-xcom-enemy-unknown-which-turns-5-today-14.jpg" />
In the end, we avoided permadeath. And after all of the extra hours, the thousands of bugs fixed, the hundreds of level playtests dissecting every piece of cover, the dozens (hundreds?) of gameplay prototypes and healthy debate that accompanied each new system, through every team meal, and in the wake of every hopeful or concerned hallway discussion, in the end, the XCOM development team emerged victorious.
We shipped the project within weeks of our original target release date, earned a near-90 Metacritic from video game journalists, garnered hundreds of game accolades, and won 13 Overall Game of the Year awards. Most importantly, a wildly creative and cross-discipline team banded together to contribute to the unlikely revitalization of a classic game, capturing the magic of X-COM for a new generation of gamers and hardcore fans of the original alike.
</div>