Having a locked script that defined characters, locations, cinematics and set pieces was great for development... once we got there. The challenge was getting there. For better or worse, I've always been Bend Studio's primary writer (often in collaboration with other writers).
Having written the scripts for all seven Syphon Filter games and for Resistance: Retribution, I thought I knew a thing or two about story and writing for video games. So when I played the first Uncharted I thought, rather naively, that it was going to be easy. It wasn't.
To start, the franchise formula is challenging. You have to find a compelling historical mystery with "gaps" that can be filled creatively; the history needs to be real and researched; you need a MacGuffin that is valuable, that people may have heard of, and that is potentially cursed; settings and plots need to be realistic, fantastic elements need to be reasonable in an X-Files sort of way; the tone should be mostly light in a summer blockbuster style; settings have to be traversable, exotic, romantic, and, well, uncharted in some way; there needs to be an army of bad guys to kill; the story has to lend itself to puzzles, twists, turns, surprises; bad guys have to be interesting, powerful, but not clichéd... and you have to do all this without repeating what has already been done in Uncharted 1, 2, and 3 (not to mention avoiding overused tropes from genre film franchises). I severely underestimated the amount of work this was going to be.
We began working on story ideas in 2009. The first version was a Drake origin story that started with him as a young dropout running scams on a college campus. He had met, and fallen for, an ambitious graduate student who'd been screwed out of an important discovery in the Grand Canyon (this character would eventually evolve into Chase).
The story involved grave robbing in Mesa Verde, lost temples in the Louisiana swamps, a museum heist during a hurricane in New Orleans and a boat chase through the Grand Canyon.
One of our earliest PS3 prototypes of the game took place on the New Orleans streets: we had built out an entire section which could be traversed, which contained our first AI proof of concept, and where we showed off our touch melee, SLR camera and interaction prototypes.
We spent a few months on this concept, working on scripts and concept art before we scrapped it. The issues were varied: First, Naughty Dog wasn't entirely comfortable with us filling in so much of Drake's backstory -- one of the dangers of working on someone else's franchise (we didn't know it at the time, but Naughty Dog would take Drake's origin in a completely different direction in Uncharted 3: Drake's Deception.) Second, the Southwest American locales weren't deemed "Uncharted" enough... it seemed like Drake would be bumping into tourists every time he climbed around a corner. And a minor concern was that the big action set piece of Drake fighting an army of bad guys in a hurricane-pummeled New Orleans was perhaps a little insensitive.
The second version of the story was more ambitious. Retaining the key Quivira plot and rainforest locations, we decided that maybe the PS Vita version of Uncharted should be completely different than the console versions. We decided to make Uncharted an open-world, hub-based game. The idea was that Drake was hired by a wealthy woman to complete her missing husband's research into the lost city of Quivira.
It would have been an open world where the player could have moved back and forth between his camp (where the player could interact with Elena and Sully), and various ruins sites, where he would have to compete for artifacts by fighting other mercenaries and local drug-running thugs. We spent a couple of months on this idea, completing concept art, prototypes, and actual mesh layouts before saner heads prevailed: Amy Hennig pointed out that it was out of character for Drake to be such a blatant mercenary for hire. He wouldn't pursue something simply because he was being paid to; he would have to be invested in the mystery in a personal way or the story wouldn't be Uncharted.
More importantly, we realized that turning Uncharted into an open-world, hub-based game might deviate too far from the established formula; players might not accept the changes (something we learned the hard way with Syphon Filter: The Omega Strain, which deviated from the Syphon formula substantially). We also realized that we had enough on our plate without trying to design and implement and open-world game system, with which we had exactly zero experience.
Slide taken from a January 2010 story presentation, the second iteration of Uncharted: Golden Abyss, when it was a HUB based game. Some of the location ideas made it into the final game. (Click for larger version.)
So we started over, salvaging as many elements from the previous drafts as possible. We spent all of 2010 iterating on the story. Locations came and went. Characters were created, iterated on, revised or discarded (see "Scope" section for more details). It wasn't until fall of 2010 that we had something close to final, and we didn't start casting until December 2010.
Looking back on the process, I'm not sure what we could have done differently. Not being the original writers on the franchise, it took a while to wrap our heads around what it meant to be Uncharted, to figure out the style, the pacing, the tone of the characters. Not being in control of the main characters' destinies made it challenging to come up with compelling conflicts and story arcs.
I wonder, with some regret, how much easier development would have been had we locked the story a year earlier. With an extra year to prototype environments, characters and movies, we may have been able to avoid the awful crunch that consumed most of 2011.
I mentioned earlier that once we got there, the locked script helped immensely by keeping production moving, and by keeping large changes to a minimum. But the locked script also hurt in an important way: once we discovered that the game was too big, we couldn't cut much more than we already had without leaving the story and pacing in shambles. Important character and plot developments happened in each world, and more importantly, key gameplay sequences. Even through the story we set out to tell in the beginning was too big, trimming it halfway through development would have weakened it, making game-play scenarios feel isolated and meaningless.
Pacing issues uncovered during the focus tests were hard to fix because we couldn't just move gameplay scenarios around. For example, there is a stretch of gameplay in Chapter 2 where the player has five touchscreen interactions in a row: a combination lock, Conquistador helmet, human skull, amulet, and grave marker, one right after the other. While we managed to add a combat scenario just before this, there was not easy way to spread these puzzles out and add different types of gameplay between them.
During the Conquistador gravesite sequence, the player has to perform five touch puzzles in a row, including the examination of this 500-year-old skull. The locked script and tight integration of story and gameplay made fixing pacing issues like this problematic.
Story-wise, there was no other place to put them, because this is where and when the actions took place. Gameplay-wise, the pacing suffered, but it would have been impractical and expensive to change. Like most sequences, narrative and gameplay were tightly interwoven and since the script was written, performances recorded, production well under way, we were stuck. This is a good example of where it would have been far better to have had a solid gameplay plan completed before the script was written.
Critical reaction to the Uncharted: Golden Abyss' story was mixed. Some of this was because I'm no Amy Hennig, who brings a very distinctive and natural style to her writing. But also, I think, players really wanted to see more of the characters they had grown to love: Elena, Chloe, and Sully. A new cast of characters, whether they were compellingly written or not -- and I let players be the judge of that -- was never going to compete with established characters from the console games.
There was strong resistance to Chase in particular because she wasn't Elena. Since this was a prequel, and players knew Drake didn't end up with Chase (she's never seen in the console games, after all), then what, some players wondered, was the point of spending time with her?
And we almost learned the hard way how much players love Sully. At one point in development, we had considered cutting Sully from the game because that would have enabled us to shorten the rainforest trek sequence (Chapters 22-27) considerably, cutting weeks from our polish schedule. Amy talked us out of cutting Sully, predicting something that players later confirmed: without Sully, it's not Uncharted. His relationship with Drake is core to the franchise. Even players who didn't like our story initially warmed to it once Sully came on stage midway through the game. I can only imagine the fan outrage we would have suffered if Sully hadn't made an important appearance.
2. Fusing PS Vita and Conventional Controls
PS Vita's Unique Control Inputs
While story and mission design were huge challenges and took months of iteration, designing controls for the many new inputs of the PS Vita hardware proved equally so. Jeff Ross, one of our Senior Designers, had this to say:
Jeff Ross: One of our greatest design challenges was to fuse the PS Vita's special inputs (front and rear touch screens, accelerometer, cameras, gyro, and compass) with conventional controls (buttons and joysticks) while simultaneously remaining true to the Uncharted franchise, and while delivering on gamer expectations of the hardware.
The cinematic action-adventure genre was crying out for tactile controls like touch and gyro: recall Raiders of the Lost Ark, where Indy had to do things like weigh out a bag of sand, touch artifacts with his bare hands to feel the engravings, or move and rotate the Staff of Ra into the light. It seemed to us that this was the perfect application of the PS Vita inputs for a franchise like Uncharted.
We had already played plenty of games on the Wii and iPhone that demonstrated the limitations of touch and motion inputs for more conventional gameplay like movement and combat -- cases where we would make use of the PS Vita's buttons and two sticks. But we felt our implementation of PS Vita-specific mechanics like charcoal rubbings, digital camera, codex, cryptex, jigsaw puzzles, and the light puzzle successfully merged fiction and gameplay.
We were even able to port existing franchise mechanics like Drake's journal and traversal to the PS Vita controls. Ranged combat benefited from subtle but integral gyro aiming that helps players fine-tune headshots. But some players and critics thought we went too far with the PS Vita controls, like when we added touch to the melee and boss fight mechanics. Our final analysis is that their strong negative reaction wasn't because of the mechanics themselves (in fact, anecdotal evidence shows that lots of gamers liked these features), but because we didn't give them a choice.
When we added PS Vita-specific inputs to other core mechanics, we gave players the option to play with sticks and buttons instead. For example, when we added touch inputs to traversal, we kept Uncharted's conventional climbing controls as well, so there was no downside to a player not liking it -- they could simply revert to the sticks and buttons instead. For aiming, the gyroscope didn't replace stick aiming, only enhanced it, and could be turned off by the player.
Touch melee, on the other hand, evaded our hybrid design approach, because focus tests didn't identify this mechanic as being a significant problem, and we saw that emerging iOS games like Infinity Blade used the mechanic successfully. In hindsight, we should have taken a global two-input method approach to every system that logically supported it. If you wanted to use touch? Great. Want sticks and buttons instead? Fine by us.
The Canoe Sequence
The inclusion of a player-controlled canoe sequence required us to design something without any existing standard to draw lessons from or to compare our results against. A kayaking mini-game was found in Wii Sports Resort, but that game focused on large body movements simulating paddling, so we couldn't draw any meaningful lessons from that. We were on our own in deciding what the first action adventure canoeing experience would be.
The only thing we knew for certain was the plot, which required Drake and Sully to quietly infiltrate the enemy-filled rainforest, and the basic human-powered canoe was all that fit their budget. Over the life of the project, the canoe implementation took on many forms from rail controls to a full-blown canoe simulation. The final version wound up employing a variety of techniques including railed navigation, player driven propulsion, calm water and white river rapids, a frantic waterfall escape, and an explosive combat sequence where players could play as Sully for the first time in franchise history.
Screenshot taken from an early version of the river level. In early prototypes of the canoe, the player could steer the canoe left and right, controlling the paddles by swiping the touch screen.
The full canoe simulation was a lot of fun, but was cut from the shipped game because it would have required too much player instruction and practice: by the time they learned how to operate the canoe effectively, the mission would have been over. This was disappointing from a design point of view, but was determined to be the right thing to do after we watched a small sample of users attempt to paddle up stream.
In the simulation, the canoe was steered by players alternately swiping up and down on the sides of the screen. To some, swiping on the left side of the screen meant to turn the boat to the right (which was the intended effect), other players assumed that swiping left meant steering left, and no amount of training could reverse this instinct.
We could have included an option to invert the controls, but there turned out to be an even greater problem that was too big to solve: the touchscreen swipe rate. Some players would gently alternate swipes on each side, resulting in straight navigation at optimum speed. These were our target gamers, the ones who could effectively control, and therefore appreciate, the simulation. Other gamers would spam short swipes on one side of the screen before realizing they were turning undesirably and they would frantically switch sides to try counter steering, often times running aground and then having to reverse their way back into the water. These problems could have been solved with a lot more training time and with helper features like "auto returning" a grounded boat to the center of the path, but given the relatively small part the canoe played in the context of the entire game, we just couldn't dedicate the time or resources to making the simulation aspects work.
We're still quite proud of the canoe's final implementation, as it was the proper balance between narrative and interactive immersion. Most gamers probably won't ever question if they're missing an awesome canoe simulation. Toward the end of the project, we questioned whether the months of development time for such a short section of gameplay was worth it, but I think if you play it, you'll agree that it's not only fun to play, but absolutely gorgeous. And who can argue against spending some time with Sully, drifting along in a canoe?
Screenshot from the final game. You can see the swipe icon to the right of the canoe. What the player lost in terms of freedom of control, they gained in freedom from frustration.