The latest issue of Gamasutra sister publication Game Developer magazine
, available for subscribers and for digital purchase now, includes a postmortem of Naughty Dog's Uncharted 2: Among Thieves
, written by the studio's own Richard Lemarchand.
, a "cinematic action game" acclaimed for its detailed graphics and compelling story, is the sequel to the Sony-owned studio's 2007 hit, Uncharted: Drake's Fortune
. It was published by Sony Computer Entertainment America for PlayStation 3 last October.
These excerpts from the March 2010 issue of Game Developer
magazine reveal various "What Went Right" and "What Went Wrong" highlights from throughout the creation of the game, revealing how the company employed considerable playtesting and a "get it done" attitude to overcome some planning issues.
Getting On With Making The Game
While preproduction can be important, Lemarchand notes that it is equally important to allow the team to get right along when it comes to actually starting development:
"The most important lesson that we took away from making Uncharted: Drake’s Fortune was one we had to learn the hard way. We spent too long making plans and not enough time simply getting on and building things. This led to a crisis that resulted in a mid-project production reset.
"We’d lost sight of the fact that theorizing about process and tools can only take you so far, and that it’s only when you build something—whether it’s a game mechanic, a tool, or a level—that you make the really valuable discoveries about what you’re doing.
"When we set out to make Uncharted 2, we kept this idea at the forefront of our minds the whole time and it served us well. For example, we shifted our level design process away from paper layout and toward iterating on prototype levels in simple “blockmesh” geometry. Our game director and one of our game designers would first sketch out an experiential flow for the player.
"The designer would rapidly build out an environment with a low level of detail to test on other team members, so we could see how navigable it was, what camera and line-of-sight issues arose, how long the experience would last, and so on. We would then start scripting interactive objects and placing enemies, and eventually give our art team the all clear to begin creating final art.
"This approach let us build out the game’s footprint very quickly, although it wasn’t without some dangers: we ran the constant risk of becoming too committed to level designs that might need changes demanded by the maturing story.
"We thought on our feet about the order in which we should tackle our new and expanded gameplay systems, and started with the ones that would have the most wide-reaching effect on the game. We took a similar approach with our tools and engine improvements, tackling the things that would give us the biggest leg-up first. By the time we finished with all the changes and improvements we made, we felt like we’d virtually reinvented our engine, and therefore dubbed it the 'Naughty Dog engine 2.0.'"
Not Quite Enough Planning
As noted, Naughty Dog likes to get right to the game-making. But while that philosophy has its strong points, it can also lead to some planning quirks:
"One of the downsides of our philosophy of simply getting on and building the game was that the line between preproduction and full production became blurred. We hadn’t really begun to plan ahead until we finished work on the first Uncharted, so we scrambled to solidify as many elements of the game’s content and story as we could in order to stay ahead of the team as they started building assets that we hoped would find a home in the shipping game.
"The story team made a lot of key decisions in a timely enough manner to provide a framework for our forward motion. What emerged from preproduction was our focus on Asia as a location for much of the story, Marco Polo’s lost fleet as our real-life historical mystery, the idea of an old friend of Nathan Drake’s who would ultimately betray him, and some big chunks of the game’s macro design.
"However, even in the absence of a story structure to frame them, the first levels took on a life of their own. Their footprints grew and their gameplay firmed up to a good degree. When the game macro was finished in the spring of 2008, the story beats that related to parts of the end of the game were still a little fuzzy. We couldn’t quite decide how the threads of the story would twine back together as we neared the end of the game. We eventually worked it out, of course, but we couldn’t quite fix all the issues that had arisen.
"So even though a lot of people who play Uncharted 2 don’t notice anything amiss with the end of the game, when we play it through we feel that there aren’t quite enough strong story beats in the monastery to match the length and intensity of the gameplay there, and it’s the first place in the game where the pace begins to flag. Hopefully we’ll learn the lesson of this minor misadventure and pay special attention to pacing issues for parts of the game whose level design starts early and whose story design finishes late."
Playtesting And Metrics
If there's one piece of advance that remains constant in Game Developer
postmortems, it's not to underestimate the value of playtesting. Naughty Dog is no exception:
"In the course of making Uncharted 2 we did more formal playtesting than we’d ever done before. We ran fifteen playtests over the last ten months of the project, compared to seven over the whole three years of the first game’s development. This resulted in fewer rough edges in gameplay than in any game we’ve ever shipped (although a couple still snuck through!).
"We ran most of our playtests in a rather jury-rigged but functional playtest room in the Naughty Dog office. We had ten TVs, each with a PS3 test station that was hooked up via video capture boxes to a PC, to record events on screen. We didn’t record video of the players’ body language, though that would have been a good addition. The TVs had 2’ by 3’ pieces of card bought at a stationary store propped up between the TVs so that the players couldn’t, even accidentally, see what their neighbor was doing in the game.
"Running our playtests in-house had the enormous benefit of allowing all of our designers and QA leads to regularly see their levels in action with new players. Of course, there are few things better for a game’s design than for the designers to watch it being played by people who have never played it before.
We got our playtesters to play through as much of the game as we had finished building, even to basic levels of completion. We didn’t let them talk to each other and were merciless about not giving them any help when they were stuck, unless we knew that something was broken.
"As they played, we uploaded metrics about their actions to a database over the network—things like how long it took them to complete each part of the game, or how frequently they died between continue points. We put the data into a spreadsheet and looked at the median values for each group. After color-coding cells with values above or below certain targets, parts of the game that were potentially problematic immediately jumped out at us. We then started looking at the gameplay videos to investigate each potential problem.
"Doing so much playtesting was particularly important because several complex sequences of gameplay only came together very late in the project, and to ensure that the things we added or changed didn’t present unforeseen problems for our players, we conducted “sanity check” playtests right up to the end."
The full postmortem of Uncharted 2
explores more of "What Went Right" and "What Went Wrong" during the course of the game's development, and is now available in the March 2010 issue of Game Developer magazine
The issue also includes a roundup of governmental game development incentives, Front Line Award finalists, a piece on the art of creating believably flawed characters, and our regular monthly columns on design, art, music, programming, and humor.
Worldwide paper-based subscriptions to Game Developer magazine are currently available
at the official magazine website
, and the Game Developer Digital version of the issue is also now available
, with the site offering six months' and a year's subscriptions
, alongside access to back issues and PDF downloads of all issues, all for a reduced price. There is now also an opportunity to buy the digital version of this edition as a single issue