Video GamesVideo Games

What the 2005 ‘Killzone 2’ Trailer Debacle Can Teach Us About Hyping Next-Gen Games

Fifteen years ago, an incredible trailer for a new PS3 game debuted at E3. One problem: It wasn’t made using the system’s engine. As we get ready to enter a new console generation, are there any lessons to glean?
Sony/Ringer illustration

Next week, millions of early-adopting gamers will get their hands on a PlayStation 5 or Xbox Series X/S for the first time. After they clear room in their entertainment centers for one of the big boxes, they’ll test out some selected titles. As they feast their eyes on the higher resolutions, faster frame rates, and decreased loading times that signal the dawn of a new console generation, some of those players will ask an age-old question: Does this game look as good as it did in the trailer?

Whenever a new console is due to come out, manufacturers hype their new hardware by showing glimpses of the games that they hope will be system-sellers. However, the hypefests often happen long before those games are in anything close to their finished state. “These initial demos are always a chance for the company to show you what a great job they’ve done of making the console,” says Iki Ikram, a longtime visual effects artist who has worked for Sony, Naughty Dog, and Guerrilla Games. “Of course, when the games finally come out, they’re not going to be as great looking as the demo.”

Related

In some cases, developers haven’t even received development kits for upcoming consoles when their games are announced, which means they’re forced to project what those titles could look and play like and fabricate facsimiles on PC before they’re actually running on the systems that they’re supposed to sell. Couple that projection with imprecise or misleading messaging about what a trailer actually depicts, and you have the ingredients for a first look that sets expectations too high.

That was the recipe for an uproar that remains one of the most memorable tech-related ruckuses in video game history: the scandal surrounding the 2005 trailer for PlayStation 3 shooter Killzone 2. And while some of the conditions that created that controversy are no longer in place, the Killzone 2 trailer taught a lesson that’s still applicable two console generations later: For better or worse, games often evolve significantly during development, especially at the beginning of the console life cycle. That labor-intensive development process may make the first glimpse of a game deceptive, although that doesn’t mean the game won’t be good.


The Electronic Entertainment Expo (more commonly called E3) in Los Angeles in May 2005 was all about the approaching seventh generation of video game consoles. (The PS5 and Xbox Series X/S will belong to the ninth.) Microsoft announced that the Xbox 360 would be out that fall, beating its next-gen competitors to market, while Nintendo teased a console called the “Revolution,” which would eventually be rechristened the Wii and debut in late 2006. Sony staged its own drool-inducing event by revealing the PS3, which would also launch in late 2006.

The company wowed audiences with footage from standout PS3 titles such as MotorStorm, Gran Turismo HD, Metal Gear Solid 4, and Heavenly Sword. But no demonstration compared to the trailer for Killzone 2. More than an hour and a half into Sony’s PS3 presentation, Rockstar’s teaser for Red Dead Redemption faded to black and gave way to a two-minute technical tour de force from Killzone developer Guerrilla. The original Killzone had been talked up as Sony’s answer to Halo but was widely seen as disappointing when it came out in 2004 for the PS2. But Guerrilla seemed to have raised its game in the trailer for the sequel, in which the point-of-view character joined an aerial landing and battled across a bridge amid a heated firefight that showcased next-level lighting and animation.

Matthew Rorie posted a gobsmacked reaction to the video for GameSpot, which would award its editors’ choice award for “best trailer” to Killzone 2 in an E3 roundup. (The Game Critics Awards, which annually anoint the industry’s E3 best in show, bestowed a “special commendation for graphics” on Killzone 2.) “We’re not ashamed to say that the Killzone 2 video has made many GameSpot editors’ jaws drop; most of us have watched it half a dozen times or so since it hit our booth,” Rorie wrote. He noted that the video looked good enough to be pre-rendered rather than running in the game engine, then continued, “If this is true in-game rendering, though, and again, we have no reason to believe that it’s not, then Killzone 2 is definitely one of the most impressive visual demos ever to appear at E3.”

There were so many people playing detective to try and get to the heart of the matter, but given the tools available at the time it was pretty much impossible.
Andrew Alfonso, IGN writer

Fifteen years later, Rorie explains why he included the caveat about pre-rendered footage. “It’s not at all uncommon for game publishers to release media well before a game comes out and try to make the trailers or screenshots match where they think the game will be at launch rather than where the game is at that point in development,” says Rorie, who notes that these fake or enhanced images or footage are colloquially called “bullshots.” The Killzone 2 trailer, he adds, “was a fair site better-looking than almost anything else at the show, to a degree that made it seem like Guerrilla had either a clear edge in technical development for the PS3 or was engaging in some forward-thinking rendering on how they thought the game would eventually look when it released.” 

Which was the truth? In the aftermath of the screening, conflicting comments led to dramatically different impressions. The next day, Mark Rein, VP of Epic Games (which presented a PS3 demo for Unreal Engine 3 at E3), said, “In addition to the Sony demos being shown by Phil Harrison, the Epic and EA presentations were the only third-party portions actually running on the PS3 in real time.” Harrison, the head of Sony Computer Entertainment Europe, seemed to acknowledge that the footage from Killzone 2 and other games was aspirational, telling gamesindustry.biz that “everything was done to PS3 spec. Virtually everything used in-game assets; some things were rendered.” Asked whether the videos were representative of how the games would actually play on PS3, Harrison said, “I think very.”

IGN writer Andrew Alfonso, who covered the controversy at the time, wasn’t fooled. “Seeing that Killzone 2 trailer was definitely wild,” he says. “I was with a group of IGN staff at the Sony press conference and we were 100 percent sure that had to be fake. The animations and physics were way too good, and it completely blew away everything at the time.” Compare the Killzone 2 trailer to that of Gears of War, which also debuted at E3 in 2005 but looked less smooth and detailed.

But Sony was sending mixed messages. A company rep told GameSpot that the Killzone trailer was running in real time. And Jack Tretton, then VP of Sony Computer Entertainment of America, told G4TV host Geoff Keighley, “It’s definitely real. I guess we’re pretty good at keeping secrets, because the dev kits were out there, the dev kits are very intuitive so people did some incredible things. And that’s one thing [Insomniac Games founder Ted Price] wanted to make sure everybody understood: That is real gameplay everybody’s seeing out there.”

Keighley followed up, asking, “So it is gameplay? All that stuff is all gameplay?” Tretton responded: “It is gameplay.” Cue the Arrested Development narration: It was not.

Guerrilla game director Jan-Bart van Beek tried to correct the record, telling the official UK PlayStation site that the trailer was “basically a representation of the look and feel of the game we’re trying to make.” But by then, many fans had formed the misimpression that the trailer depicted gameplay, and the debate became partisan, a proxy for the larger console wars. The game wasn’t actually released until February 2009, which allowed a lot of time for the back-and-forth to fester. “I remember following threads on various game forums and it was like a war zone,” Alfonso recalls. “There were so many people playing detective to try and get to the heart of the matter, but given the tools available at the time it was pretty much impossible.”

Tretton had inadvertently done as much as anyone to fan the flame wars. The Sony executive, who was promoted to president and CEO of SCEA in 2006 but left the company in 2014, now says, “I’ve had a lot of experiences over my 34 years in the Industry and the Killzone 2 trailer does not stand out in my memory. I don’t feel I’d have anything of value to comment on that situation. … It’s too long ago to have any informed perspective on it.”

Part of the problem, though, is that Tretton didn’t have an informed perspective on the trailer at the time. In the years since, various Guerrilla representatives have hinted at how the trailer miscommunication occurred. In 2008, a Guerrilla developer clarified that the trailer was a “target render of what we thought would be possible on the PS3” and blamed the false advertising on “one confused Sony rep” who was “blitzed out of his mind on fatigue, jet lag, and the madness that is E3.”

In 2010 and 2011, the managing director of the Dutch studio, Hermen Hulst (who’s now the head of PlayStation Worldwide Studios), largely confirmed that account. In his August 2010 comments as paraphrased by GameSpot, Hulst said that the trailer for the sequel, which was originally envisioned as a PS2 title, “was created with the intention of being shown in small, behind-closed-doors settings to select members of the press” but that “Sony showed so much confidence in the trailer that it wanted to air it to the entire gaming world at its press conference, and Guerrilla, being an ambitious studio, was proud to see its work on the big stage.”

Lastly, in a 2017 documentary about a later Guerrilla game, Horizon Zero Dawn, Guerrilla executive producer Angie Smets—who recalled that Guerrilla had received its first PS3 dev kit just before E3—echoed Hulst’s recollection and went on to describe the internal reaction to Tretton’s embellishment. “We were watching this back home, going, ‘No!’” Smets said. “‘What did he just say? It’s not true!’ Then we figured, nobody will believe that, because it’s obvious that it’s all [pre]rendered. Then we went online, and we found that lots of people believed it.”

As Hulst suggested in 2010 and 2011, though, the effect wasn’t all bad. It’s true that the hype put pressure on Guerrilla (which was acquired by Sony in 2005), but it also let the developers know that there was an audience for the game and gave them a lofty target to aim for, which may have helped them do their best work. Multiple Guerrilla veterans referred interview inquiries about the trailer to Sony, which declined the request. But former Guerrilla artist Dirk Boer, who joined the studio as a technical artist shortly before E3 in 2005 and was later credited for cinematics work on Killzone 2, says, “The main thing I remember is that it felt quite exhilarating. The feeling like before you go into a rollercoaster when you were a kid. Little bit scary, but definitely energized.”


Ikram didn’t join Guerrilla until January 2007, when he began working on Killzone 2 as a senior visual effects artist, but he backs up what Hulst and Smets said about the trailer being a pre-rendered, inspirational video designed for Guerrilla’s eyes only. Sony, he says, “needed to show content. So they showed that trailer.” But Ikram, who worked on the eventual in-game version of the mock cutscene that was shown at E3, says the staff at Guerrilla “were very well-spirited about it because they had already been aiming high. Now the pressure was on to try and provide results.”

Ikram, who has worked on blockbusters like Uncharted 3, Uncharted 4, and The Last of Us, says development targets are most subject to change when working for the first time with brand-new hardware. In some cases, developers simply decide to go in a different direction after trying one approach, which works out well at times. For instance, in its 2000 and 2011 tech demos for the GameCube and Wii U, respectively, Nintendo tried ultrarealistic takes on The Legend of Zelda that whetted fans’ appetites for games that looked like that. But the company opted to pivot to cel-shaded visuals for its next adventures starring Link, The Wind Waker and Breath of the Wild. Some fans found those decisions disappointing at first, but both games are among the best-loved entries in the series. The Wind Waker’s Toon Link,” which held up well long after a GameCube-era “realistic Link” would have looked dated, was a direct response to the tech demo artists’ realization that the initial style made a singular franchise look like every other game.

Many mid-development changes are driven by necessity, not creative epiphanies. Early in his career, Ikram helped design Dropship: United Peace Force, which was announced as a PS2 title before the PS2 was released outside of Japan. The game came out in 2002, and while it received mostly positive reviews, it wasn’t the epic creation that the team had conceived. “Our vision for the game was absolutely huge,” Ikram says. “But in the end, the technology dictated what we could do. And I think a lot of the reasons why the game turned out the way it did is because we had to sacrifice so much. Instead of having a smaller scope and focusing and concentrating and making it look amazing and feel amazing, we literally had to water down everything.”

That’s not an unusual developmental tale. Ikram says that while he’s never worked for a studio that intentionally made false claims for marketing reasons, it’s pretty routine for technological limitations to impose reductions in scale and graphical fidelity. During his time at SCEE, Ikram contributed uncredited work to The Getaway, a big-budget Grand Theft Auto competitor that was originally slated to launch alongside the PS2 but was delayed until December 2002 in Europe and January 2003 in North America because of the technical hurdles inherent in the game’s goal of re-creating London block by block. The game’s E3 2000 teaser video promised “over 50 square kilometers” of playable territory, but the end product included approximately 16. “It was scoped down because you just couldn’t put that on the PlayStation [2],” Ikram says. The studio’s “dreams were bigger than what the console could handle.”

The Getaway designer Chun Wah Kong acknowledges that not everything the team tried worked and that whittling the game down from pie-in-the-sky concept to shippable product involved “laying everything out and putting priorities on the key features, then balancing them out by choosing from a million features on a wish list pending time and resources.” But he says the prerelease hype was more of a blessing than a curse, even though the game (which sold well) received lukewarm reviews.

“It was brilliant to know the excitement it received from the press during production,” Kong says, adding, “to see the amount of interest the press had dedicated to it was definitely very motivational. I think it also gave our Sony executives the faith and confidence to keep feeding us.” He doesn’t lament the pressure to show the game publicly before it was clear that the system would support its ambitions. “Everything was possible, since we didn’t know the limits,” he says. “I would always prefer to aim high and try my hardest getting as close to it as possible than to set safe boundaries from the get-go. Everything’s impossible until it isn’t.”

According to Ikram, the resolution and look of the Killzone 2 trailer was probably impossible to match in a game engine running in real time prior to the PS4. To make Killzone 2 run on PS3, he says, “We had to sacrifice, obviously, a lot.” (On the PS5, he adds, “we could totally blow that away.”) The team at Guerrilla knew that the public would compare the finished product to the much-publicized trailer; as Ikram says, “That teaser was still in the minds and brains of everybody out there. So when the game was released, they were totally going to reference it. They were totally going to compare it.”

But Guerrilla wasn’t worried, both because the studio was confident that the game would deliver the frenetic experience the trailer had promised and because the game looked great relative to its competition, regardless of whether it lived up to the target render. Sure enough, Killzone 2 became one of the best-reviewed titles on PS3 and also went on to be what was then the fastest-selling first-party PS3 game, moving a million copies worldwide within two months of its release.

In the years since the Killzone 2 debacle, gamers and media members have gotten better at recognizing when developers are leading them on. “These days places like Digital Foundry can easily analyze trailers, so I guess the ‘magic’ of trailers is kind of gone,” Alfonso says. (Digital Foundry, a technical analysis site hosted by Eurogamer, has also produced a retrospective analysis of the Killzone 2 trailer and Sony’s PS3 reveal event, which concludes that “Sony went too far at E3 2005.”) Alfonso notes that the audience’s elevated technical knowledge “makes it harder to pull a fast one on people.”

But companies have also learned to restrain themselves until games are ready for their close-ups. When Guerrilla unveiled Killzone 3 in June 2010, it followed a stylized teaser trailer with a full-length trailer that Hulst specified was “made completely out of in-game footage.” A stage demo featuring early code soon followed at E3. Other companies, such as Ubisoft, have similarly adapted after painful first impressions. The publisher showed a trailer for Watch Dogs at E3 2012 that was based on assumptions of next-gen consoles’ capabilities and looked better than the disappointing final version released in 2014. Ubisoft CEO Yves Guillemot later admitted, “You want to create the images of what you dream. At the end of the day, we didn’t get exactly what we wanted.” The company dialed down the hype with Watch Dogs 2, which was much more successful.

Companies have also improved at being up front about what’s on the screen. Sony’s major misstep in 2005 was not divulging the nature of the Killzone 2 footage the way it did with the Final Fantasy VII tech demo at the same reveal event. These days, Rorie says, “early trailers are more likely to be clearly labeled as CGI if they’re not representative of gameplay.” But the precise description still matters.

“When you see on the screen, ‘running on a PS4,’ that doesn’t mean that’s the game,” Ikram says. “That means it’s still running in the engine. It’s just a hyped-up version of it. So for scenes like that, for cut scenes, we can spare a lot more geometry, particles, effects, everything, because we’re not running the game per se. When you’re running the game, it’s a whole different ball game. So the language is very important.”

As always when a new console generation begins, gamers should temper their initial expectations for launch titles, some of which are rushed to capitalize on new owners’ need for software. Certain studios, Ikram says, employ people who can “really reverse engineer the console and find out where they can pull all the power,” allowing them to improve on the previous generation at or close to launch. But most studios don’t, which means that the best-looking games typically appear later in consoles’ life cycles. What’s more, some new systems are more like iterative upgrades than revolutionary leaps, and the visual jumps between generations aren’t what they were when games were going from eight bits to 16 bits or from 2D to 3D. But that’s because almost every new game in 2020 would have looked miraculous in 2005.

“Games in general now are approaching, if not photorealism, then at least a level of graphical quality that at least rivals really good CGI trailers of the past,” Rorie says. “So the need to massage the way they look for trailers might not be as imperative.” In other words, we’ve become accustomed to playing in virtual worlds that were once too pretty to be true. 

Ben Lindbergh
Ben is a writer, podcaster, and editor who covers culture and sports. He hosts ‘Effectively Wild’ at FanGraphs and previously wrote for FiveThirtyEight and Grantland, served as editor-in-chief of Baseball Prospectus, and authored ‘The MVP Machine’ and ‘The Only Rule Is It Has to Work.’

Keep Exploring

Latest in Video Games