The Left 4 Dead co-lead looks back on "paying for the debt" of the game's broken engine—and how that led to a rapidly-developed sequel.
As he recounts his work producing and managing the original Left 4 Dead game launch, released 15 years ago this week, Chet Faliszek pauses our conversation at his Seattle home office to double-check an exact date. "September 7, 2005, at 1:10 p.m." This was the exact moment he was formally introduced (via e-mail) by his boss, Valve co-founder Gabe Newell, to Turtle Rock Studios founder Mike Booth, to “consult” on a new, Valve-funded video game project. Its codename at the time was “Terror.”
Faliszek was a novice staffer at Valve at this point, joining the company earlier in the year, but he had a good hunch about this lead. Speaking in both the past tense and present, Faliszek offers the following about the man at the other end of that e-mail chain: "If Mike Booth said he was releasing a video game tomorrow, I would buy it.”
The development process that followed began on an incredibly small scale, with a much different default playstyle, and the project ran into bumps, snarls, and iconic, frothing-zombie screams before launching as Left 4 Dead a little over three years later. On the eve of the legendary, horror-fueled co-op game’s anniversary, Faliszek sat with Game Developer to talk about its development process–and the learnings he’s taken in making his own Left 4 Dead-like game in the years since.
When it launched in 2008, Left 4 Dead created a monumental shift in the online game space. It was the first truly dedicated co-op shooting game, and in a rarity for the time, it was tuned to be impossible to beat without help and cooperation.
In the years that followed, major series like Halo, Gears of War, and Call of Duty introduced their own "collaborate, shoot, and survive” modes (with CoD outright lifting the zombie conceit), but Left 4 Dead put an indelible mark on the concept that remains popular to this very day. Yet Left 4 Dead started in a much different shape than how it eventually turned out.
The standalone game studio Turtle Rock, headed by Booth, originally supported Valve in an outsourced capacity. After working on an expansion pack for Valve’s tactical shooter series Counter-Strike, Turtle Rock remained on Valve’s payroll to work on the series’ computer-controlled enemies in its offline, single-player modes, along with a focus on a console-compatible, higher-fidelity version of the game dubbed Counter-Strike: Source. But Valve budgeted additional payments to Booth and Turtle Rock for work that Faliszek describes as “this kind of outside person experimenting and pitching other [video game] projects.”
One of those projects was a zombie-themed offshoot of Counter-Strike, code-named "Terror,” which Faliszek says began life as a mod for the newer CS:S engine. In the mod’s earliest tests, a group of human-controlled players faced off against waves of AI-controlled zombies. "Mike was an AI guy, so this was a logical extension,” Faliszek says. (If you’ve kept up with Left 4 Dead history, you might have seen this initial Terror test in the form of a leaked CS:S map and assets.)
Terror’s gameplay evolved during its pre-Valve prototyping phase until Faliszek and Booth were introduced, at which point the game had settled into something a little more CS-like: four players controlling human survivors with an objective (run all of the way through a map, or set off tanks of gas), versus four players controlling constantly respawning, super-powered zombies, with the latter getting help from waves of weaker, AI-controlled undead.
Image via Valve
Everyone in the earliest conversations seemed to agree that out of Turtle Rock’s prototypes of the time, Terror had the most potential as a full game that Turtle Rock and Valve would collaborate on going forward.
Faliszek suggests a substantial period took place between Terror’s limited prototyping phase and Valve’s larger involvement in what would eventually be renamed Left 4 Dead. While working primarily as a writer between 2005 and 2007, Faliszek additionally "playtested anything that came through [Valve’s] doors that I could." He recalls being immediately intrigued by Terror.
"The first time I played Left 4 Dead, I told Gabe, ‘This was the game I’ve always dreamed of,’” Faliszek says. He immediately recognized two touchstones in the game’s early prototypes: “midnight horror, shit-show craziness” films like Dawn of the Dead, and his early experiences with the FPS classic Doom.
Faliszek recalls a mid-‘90s group of friends “playing that as a four-player co-op game by sitting in a warehouse in a super-bad neighborhood in Cleveland, with a case of beer between us, all back-to-back playing Doom all night, to, like, 5am. And then we were scared to go down to our cars, because we'd gotten so amped up.”
With this excitement, Faliszek began casually dropping hints about Terror to Valve colleagues and showing them what state it was in, suggesting they join the project. (Valve famously relies on a "flat” organizational structure even to this day, allowing someone like Faliszek to ask colleagues to wheel their desks toward a new project if it excites them.) Turtle Rock’s internal work on the game ramped up enough to merit a November 2006 announcement by Valve, along with a decision to rename the game Left 4 Dead.
Turns out, one of the original leading names for the game had been “Blood Run,” which Faliszek personally resisted, asking for “one week” for him and his fellow writers at Valve to come up with something different. "Erik (Wolpaw) came up with ‘Left For Dead,’” Faliszek says. “Marc [Laidlaw] added the "4.’”
Yet it wasn't until roughly a year after that announcement that Valve’s own internal focus and staffing on the game seriously expanded. Faliszek recalls the period after The Orange Box shipped, when various employees peeled off to work on the Directed Design Experiments–a multi-month prototype initiative at Valve meant to foster creativity and new ideas.
Image via Valve
"I'm the only person who didn't do the design experiments at first,” Faliszek remembers. “And I didn't do them because I wanted to go work on Left 4 Dead.”
This is when Faliszek began more aggressively enlisting Valve staffers to contribute to the game that Turtle Rock had been leading. He specifically recalls Valve artist Randy Lundeen joining Left 4 Dead after finishing his DDE work–and how Lundeen understood the game’s horror-film vibe instantly. "The first thing Randy did was in the No Mercy [level] apartment building, having it look like these inner-city run-down apartments like in the original Dawn of the Dead,” Faliszek says.
Once he’d attracted a dedicated team, Faliszek and this team relocated to the 11th floor at Valve’s office, away from other colleagues. "We tried to make it our own thing for a little bit, just so we could concentrate,” Faliszek says.
This headcount initially grew to roughly 20 Valve staffers and 6 employees at Turtle Rock, with email serving as the primary comms tool between Bellevue, WA (near Seattle), and Lake Forest, CA (in Orange County). Faliszek suggests that limited, focused communication ultimately worked out, even as the Left 4 Dead team grew to Faliszek’s estimate of roughly 100 at Valve’s Bellevue HQ and 10 at Turtle Rock, so long as both teams could regularly engage in an out-loud manner.
Faliszek says in Left 4 Dead’s case, a remote-friendly playtesting environment facilitated regular, organic communication for the split-up offices. Nearly every day, one of the offices (usually Valve’s) would host a playtest, with the resulting gameplay being beamed to a video-link conference environment in both Bellevue and Lake Forest.
Doing this meant that relevant staff members weren’t in the same room as the players, so they could talk more freely about the game’s human-versus-zombie face-offs. More importantly, Faliszek says, regular real-time conversations with gameplay as an anchor "helped the team feel that they were all on the same page. You feel unheard If, after a playtest, you can't voice the problems that you saw, or talk about your ideas and how to make things better.”
Before Left 4 Dead’s team had grown to over 100 staffers, Faliszek recalls the moment when he was inadvertently christened as Left 4 Dead’s project lead on the Valve side.
"Aaron Seeler was the engineer that was helping do the Xbox 360 version, because [Left 4 Dead] was gonna be our first Xbox game originally. Gabe was asking him some questions, and [Aaron] was like, ‘Well, you should just talk to Chet, Chet's kind of the person.’ Like, no one ever told me I'm running it!” Yet from that point, Newell made Faliszek the point person for Left 4 Dead’s production and management, along with constant contact with Booth as the game’s original creator and director, up until its launch.
Image via Valve
In his sudden ascent to a major player in Left 4 Dead’s development, Faliszek found himself feeling over his head with a particularly unsuccessful playtest: the first with the entire Valve staff. Left 4 Dead’s team had reached roughly 25, with much of the company “coming in cold” on the game’s concept (remember, its early builds had enjoyed some seclusion on the 11th floor). Before the playtest even began, its matchmaking system was crashing, even on Valve’s internal network, and Faliszek remembers his boss Newell pouring fuel on the situation:
“Gabe just asked me very seriously, "why don't I just play Counter-Strike instead? That's working.’ And I thought he was joking. Like, ha, ha, ha. But then he made me watch him play it, he made me sit in his office and watch and play it. He says, "No, seriously, this is every single one of your customers, asking this question right now. What do you tell them?' Like, fuck, good question.”
The internal test didn’t get better once the matchmaking system was resolved, as Faliszek’s concerns over the game's emphasis on four-on-four versus combat–again, four human survivors against four super-zombies–were inflamed by the Valve staffer test running into crushing difficulty. "What we found out was, it was really hard to balance the game, because one really good player on the infected side could ruin the game for the other team,” Faliszek says. “Also, I didn't pick the team that [Newell] was gonna be on. I thought they would play better with him, and they didn't. It was a mess.”
Faliszek doesn’t remember exactly when the team shifted away from a four-on-four versus default, and towards co-op play against only AI enemies, but it was around this time that the game’s core philosophy and systems began to gel together into a more confident product.
"The underlying idea that Mike [Booth] had was, ‘lone wolves die,’” Faliszek says. "That was the knife that everything [in Left 4 Dead] was cut on.” He remarked that the game’s unique tension came from dire consequences in a zombie apocalypse, something that didn’t feel as potent in die-and-retry video games.
"I remember at one point an animator at Valve telling me that we would never ship the Hunter, we simply never would.” Faliszek adds. "Because nobody will ever put up with something that can just jump on them and take them down, and you have no way to get it off. ‘You need to have a quick time event! You need to do something.’ And I'm like, ‘Nope! Nope! we're gonna ship that! I'll bet you money.’”
Faliszek lowers his voice. “I should've bet a lot of money, because he's rich.”
Shifting the game’s default mode from versus to friends-against-AI helped sell this hardcore approach to cooperative gameplay–where players knew that if one squadmate was pinned or restricted by zombies, everyone else needed to step up and save them. Learning and digesting this was a lot easier without having to factor in any frustration at an online opponent.
Still, the early game’s emphasis on versus contributed in part to the development of "super” zombies, who had more health and abilities than other basic zombies. This aspect remained all the way to Left 4 Dead’s final version, though zombie types and abilities constantly evolved during development.
This aspect of Left 4 Dead is arguably its most well chronicled, thanks to Valve adding an in-game commentary feature that shared details about its development history. (In particular, a super zombie dubbed the "Screamer" was removed from Left 4 Dead early in development, and its abilities were transferred in part to the exploding "Boomer.")
Faliszek indicates there was another super zombie that wasn’t previously chronicled: “The Hunter for a while would just go invisible if it didn't move, and then a player had to back into them or touch them [to make it appear again], which is just super aggravating for a player–like, 'what-did-who-now, dammit, he got away.'"
Faliszek recalls putting his foot down about other key Left 4 Dead design decisions, with one early argument revolving around whether the game’s human characters should be broken down into classes—a feature that would later show up in Alien Swarm. "I dearly love the team from Alien Swarm, but I don’t know how many times where I was told to make [Left 4 Dead] more like Alien Swarm.”
"I kept saying, 'no, that is the wrong way to go. This game is more mainstream and accessible.' I didn’t want players to need a specific class or else they couldn’t, like, open a door or have a blocking ability."
He also recalls an internal argument that was flared by having to show the game at a European gaming expo. "Germany has a law where you can’t hang people in a video game,” Faliszek says. “The Smoker zombie used to grab people by the neck.” In putting his foot down about moving the zombie’s choking point from the neck to around the chest, for the sake of global compliance, “I remember that was the first day that I heard, to my face, "Valve is ruining the game.’"
Image via Valve
"I understand where this comes from,” Faliszek continues. "It’s not meanness. You want to make the best thing that you’re in love with. But there was conflict there, we had to smooth it out. And of course, if you ask anybody now, they all think that the Smoker is choking you around the neck. Nobody really cares. It was such a non-event."
Other crucial tweaks and small touches settled into place over the course of the game’s development to make the co-op experience more enjoyable for average players. Each campaign eventually received break moments that required someone to hit a button or activate a door–which both raised tension for what was to come and allowed four-player groups to pause for a drink of water or a friendly conversation, since zombies didn’t pester players in the safe space before hitting the button in question.
Valve’s underlying technology, the Source engine, received at least three major engineering updates on its audio end during Left 4 Dead’s development to support Faliszek’s massive dialogue trees between all its main characters, which he insisted on to support a "bed of sound" sensation of dialogue in the game, "so that voices don’t come out of nowhere.”
(This is the main reason Left 4 Dead’s characters talk about reloading a lot, if you’re wondering.)
Speaking of audio: Faliszek confirms that many of the most crucial zombie voice clips in Left 4 Dead, voiced by Faith No More and Mr. Bungle lead singer Mike Patton, were sourced from his initial three-minute demo tape, perhaps more than the four hours of final audio he recorded with Valve in a studio later on.
He recalls Patton having a good knack for making zombie noises that had some nuance between the guttural screams and screeches. "You want to have a zombie sound alert, or all these other flavors, but you don’t want to be Scooby Doo,” he says. "You can’t have a zombie say, ‘Ruh-roh.'"
Faliszek points to a massive influx of Valve engineers and back-end programmers chopping away at the game in its home stretch to get it to its performance goal: a fluid frame rate, and up to 30 zombies on screen at any time on any map.
Preserving that high number of zombies onscreen is a major reason why zombie corpses disappear shortly after dying, and why some zombies might not have visible or detailed legs at times.
He also credits that same engineering team with getting versus mode back into the game after it had been jettisoned. Turns out, after tweaking general gameplay and super zombie types in Left 4 Dead’s co-op mode, the fun factor of versus had settled into something less frustrating–and Faliszek convinced a few pessimistic Valve engineers by putting together one more internal four-on-four versus playtest with only a few months of development time left. "I remember one engineer walking out, Brian [Jacobson], and saying, 'yeah, we have to ship this.'"
Image via Valve
Yet despite the team’s impressive work on optimizing the game for weaker PCs and the limited Xbox 360 console, Valve ran into one major technical brick wall.
"I don’t think outside people can appreciate how broken the Left 4 Dead engine was but still shipped,” Faliszek admits. "It loaded each map two or three times in the background.” As the game approached its certification window, an engineer took it upon themselves to try and fix this issue. Their solution didn’t work: During nearly every session with the fix, “a survivor just disappeared.”
This was the primary reason Left 4 Dead 2 was pitched as a full, standalone sequel. "Left 4 Dead was such a broken thing that nobody wanted to touch it,” Faliszek says. "That game iterated so quickly that if it meant breaking something horrible, where you had to load a map [two] or three times but you could playtest it today, we did it. That meant at some point, you had to pay for that debt. There was no way you were going to support mods for Left 4 Dead in the same way we did for Left 4 Dead 2 without a big reset."
When asked why Valve didn’t communicate this more clearly when the sequel’s announcement drew controversy, Faliszek is blunt. “When people kill themselves to ship a game, you don't really want to say that there were problems with it. It was a lot of patching and Bondo-ing to get it through the door. To be appreciative of that, I’d rather just have somebody mad at me because they thought it was my idea.”
After leaving Valve in 2017, Faliszek stomped forward with collaborative, multiplayer gaming as his primary interest, and nowhere is that better evidenced than his latest game, The Anacrusis, which exits Early Access on December 5.
This co-op, survivors-versus-aliens shooter owes a tremendous debt to the gameplay formula established by Left 4 Dead, and he admits that The Anacrusis swiped one idea that hit Left 4 Dead’s cutting room floor: random reward drops after the completion of each portion of a longer campaign. "The idea was right, but the implementation was wrong,” he says of Left 4 Dead’s stab at the concept.
"I knew the stats of how Left 4 Dead was played, and we saw that people predominantly played co-op, along with the 'Holdout' mutation mode [which resembled Halo’s Firefight, where players hunker down and deal with waves of enemies in a particular spot], and then just a little versus. So my focus [on Anacrusis] was, making it about friends hanging out together, playing together. What does that mean?”
For Faliszek, the answer in his newer game is smaller-yet-crucial tweaks inside of a similar mechanical core: Allowing players to revive a dead friend sooner, albeit with the penalty of a dangerous horde for doing so. Giving players more items to assist each other against waves of enemies, like shields and fantastical grenades. Having an AI "director” system that accounts for variables like player ability, amount of health, or needed supplies, then mixes up gameplay accordingly. (Here, Faliszek admits that the “most genius” development twist of Left 4 Dead was in naming its almost entirely random AI system The Director, and allowing fans to ascribe a personality to it that never existed.
“It didn’t really care about the player,” Faliszek now says. "It was a random number generator that kind of just worked its way out.”)
He admits that when he’s not working on or testing his own game, Faliszek still leans on Left 4 Dead... or, more accurately, its sequel. "It’s still fun. It still clicks. And there are so many crazy mods. Oh my god. You can have an understanding of how the game works and then turn it on its head, and it’s just fun and goofy. If you just let people have time with something for long enough, they’re gonna go insane.”
"But that had to be on Left 4 Dead 2. Left 4 Dead 1 would not have supported that. It would have been crashy, crashy, crashy.”
You May Also Like
