Valve explains the reasons behind one of the greatest cliffhangers in gaming history.
Ahead of the release of Half-Life: Alyx, I spoke with level designer Dario Casali – a Valve veteran who has been with the studio since 1996 – about Half-Life 2, its development, and its influence on Valve’s new VR prequel. During that chat, we also discussed why Valve decided to make Episodes instead of a full sequel, and why the now infamous Episode 3 (and a sequel in general) never arrived.
His answer is, frankly, the clearest I’ve heard on the issue yet, but there’s still no single, simple reason it never manifested. It was partly due to Valve’s worry about “scope creep” in what were supposed to be smaller expansions, partly the studio’s desire to begin development on the Source 2 engine, and partly the lack of a creative spark (and unsatisfactory internal experiments) worthy of carrying the Half-Life name.
I initially asked Casali what lessons he felt Valve had learned from the development and release of Half-Life 2, and he says one of the main ones was that trying to build a game from the ground up while also developing the new game engine it was running on was a bad idea. “When we put Half-Life 2 out, of course it was a really long time,” Casali says, “that's six years, and we were developing the Source Engine alongside the game design.”
Casali tells me they had to throw out a lot of the work they had already done on Half-Life 2 as they experimented with what Source could do, played around with the physics system, and tried to push the limits of their new tech. “I think our main take away from that is ‘get some stable technology and then build a game on top of it,’” Casali explains, but the eventual completion of Source meant they were finally able to do just that - even if it did take longer than they had originally hoped.
“After working on Half-Life 2 for six years we decided we didn't want to go dark for so long. That's why we started doing the episodes where we thought, ‘well, we have the stable technology now. We understand the characters, we understand the story, we have most of the mechanics. Let's just bite off little chunks and then release more often. We think players are going to prefer that from waiting six years and going through however many delays we went through.’"
Of course, I pointed out the irony of him saying Valve disliked going dark for six years when the gap between Episode Two and Alyx ended up being more than double that, to which Casali jokingly replied “yeah, it's like we adjusted to an extreme” after they moved away from the quicker episodic format.
But, regardless of how it ended, a plan was set to develop and release each episode in a year, designing them as shorter additions to the story to keep players satisfied more frequently. That plan didn’t work out entirely as Valve had hoped. While Episode One was successfully developed in about a year, Casali says “scope creep” became a problem. “We found ourselves creeping ever forward towards, ‘Well, let's just keeping putting more and more, and more, and more stuff in this game because we want to make it as good as we can,’” he explains, “and then we realized these episodes are turning more into sequels.”
Episode Two actually took two years to make – Valve started work on it at the same time as Episode One. The plan for smaller, faster releases didn’t line up with the studio’s ambition for the project, and the scope of Episode Two increased past its original concept. After Episode One shipped, some members of its team even joined the Episode Two team to help out. “I think at that point we realized, ‘Okay, maybe this episodes thing, it was a good concept, but we're not executing terribly well as far as getting things out quickly enough,’” Casali explains, so the team started rethinking things after Episode Two.
So that’s why Episode 3 never arrived (though former Valve writer Marc Laidlaw did post what is seemingly a gender-swapped synopsis of what it could have been it back in 2017) but why did Valve’s re-evaluation result in an indefinite cliffhanger (and innumerable memes) instead of a proper Half-Life sequel? Casali ties it back to two things: the start of Source 2’s development, and Valve’s goal of making Half-Life games more than just another release.
Both Casali and Valve co-founder Gabe Newell explained to IGN that Valve uses Half-Life games explicitly to push technology forward and turn heads. In a new interview with our own Ryan McCaffrey, Newell said “Half-Life games are supposed to solve interesting problems,” and explained that Valve doesn’t want to just “crank Half-Life titles out because it helps us make the quarterly numbers.” Casali similarly says that they were “looking for what is going to make that next big impact” after Episode Two.
In the time since then, Valve has worked on loads of different projects: Steam, Dota 2, CS:GO, multiple VR headsets, and plenty more, many of which the outside world never saw. Casali confirmed something Valve has already publicly stated elsewhere, that some of those projects were Half-Life-based and never saw the light of day. He explains that, “we were never really that happy with what we came up with.”
Casali says Valve doesn’t move forward with projects that don’t seem promising or aren’t working out. “Our judge and jury is always the playtesting,” he explains. “It never comes from us. It always comes from somebody outside. And they always tell us how we're doing. And no matter what it is that we're doing, we get validated by that playtesting process, and we stick to that religiously.” Simply put, if we never got to play the Half-Life games Valve was messing around with, odds are we wouldn’t have wanted to anyway.
The other reason for the long delay in Half-Life’s return was the creation of Source 2, the follow-up to the Source engine used in Half-Life 2, Team Fortress 2, CS:GO, and lots of other games (including the Titanfall series). By the end of Episode Two, Valve was already looking towards its next engine, and had already learned the hard lesson not to develop both a Half-Life game and its engine from the ground up at the same time. “We [didn’t] want to make that same Half-Life 2 mistake again,” Casali explains, “of working on Source 2 and the next Half-Life game at the same time, because that created a lot of pain the first time we tried to do that."
To break the timeline down for you, Half-Life 2 was in development for six years, starting just after the first Half-Life’s release in 1998 and ending in 2004. Episode One followed roughly a year and a half later in 2006, followed by Episode Two at the end of 2007. At that point, Valve knew it wanted to make Source 2 and didn’t want to start work on a Half-Life game using it before it was ready – and knew it still wanted that follow-up to make an impact.
Seven years later, Source 2 was made available in Dota 2’s Workshop Tools in 2014 before the entire game was ported to the engine in 2015. Meanwhile, Valve tells me Half-Life: Alyx has been in development for roughly four years, allowing the studio to start working on it around 2016 with a Source 2 engine that Casali says was nearly complete by that point.