It's hard make a good sequel. Whether you're working in film, games, or any other media, it takes a lot to recreate the success of a strong debut.
Valve had it particularly rough when it set out to make a sequel to Portal -- easily one of the most beloved games the studio has ever produced. At GDC 2012, Valve writers Chet Faliszek and Erik Wolpaw recalled the difficulties in making their 2011 follow-up, Portal 2.
"Lots of players thought [the original] was perfect," recalled Valve writer Erik Wolpaw. Audiences praised the game for its innovative gameplay mechanics, it's surprisingly well-crafted story, with many calling it a one-of-a-kind experience. Making a new game would be like making lightning strike twice, so why did Valve even want to do a sequel?
"The bottom line was that Portal 1 was made by a team of students [we hired from] DigiPen, and that left 150 people who didn't get to work on it," said Wolpaw. Valve decides its projects based on internal interest, and "there were a lot of people who wanted to work on a Portal game, and it was as simple as that."
What wasn't so simple, however, was figuring out what this sequel would entail. Valve wanted to surprise players all over again, and their first attempt at a sequel involved throwing most of the old concepts out the window.
In its early days, the only thing Portal 2 had in common with its predecessor was the Aperture Science laboratory itself. That meant it had a new protagonist, no Glad0S, and even no portals -- despite the obvious dissonance with the game's title. "We said we'd work that out later," Faliszek said.
Instead, the game used a brand new mechanic that Valve called "F-Stop" -- Wolpaw and Faliszek didn't explain what it entailed, as they hope it will be used in a future game.
"We felt pretty good, and we liked the individual component ideas," said Faliszek. "The feedback we got was that it was good, but it wasn't Portal 2."
Thus, Valve decided to go back to the portal mechanic, as well as the characters that players knew and loved. "Lesson one: you don't need to burn everything to the ground," Faliszek said.
The only problem with bringing back portals was that Valve had to teach new players how the mechanic worked.
"The problem was that 80 percent of Portal 1 was a training arc... we needed to teach players while still being interesting to old players," Faliszek said.
To streamline the training curve, Valve looked for new ways to provide crucial feedback to the player. For instance, Valve replaced the bouncing energy balls from Portal 1 with laser beams, which stayed in one place and always let the player know how they will affect the environment.
The team also reworked some tutorials from Portal 1, switching timing-based puzzles with switch-based ones that gave players more control over the puzzle chambers.
A major key to welcoming new and old players was the new companion character, the bumbling robot Wheatley. This new apparent sidekick gave Valve a means of communicating story beats, teaching new mechanics, and setting up the game's tone right at the get-go.
"Portal 1's story sneaks up on you. Players don't know there's a story till two-thirds of the way through, and you can't do that twice," Wolpaw said. "And Wheatley sets the tone right away, for instance, you know it's a comedy."
Wheatley, however, went through a number of iterations and story arcs over the course of development. Early on, Valve planned to kill him off at the beginning of the game, while introducing a handful of new characters along the way.
One such character was a philosophical robot dubbed the "Morgan Freeman sphere" that held incredible wisdom -- but only in terms of the 20x20 room in which he lived. Once the player took him beyond that space, "his mind was blown," as Wolpaw put it.
Eventually, Valve decided to stick with Wheatley throughout the game's narrative, and making him an eventual antagonist gave the game a satisfying and logical story arc.
Even with Wheatley's character fully developed, the developers still had one major problem: they couldn't come up with an ending.
"Portal 1's ending came late, but it was one of the most highly regarded sections, so we at least had that going for us," Wolpaw said. Given this precedent, Valve hoped to pull something together in a similar manner.
While working to come up with a conclusion, the team decided to create simultaneously several "joke endings," which would prematurely end the game if players sought them out. One such ending included shooting a portal to the moon, and players would asphyxiate in space as a sad song lamented their demise.
Unfortunately, those optional endings would have taken far too much work for minimal return, so Valve cut them from the game.
Eventually, the team realized that the cut moon ending would serve as a great climactic scene, and the team used it as inspiration to make the shipping game's final over-the-top ending sequence.
"It was a perfect mix of being totally awesome and completely stupid, said Wolpaw. "It was an obvious solution that was sitting there all along."
Reflecting on this experience, Wolpaw advised developers, "Try to give yourself enough time for the obvious to become obvious, because oftentimes the solution is buried within what you've already done."