Portal 2

Sequels must balance a fine line between predictability and novelty. They must capture the essence of what made the original so great, while standing on their own as a complete experience. The first Portal succeeded because it surprised players. It took something vaguely expected--a challenging, dimension-altering puzzle game--and subverted that with dark humor, an emotional and physical rebellion of the artificial intelligence overlord, Glados, and the manifestation of a world beyond the immediate testing chambers. We expected to be trained in portal technology--we didn't expect to go an emotional journey through a science fiction narrative.

With Portal 2, we come with many expectations. At the end of Portal, Glados is destroyed, we escaped (or did we?), and Aperture Science stood as the enigma facility that it was. So how do you innovate on the concept of portal gameplay, testing chambers, aperture science, and the artificial intelligence overlord, Glados?

The beginning of Portal 2 gets this right. We are awakened by a new voice, that of Wheatley, in a foreign location. This creates an expository "gap-filling" in the player's mind, causing them to interact with the game narrative by filling in the missing details between Portal 1 and Portal 2: how did I get into this 1950s hotel room? Where am I? Who is Wheatley? And how does he function? This gets us interested in the narrative in the same way that Portal 1 interested us: Who/where are we? Why are we testing with portal guns (and so on)? If Portal 2 began outdoors, with the wreckage of Glados, then it would not be nearly as engaging, because that is the expectation. Portal 2 subverts that with a new, disembodied character in a foreign environment, and thus we are motivated to keep playing and alleviate this narrative dissonance.

Unfortunately, the game becomes more predictable from here on out. With the revival of Glados, she re-assumes her role from Portal 1, and expectantly takes vengeance upon us. Having Glados serve as our overlord does not offer us anything new, because we already experienced this in Portal 1. We know how she will act as our testing proctor, we know how the test chambers will proceed, and so this decreases our motivation as a player. Even though it was fun to fight Glados the first time, we evolved from that. We overcame Glados. We defeated her. And now the story and gameplay should move on.

Moreover, reviving Glados this early in Portal 2 disinterests players by removing that great "gap-filling" narrative dissonance. By entering the narrative this early, she is able to spout direct exposition at you, removing key mysteries in the narrative: What happened since she was gone? Was she really not even angry? Was she being so sincere right then even though you broke her heart and killed kill her? If you do not allow Glados to answer this by keeping her dormant, this leaves those open questions to the helm of Wheatley. This would make it much more interesting, since Wheatley speaks to you with the clout of unfamiliarity, not directly acknowledging the events of Portal 1. Continuing this would have made the story much more engaging due to the lack of narrative closure, motivating the player to continue playing and find out the answers to the dissonant questions between Portal 1 and Portal 2.

In contrast to reviving Glados early, a better way to have subverted player expectations in Portal 2 would have been to treat Wheatley as the main villain during the first act of the game. Instead of reviving Glados, simply input Wheatley into the dead Glados core early and have him betray you at that point. Then, leave Glados out of the narrative until the mid-section of the game when you explore Aperture Science's history. If you let Wheatley serve as the villain during the first act, it gives the player the firsthand experience of his villainy so that players can learn to despise him as a character. You can use him as a different test proctor to provide a new testing experience for players instead of rehashing Glados in the first act. Then Wheatley's evil nature would be directly impacted on the player's tests, motivating the player to overthrow him as the facility overlord. This makes the player much more motivated to defeat Wheatley in the second act by finding and allying with the player's previous enemy, (and the lesser of two evils) Glados.

Moreover, you can use Glados' second act introduction as a catalyst for the backstory of Aperture science. Just say that Glados has a reboot copy of her personality hidden deep within the mining sections of the facility. This gives a natural story incentive to explore these back sections of Aperture Science other than just, Let's run with Potato Glados, escape the underground, and defeat Wheatley while learning about the history of Aperture Science on the way! You can then tie in learning Aperture Science's personal history into discovering the location of Glados as a story means. Then, the back-story of Aperture Science would be embedded into the story goal of finding Glados' backup, rather than the side-juxtaposition it is now in the happenstance discovery during your mining escape.

Finally, the last act of Portal 2 subverts the player's expectations better, but by this time in the narrative, we can already see how the game will end. The original Portal was great because we started the game only with the expectation of completing test chambers. However, this was totally subverted towards the latter chambers, as notions of an escape crept in on us. By the time we switched our goals from testing to escaping and got to Glados' final test chamber, the realization that we were going to destroy her finally became a fulfilled reality. In contrast, in Portal 2, our expectation at the beginning is already to escape. Even after this is thwarted with Glados' resurrection, our goal never changes from this, even after Wheatley betrays us, and we get dumped into the back-sections of Aperture's history. Despite the fact that the during the last act, the game uses a great set piece (Aperture Science 1950s-1970s) and the new gameplay devices of the gels (which are implemented very intuitively), it is all progressing toward a predictable conclusion. Because of this predictability and the fulfillment of our expectations in our narrative goals, we are not nearly as motivated to escape as we were in the spontaneous rebellion in the last act of Portal 1.

CONCLUSION
So how do you successfully subvert player expectations when creating a sequel? Valve already did it with a different "2" game, Half-Life 2. That game took the original source material of Half-Life and expanded on it to create the oppressed, post-apocalyptic world of City 17, where aliens have enslaved the human race. It did not retread its old material at all--an alien invasion in a subterranean New Mexico facility--rather, Half-Life 2 used its source material as a foundation to go further, to explore new areas (Ravenholm, the Coast, Nova Prospekt), new characters (Alyx, Eli, Dr. Breen), and new forms of gameplay (the Gravity Gun) not restricted by the original concept of Half-Life. Half-Life 2 subverted our expectations as a player, and because of that, it broke new ground and made Half-Life 2 an arguably even better game than the first.

Portal 2, however, mostly sticks to the expectations from Portal 1: the return to Aperture science, the continuing use of test chambers, the vengeance of Glados, and more portal puzzles in a linearly progressing chamber sequence. We experience the same type of portal gameplay as Portal 1, yet unfortunately we don't have the succinct emotional beats of isolation, rebellion, and escape or even that sense of mystery which motivated us to escape. Instead, we experience the expected: a prolonged and failed escape, a continuous use of testing chambers, our submission to Glados, and a foreshadowed betrayal with Wheatley. Portal 2 does do many things right, such as the role-reversal of Glados as our inferior, the handling of the gels, and the ultimate relationship with Wheatley, but its failure to subvert our overall expectations results in a game that cannot match the same awe as in the first Portal and the surprises that it gave us.

Visually Directing the Player

A while ago in 2009 I wrote an article called "Visually Directing the Player" for HardyDev.com.

The article has been slightly updated for the UCLA Game Lab in February 2011 and was just posted online here.

In the past, the article has been used in Carnegie Mellon's Graduate Level Game Design Course (Spring 2010 and Spring 2011)
and also Michigan Tech University's Intro to Game Design Course (Fall 2010).

The most current version can be found at the UCLA Game Lab.

Happy reading!