As always, you may click on images to view larger versions.

I’ve played Portal many times, but I only ever played Portal 2 once, back in 2011 when it was released, shortly before starting this blog. My decision to return to it now was not actually prompted by the game’s recent 10th anniversary — or perhaps it was, but indirectly. A friend mentioned they’d replayed it (they did not say if the anniversary prompted them to do so) and sung its praises, telling me that they hadn’t thought anything could be better than the original game, but Portal 2 managed to do it. This surprised me, because my own memory of Portal 2 was that it was good but couldn’t quite live up to the high bar set by its predecessor. But maybe I was wrong? After all, I’d never gone back to Portal 2 like I had with the first game. Maybe, playing it again now, I’d find myself admitting that it surpasses the first Portal. I decided to find out.

I mostly want to write about the narrative structure of the two games, so in a break from tradition this post will be full of spoilers for both games. Read on at your own risk!

In 2007, when the original Portal was released, it was unusual for a game to be so short. Games that only take a few hours to play are far more common now, but at the time it was a novelty, and also one of Portal’s biggest strengths. Players got to test out the titular portal device, which can “fire” two types of portals — cyan and orange, of course, because in videogames no other colors exist — at walls, floors and ceilings to make an instantaneous passage between them. This alone provides the basis for a set of fabulous navigation puzzles, but Portal smartly ends while players are still excited and engaged by the portals, before tedium has a chance to set in.

The lateral thinking required to deduce solutions using the phsyics-defying portals is enough to entertain players for a good while, so Portal’s narrative is initially thin. The protagonist, Chell, is in some kind of high-tech facility covered in pristine white paneling (it soon becomes clear that the white panels are needed to support portals, which will not “stick” to other surfaces), and proceeds through a series of test chambers while being guided by an artificial intelligence, GLaDOS, speaking over a PA system. It seems to be some kind of training course, testing Chell’s ability to traverse various obstacles using the portal device. Perhaps Chell knows why she is being tested, but players do not; I assumed that portal devices have some advanced real-world application, and operators must be trained to use them effectively before they are cleared for work. As the game proceeds, however, it slowly becomes clear that something more sinister is going on. Test chambers start to feature deadly hazards, and GLaDOS’ initial detached, emotionless demeanor transforms into deliberate malice. As this narrative curtain lifts, so do players break free of the carefully controlled test chambers, navigating maintenance shafts and corridors that they weren’t meant to see as they take the fight to GLaDOS itself. Oh, and did I mention that Portal is also hilarious? It’s filled with dark humor based around GLaDOS’ inability to understand humans, plus a few jabs at the dangers of unchecked capitalism.

All that, in just a few hours. Even just summarizing it here brought to mind a full-length videogame plot, stretched over a dozen hours at least, but if Portal had done that it would have been far less effective. Manipulating portals remains fun and surprising throughout, and the later segments brilliantly let players use their newfound portal mastery to wrest control of their environment from what seemed to be an all-powerful antagonist. Sure, these sections are still puzzles that the designers made for players to solve, but they feel subversive, and as such are a wonderful capstone to the short game. If it had gone on longer, and traversing environments via portals became routine, the magic would have worn off before the end.

I remembered Portal 2 being much longer than Portal, and suffering for it. Replaying it now, I found that it’s not as long as I thought. The first time through, I spent more time stuck on some puzzles, which really messed with the pacing. It is still longer than the original, however, and packs in more story with a larger set of characters. The writers were in the unenviable position of following Portal’s brilliant exploration of portal-based puzzling and a pitch-perfect villain in GLaDOS, and the only viable way to do so is to introduce a bunch of new stuff. Portal 2’s opening is striking, casting Chell (who had been in stasis) far into the future, where the subterranean Aperture Science laboratory lies in apocalyptic ruin. The opening segments are almost entirely devoted to scene-setting, with players occasionally firing a portal or two just to cross otherwise impassable debris, as they are introduced to their new guide, a personality core named Wheatley.

I love the aesthetic of the crumbling facility. The pristine test chambers from the original game are now overgrown with ivy and moss, filled with collapsed scaffolding and detritus. These places deftly avoid repeating Portal’s look, while remaining recognizable as the Aperture facility. It’s a strong opening, made somewhat awkward when players find themselves solving (half-ruined) test chambers again. The original Portal kept everything beyond the test chambers well hidden, only revealed at key moments for narrative effect, and players were only let loose in these forbidden areas in the finale. But that trick can’t work twice, and indeed Portal 2 begins in those same corridors and maintenance shafts, the veil lifted from the very start. That players must re-enter and solve test chambers in order to proceed to the next part of the story feels remarkably artificial, in these circumstances.

This problem is only exacerbated by the reintroduction of GLaDOS. Reactivated accidentally, GLaDOS’ appearance should be a terrifying moment. But in Portal 2, GLaDOS is never scary. Chell finds herself at GLaDOS’ mercy, with nothing to stop GLaDOS from exacting revenge… by making Chell do more test chambers? They’re not even particularly dangerous or difficult test chambers, since it’s still early in the game. What follows is just a set of puzzles punctuated with constant passive-aggressive insults from GLaDOS. It’s an attempt to recreate the humor of the first game, with GLaDOS trying and failing to craft emotionally devastating jabs, still not quite understanding the human psyche. People don’t like it when you call them fat, right? Or if you make fun of them for being orphans? The insults are reminiscent of children at a playground, and only serve to make GLaDOS seem toothless. The first game was darkly humorous, but GLaDOS still felt dangerous. No longer. This may be Portal 2’s greatest mistake.

These test chambers even look bland, because GLaDOS starts cleaning up the place, removing the vegetation and grime that made the opening so visually striking. Fortunately, players break free from GLaDOS’ endless testing routine, because of course they do. This leads to a pale imitation of the original Portal’s finale, not just because players have already been traversing Aperture’s hidden corridors and forbidden passages already, but because this time Chell isn’t taking matters into her own hands, she’s just following Wheatley’s orders. Of course, this isn’t the finale of Portal 2, it’s just the beginning. Wheatley is installed in GLaDOS’ place, but quickly becomes drunk with power, sending Chell plummeting into the bowels of the facility instead of letting her go. He also reduces GLaDOS to a tiny chip powered by a potato battery, dropped into the depths along with Chell. Before the fall, GLaDOS recognizes Wheatley as a personality core specifically designed to come up with terrible ideas, as a (failed) attempt to limit GLaDOS’ power. The game itself has undermined this revelation, however, by depicting Wheatley as far more resourceful than GLaDOS. Sure, Wheatley is a bit inept and clumsy, but he’s constantly improvising new solutions, and he’s prepared well to face GLaDOS, leading Chell to sabotage the turret assembly line and neurotoxin supply so GLaDOS cannot use them. All GLaDOS has demonstrated are weak insults.

The next segment of Portal 2 is the biggest departure from the original game. There were new puzzle elements in GLaDOS’ test chambers before this point (redirectable lasers, light bridges) but the puzzles still felt familiar. Now, Chell finds herself entering a massive, sealed vault deep within the facility, discovering the original Aperture Science testing area from the 1960s. In stark contrast to the reconfigurable paneled rooms above, this is a vast, open cavern that dwarfs the testing apparatuses within, which are built from simple metal scaffolding, plywood, and concrete. I love the look of this place, scrappy and practical and completely lacking the cohesive and dare I say branded visual style of the modern test chambers. I loved traversing huge distances in an instant by firing portals at barely visible patches of white wall on the other side of the cavern, and emerging in control rooms complete with racks of vintage electronics and 1960s style furniture. And I enjoyed the new puzzle elements, in the form of gels.

Just as the original Portal famously adapted ideas from the student project Narbacular Drop (by hiring the team that made it), the gels in Portal 2 are based on the student project Tag: The Power of Paint, its developers hired by Valve. The gels do indeed resemble paint, splattering surfaces with patches of color that also imbue them with useful properties. The blue repulsion gel acts like a jump pad, catapulting Chell to otherwise unreachable heights, or letting her bounce back up after a fall. The orange (remember: blue and orange are the only colors that exist) propulsion gel increases Chell’s running speed, acting like a speed strip in arcade racing games. Players must use portals to coat (or sometimes re-coat) surfaces with the appropriate gels so Chell can make death-defying leaps between platforms to reach the exit.

The first time I played Portal 2, I got stuck on a few of these puzzles, making this section seem much longer than it actually is. This time it clicked with me more easily, and I was having a good time flying around the cavern. But I did still get stuck in one place again, bringing flashbacks of my original frustration: the introduction of conversion gel, which coats surfaces in white and allows them to host portals. This gel is introduced in a large, blank room, and players can spread the conversion gel all over, letting portals be placed anywhere. When puzzles up to this point had relied on carefully considering which surfaces could hold portals, the sudden freedom to put portals everywhere is overwhelming, and I found myself lost as to how to proceed. I managed to sort it out with less pain than I did the first time around, but it’s still a strange moment in the game, and one which is never truly repeated, even in later sections that feature conversion gel.

Overall, however, I enjoyed messing around with gels. I was less enamored with the narrative accompanying them. Deep in this forgotten vault, players are introduced to the founder of Aperture Science, Cave Johnson (yes, his name is Cave, while players are exploring a cave; the writers actually did that). He’s left pre-recorded messages to guide test subjects through Aperture’s original testing chambers, so players are able to glean the story behind the origins of Aperture Science. Which is basically: Cave Johnson is an asshole. His character is clearly inspired by American capitalists like Rockefeller or Ford, his pronouncements full of promises of progress and greatness, of conquering challenges and attaining glory. He is arrogant, displays a callous disregard for human safety by willfully sending people into dangerous tests, randomly fires employees in the middle of his recordings, and regularly ignores advice from his teams of scientists and engineers (who are clearly the ones doing all the actual work). Pretty much like Elon Musk, so kudos to the writers for nailing that years before popular opinion started to turn on that guy.

But here’s the thing: so much of what made the first Portal work was about GLaDOS being an AI, and therefore having a fundamental misunderstanding of humanity. At first the tests seem like they are aptitude tests for using the portal device, but later it’s revealed that they’re just some sort of sadistic gauntlet, “testing” with no actual purpose, an AI doing what it thinks is science but is actually meaningless. GLaDOS thinks that painting a heart on the side of a cube and calling it a “companion cube” will cause Chell to develop an emotional attachment to it and be unable to sacrifice it to proceed, an amusing misunderstanding of how anthropomorphizing objects actually works that, ironically, still worked on a lot of players. Even after GLaDOS’ murderous intent is revealed, GLaDOS still thinks that Chell can be motivated to finish the deadly test course with promises of cake. GLaDOS is insane (or rather, isn’t) in a way that only an AI can be. But now, in Portal 2, we find Cave Johnson acting in exactly the same way. What is he even testing for? The gels already work, as could be easily verified using inanimate objects. What “science” is actually happening in these test chambers? Why bother building this huge underground complex in the first place, and coercing people through its gauntlet? Sure, Cave is clueless and full of terrible ideas, but I was still hoping for some semi-coherent purpose to it all. How fascinating it would have been to see Aperture Science transform from an institution with noble, if misguided, intentions into the monstrosity it is today? Instead, all we learn is that Aperture Science was always crazy and dangerous, even when humans were still in charge.

The big reveal, as players move through this section, is the origin of GLaDOS. Reunited with GLaDOS in potato battery form, Chell forms a tenuous alliance with the AI since they now have a common enemy in Wheatley. As the pair climb upwards out of the vault, they move through different eras of Aperture Science, from the glory days of the 1960s to harder times in the ’70s, and ending in 1986. Here we learn that Cave has flaunted safety guidelines and given himself a lethal dose of moon dust. Terminally ill, he decides to start research into converting human minds into artificial intelligences that can live forever. If he doesn’t survive long enough, he wants his assistant, Caroline, to be digitized and run the place. There were several hints about this already, but now it’s clear: GLaDOS grew from the psyche of Caroline. This reveal undercuts so much about the first game, as GLaDOS can no longer be considered an AI whose thinking is alien to ours. GLaDOS is human (or at least, was human) after all. And Caroline as a character is barely established. She speaks up a few times during Cave’s recordings, but all we learn is that she’s very enthusiastic and loyal to Cave. She might be the one who’s really pulling the strings at Aperture, or she might just be a toady. Without more to go on, the reveal isn’t that impactful.

There’s one more thing about this section that I’d like to complain about. I understand the symbolism of Aperture’s forgotten past sealed away miles below the modern facility, but it doesn’t make any sense. Even then, the facility was underground. Surely the first test chambers would have been just below the surface, and then as Aperture needed to build more they would have dug deeper, such that the deepest constructions are the most recent? It would have taken a lot of foresight to start construction that deep, just so newer facilities could be added above later, and Cave doesn’t seem like the kind of guy who thinks that far ahead. Also, I liked playing around with gels as examples of early Aperture technology, but the puzzles can’t be solved without portals. Are we to believe that the portal device, which seemed a pinnacle of modern technology in the first game, was actually already in use as early as the 1960s, perhaps dating all the way back to Aperture’s founding? If so, Aperture’s accomplishments since then are disappointing.

Climbing out of the vault, players enter the final act of Portal 2, which is by far its strongest. Spotted by Wheatley again, Chell is once more forced to work through test chambers, as we learn that Wheatley has an “itch” that gives him the need to test. We also learn that Wheatley gets the equivalent of a dopamine hit whenever a test subject solves a puzzle, so at first he designs insultingly easy puzzles for Chell to solve. He even tries to send Chell through the same test twice, only to find out that this doesn’t work, each test must be new. So he tries to make new ones by smashing together a few existing chambers, before giving up and just using some of GLaDOS’ designs. All the while, Wheatley is finding diminishing rewards as he builds up resistance to the virtual dopamine, slowly becoming more hostile. Also, he is woefully unequipped to run the facility and systems are going haywire all over the place. Chell has to get GLaDOS reinstalled or the whole facility is going to blow.

This is the first time that Portal 2’s puzzles don’t feel entirely separate from its story. Watching Wheatley adapt to his new role of test administrator is like seeing an accelerated timelapse of GLaDOS’ own development. Wheatley starts out nice, generously offering easy solutions for Chell, but soon becomes increasingly frustrated and projects that anger onto Chell. Each test reveals more of what Wheatley is going through, and soon it’s clear that he’s no longer interested in having Chell run any tests, as he starts dropping blunt and obvious hints that he simply plans to kill Chell and GLaDOS. Players get a glimpse of GLaDOS’ side of the first game’s story, which doesn’t inspire sympathy but does yield some understanding. And GLaDOS gets to see what it’s like to be the one running the gauntlet.

It’s inevitable that Chell will manage to escape Wheatley’s deathtrap and eventually challenge him, but this time around it’s so much more satisfying than taking down GLaDOS earlier in the game. This time, players take the lead rather than simply following instructions. GLaDOS wants Chell to get to Wheatley’s chamber in order to be reinstalled, but how to get there is up to Chell, and the sequence is a decent rehash of the first game’s finale. I spent a lot of time sneaking around through walkways and passages in Portal 2, but none were as enjoyable as these, in no small part because I got to exploit malfunctioning gel pipelines to propel Chell across huge chasms. And once again, I enjoyed Wheatley’s approach to problems. He generally has the right idea, namely removing any portal-capable surfaces to leave Chell helpless, something that GLaDOS conspicuously failed to do in the first game. But his execution is inept as ever, so there’s always a loophole Chell can exploit to circumvent his defenses. The final confrontation with Wheatley is more rewarding than the duel with GLaDOS was in Portal, and the ending sequence of Portal 2 is absolutely brilliant, so much so that even in this spoiler-filled post I can’t bring myself to spoil it. Valve certainly nailed the finish.

But if there’s one thing that Portal 2’s final act makes clear, it’s that the rest of the game feels like a story that keeps interrupting itself with puzzles. Even as I realized that Portal 2 isn’t as long as I remembered, I was still wishing its earlier parts were shorter. I recently saw a comment on social media arguing that there’s no such thing as a game that’s too long, only a game that’s too boring. Portal 2’s early sections are too boring. Or, put another way, the pacing of its puzzles and its story don’t match. Working through test chamber after test chamber while suffering weak insults from GLaDOS is not compelling; after the strong, story-focused opening scenes, players want to see what will happen next, but instead the story is put on hold while they complete a bunch of simple puzzles.

The first Portal got around this problem by seeming to be nothing but a puzzle game at the start, with story only creeping in at later stages. Portal 2 wants to tell its tale right out of the gate, but it also wants to have players complete test chambers, and those two things don’t work together. Why bring in test chambers in the early acts at all? Sure, seeing their ruined forms right at the start was kind of cool, but once GLaDOS is reactivated, it would have been much more interesting to control Chell on the run as GLaDOS tries to kill her again. Maybe Wheatley could even “help” by messing up something so badly that it accidentally ruins one of GLaDOS’ attempts at murder, or maybe the sorry state of the facility inhibits GLaDOS and lets Chell escape. Imagine a trap sprung, only to get caught amongst the vegetation and fail to properly deploy. Players could still move through a set of puzzles, one step ahead of GLaDOS, who can only throw insults as the malfunctioning facility stymies any direct action against Chell. Then, just as GLaDOS is finally regaining full control, Wheatley and Chell strike.

That wouldn’t solve all of Portal 2’s problems, but it’s an example of the kind of narrative design that it needs. The finale pulls this off, but the opening and the sojourn into Aperture’s past could do with a rethink along these lines. And so, at the end of this lengthy post, I find my original opinion of Portal 2 is unchanged. Early pacing issues and some specific problems in the writing (particularly for GLaDOS and Aperture’s past) mean it falls short of the brilliance of the first game. But that’s a very high bar to set, and I fear that all of my complaining may leave readers thinking I dislike Portal 2. Not so! It’s still a great game, one I highly recommend. But it has some flaws that keep it from perfection, whereas the original Portal is very close to perfect. I don’t envy the developers for having to follow that masterpiece, and what they made is a valiant attempt, but going for something bigger turned out to be a double-edged sword. The high points of Portal 2 are well worth it, but I can’t help but wish that the rest of the game were elevated to the same level.