Intermittent Mechanism

Ian here—

A quick break from the regular flow of student blog posts to announce that the fifth episode of the Shape Up! series is posted. It is on Save the Date and Elsinore, both games that I have taught before but never written about myself on the blog before, despite them being favorites of mine. More videos in this series are forthcoming, although it will be awhile before I can return to the one-a-month schedule.

Full script below the jump.

Hello everyone, and welcome to another episode of Shape Up!—an ongoing series about form, structure, and pacing in games. This video is a continuation of a discussion about “off-ramps,” a structural aspect of games I discussed in my previous video on P.T. If you didn’t watch that video, don’t worry—I’ll be introducing it from scratch here.

Contemporary single-player games, across a wide variety of genres, tend to be stuffed with features and content: collectables, optional challenges, sidequests, unlockable difficulty levels, and these days more and more live service elements. And of course, not every player is going to be interested in all of these things. So games typically offer cascading tiers of closure. Most modern single-player campaigns will have a basic ending, where the credits role. Putting the controller down at this point is different than putting the controller down after you’ve died in the first really tough boss battle. It’s not socially coded as “giving up.” You’ve reached an ending that, while it might not be the only ending, is an ending that is officially sanctioned by the developers as being something other than a fail state. You get a little pat on the head, and you get some sense of closure. “You’re done here,” the game announces, “if you want to be done.” This is the game’s first major off-ramp.

But some players are compulsive completionists, who won’t consider the game “done” until they’ve played every sidequest. Or seen every ending. Or acquired every car. Beat every optional boss. Gotten an S+ rank on every level. Collected every pigeon—whatever. Different players have different play styles, and different ideas about what it means to “finish” a game. And developers realize this. Offering multiple off-ramps—one off-ramp when you see the credits the first time, another off-ramp when you’ve seen every possible ending, or when you’re romanced every datable character, or when you’ve unlocked every cosmetic item—this helps everyone customize their gaming experience around their own desires, and then get their own unique, tailor-made moment of closure.

To facilitate this sense of closure, the predominant philosophy in single-player game design these days trends toward transparency of content. A lot of games will tell you exactly how far along you are to blacking them out completely, sometimes down to the 100th of a percentage point. 97.43 percent of the time, I’m not going to do everything, because I don’t find the content interesting or rewarding enough to engage with. But I understand the function of these numbers. It’s to give players who are hopelessly fixated on scratching that completionist itch a way to gauge their overall progress. It’s one means of answering that question I mentioned in the introduction of this series—that recurring question, “how much longer is this game?”

There are entire genres of game that I don’t really think would work without this sort of transparency. Years ago, I did a video on Nine Hours, Nine Persons, Nine Doors, where I pointed out that the main menu screen that tracks how many endings you’ve seen serves a vital function in that game, by encouraging players to keep trying until they find a good ending. The narrative twists and turns in that game just wouldn’t work if the game didn’t give players some sort of motivation to keep trying despite failure, and the endings grid screen fulfills that function. A lot of visual novels and interactive-drama-type-games are like this. These games typically don’t offer much in terms of traditional gameplay challenges—instead, they’re all about narrative exploration. Giving players a node map gives them something to hold on to, some sense of progress and satisfaction. (Even if it’s only the meager satisfaction of checking things off a checklist.)

So: transparency makes sense in a lot of cases. It makes sense in 999, and I wouldn’t want to live in a world where that game and the subsequent Zero Escape games didn’t exist. And the first video in this series, I praised the menu of the game Vignettes precisely because it provides clarity of scope, warding off the boredom that can come if exploration is too free-form and unstructured. That said, every aesthetic technique has its place, and there are also games that are compelling to play, and keep playing, precisely because they aren’t transparent about how much content they have—that toy with and problematize their players’ completionist impulses.

The Stanley Parable, for instance, engages in a kind of asymmetrical warfare: it tracks your movements carefully, so as to provide unique dialogue for many different behaviors. And yet it steadfastly refuses to let you track it. It doesn’t provide an in-game flowchart of endings. It doesn’t give you a percentage completion screen. Even its achievements don’t give you a good sense of what you have and haven’t seen yet—instead, they’re all weird and jokey. The Stanley Parable can’t outright tell its players the numerator and denominator of the endings they’ve unlocked, because as the loading screen says, “the end is never the end, is never the end, is never the end.” Playing The Stanley Parable is engaging and fun precisely because it doesn’t feel like you’re just going down a checklist. There’s still a little excitement there—even playing all these years later, when the internet is full of fan-made flow charts. There’s still the feeling that you could boot it up and discover a new surprise, purely by accident.

Over the course of the rest of this video, I’m going to discuss two games that are unusually dedicated to interrogating their players’ desire for closure. These games understand the expectations about off-ramps that players have formed playing other narrative-based games and visual novels, and they deliberately subvert them. They’re also both about time loops, which I don’t think is incidental.

Fair warning that because this video is an analysis of how these games structure the search for an ending and players’ ultimate decisions of when to stop playing, I will of course be massively spoiling them. There’s something to be said for the experience of negotiating your own sense of closure as you play both Save the Date and Elsinore, and that experience will be severely compromised if you watch this video first. So consider yourself warned. (And also, Save the Date is a free download for Windows, Mac & Linux, and only takes about an hour to play, so consider that before you spoil it.)

OK? You ready to continue?

Save the Date is a 2013 visual novel created by Chris Cornell, operating under the moniker Paper Dino Software. It starts with your player-character getting a call from Felicia, a woman who you’ve agreed to go on a date with. Your first choice in the game is where to go: out for burgers, Thai food, tacos, or cancel the date altogether. Choosing this last option leads to the absurdly quick arrival of a bad ending.

OK, then! Back to the dinner choice. If you go out for burgers, after a few bites of your dinner and some conversation about family, Felicia gets shot during a breakout of gang violence, and dies. You get another bad ending.

Hmm. This game seems to be filled with abrupt and absurd bad endings. The Thai route and the taco route don’t end well, either: if you go to the Thai restaurant, Felicia will die from an allergic reaction to peanuts, and if you go to the taco place she ends up on a deck that collapses and falls into the sea.

And that was all four of your starting options—so something’s got to be going on here. There has to be some way of using the knowledge we now have to get a less absurd ending, right? There has to be some way to … save … our date … oh. Get it?

From here on in, Save the Date turns into a game of applying knowledge gained from previous failed runs. A Groundhog Day scenario, essentially—and the game even name-drops that film as a reference point. And while we’re on the topic of Groundhog Day, I want to make a point about its narrative structure. Yes, Groundhog Day is about a time loop. Yes, it has an unusual screenplay in many respects, most notably for its high degree of redundancy. But despite all of this, its structure isn’t radically avant-garde, and in fact once you examine it, it really conforms to the expected shape of Hollywood’s three-act structure.

Now, a quick aside here: it’s confusing, but even though screenwriting gurus love to refer to the supposed “three-act structure” of Hollywood films, the most popular manifestations of this structure actually have four parts, not three. Pretty much everyone agrees that “Act II” is twice as long as acts I & III, and most manuals also say that there should be some sort of action at the midpoint that presents either a boon or a setback to the protagonist. It’s more because of momentum than anything else that people call this a “three-act structure” instead of a “four-act structure.” Kristin Thompson, whose analysis of Groundhog Day I’m drawing from here, does propose that four acts is the right way to think about things.[i] I’m sympathetic to her point, but it’s not worth turning this into a whole thing, and the only reason I addressed this at all is to make sure the next bit make sense.

Anyway, according to Thompson, the first large-scale portion of Groundhog Day consists of the scenes that take place pre-Groundhog Day, then the first Groundhog Day, and the first two repetitions where Phil first realizes his predicament and then fails to get a medical explanation for it. This is the “setup,” and it ends when Phil realizes that his previous goal of moving on to the next phase of his career at a major network has been rendered completely impossible. Next is the “complicating action,” where Phil makes a new goal for himself: he realizes a life without consequences can be a life of hedonism. He seduces Nancy by manipulating her across loops, and tries to do the same with Rita, but never can quite get it to succeed. Increasingly desperate, he kills Punxsutawney Phil, and himself, multiple times. Third comes the “development.” Phil changes his goal again, and now tries to better himself, live a perfect day, and become Rita’s ideal man, rather than simply tricking her. Finally we have the climax/dénoument, the final perfect Groundhog Day and Phil’s escape from the time loop. There you go: many loops, but four large-scale parts, and a nice clean character arc, signified by Phil’s shifting goals.

And we can perform a similar large-scale segmentation of Save the Date. Sure, everyone’s playthrough is going to differ in exact order of things tried, and re-tried. But as a whole, the game is structured around a series of tactics that present themselves before exhausting their utility.

So, first players will try all three restaurants—and we’ve already seen how that goes. From here, there are two alternate tactics you can discover as possible ways forward. One is to warn Felicia about her death during a restaurant scenario, via new dialogue that opens up once you’ve seen her die there once. This new dialogue option appears in each one of the scenarios, and if you select it, you postpone her death a little longer. But she’ll still die in another way moments later, in all three restaurants. So, that tactic’s a dead end.

Then there’s the other tactic: Once you’ve tried all three restaurants at least once, the very first decision point changes, and now you’ll have the option to warn Felicia that going out to dinner is too dangerous. She doesn’t believe you, but the game offers three different dialogue paths of what you could possibly say to convince her, before she abruptly dies again by a plane crashing on top of her. By exploiting information you gain in other playthroughs, you can do each of the thee things that might convince her—but none of them actually work. In each instance, she’s weirded out by your behavior and hangs up. Another dead end.

Both of these tactics are dead-ends, but they are necessary dead-ends, because you can combine them into a new tactic. If you use the moments after you’ve saved Felicia’s life in one of the three restaurant scenarios to tell her about the trouble you’ve been having convincing her over the phone, she’ll be more receptive to your claims of supernatural abilities. If you continue this same line of conversation across at least two life-savings, she will ultimately suggest a new tactic that involves leveraging her childhood fantasy of a local Hogwarts pickup stop.

Finally, you’re able to meet Felicia on a hilltop outside of town, for the game’s extended climax. Here, the two of you have a conversation about death, and videogames, and her deaths in this videogame, and for awhile it seems like this is it: you’ve found the good ending, where the game goes full meta. Except Felicia doesn’t stop dying: she can die by a meteor strike, and then if you successfully prevent that, she dies during an alien invasion, which seems to represent the outer limit of her fate, one you can’t possibly change. So with no way to save her, all you can do is keep talking to her, be frank with her about your fruitless search for a good ending. The resulting conversation you have, across multiple dialogue branches that have to necessarily be spread across multiple playthroughs, gets increasingly philosophical. Felicia asks you why you think a “good ending” exists at all. You can point out that there’s an implicit contract when you play a game that it will have some solution that provides both mechanical and narrative closure. But she challenges you with the idea that maybe this game doesn’t have a “good” or “true” ending—and maybe the best thing for you to do is just cut your losses and stop playing. After all, your continued playing means her continued dying. If you want a happy ending, maybe the onus is on you to quit the game, and write that ending yourself.

Now, maybe that seems stupid, and not at all like what a game or a story should be. But by this point you’ve seen every branch, mapped every node, exhausted every option. So there’s nothing left to do but dwell in your own dissatisfaction, and think about exactly why it is that you have such a strong desire not just for closure, but closure that is sanctioned by the author, in a way that makes it more official, or more real, or more (literally) authoritative. Why is it that unlocking a particular “good” or “true” gives us a sense of accomplishment? Wouldn’t we get a deeper, more gratifying sense of accomplishment out of undertaking a creative task, like writing our own story? On the one hand, I want to praise Save the Date for being rich and humanistic and prodding its players to consider the value of creative work—including the value of interpretation as creative work. But on the other hand, playing it is also unsettling, because it reveals just how accustomed I have become to the cheap satisfaction of closure: that little dopamine rush that accompanies the final hit on the final, hidden boss’ HP bar, and the chime of the achievement or trophy that acknowledges my accomplishment. The percentage counter ticks up one last time, and meanwhile I’m one tick closer to the grave.

God, that’s depressing, sorry. Let’s change the subject to something lighter, like … Shakespearian tragedy.

Elsinore was released in 2019 by Golden Glitch Studios. It is a riff on Hamlet, in which you play as the character Ophelia as she experiences a four-day time loop, beginning on a Thursday morning positioned between scenes I and II of Hamlet’s first act, and ending on a Sunday evening during which the events of the climax of act five take place. The game shares some immediate similarities with Save the Date—both are about leveraging information learned during previous failed runs, and both concern a seemingly inexorable slide toward death and tragedy—but unlike Save the Date, Elsinore is no bare-bones visual novel. It’s a full-fledged adventure game, one that harkens back to some of the greats of the genre. (In particular, its emphasis on following characters around, learning their schedules, and spying on them recalls The Last Express.) But it’s also not content to sit on its laurels. It boasts an ambitious and fascinating “story engine” system, which replaces the item inventory of traditional adventure games with a storehouse of information and rumors, and swaps traditional locked doors with characters whose actions can be altered by changing their beliefs about the world.

Much like I did with Save the Date, I want to start off my discussion of Elsinore by revealing the large-scale segmentation that lurks behind its many, many, many loops. Your first run through Elsinore is going to track exactly onto the plot of Hamlet, and the first two characters to die in Hamlet are Polonius (your father) and Ophelia (your character). In Elsinore Ophelia’s death is retconned not as suicide, but as a murder by a conspirator in league with Fortinbras attempting to sew chaos among the court, that is subsequently disguised as a suicide. Once Ophelia dies, she’s sent back to Thursday morning, and begins contemplating her predicament. Your first, most proximal goals are going to be preventing your own death, and preventing your father’s death. Preventing Ophelia’s death is a matter of espionage: spying on everyone to discover who the assassin is, so that you can then have them locked up on a subsequent run. Preventing Polonius’s death is a matter of playing with the game’s social systems, finding out which bit of information dropped to whom at what time will prevent him and Hamlet from being in Gertrude’s room at the same time. There are several ways to accomplish this, with different downstream effects.

Once you’ve saved yourself and Polonius, you’re able to witness to the rest of the events of the play, all the way to Sunday evening. And you can also start rejiggering aspects of it to achieve different outcomes. Get Claudius to confess to his brother’s murder, and Hamlet will stab him in the back at his earliest convenience, without wasting time skulking around and feigning madness. Then you can become queen and rule by his side—for a few moments, at least, until Fortinbras invades. Pretty quickly during experimentation it becomes clear that Fortinbras’ invasion and coup puts a hard limit on your narrative horizons, so maybe your next goal should be snipping it in the bud somehow. This is easier said than done however, because even if captured, Fortinbras will escape and find a way to lay siege to Elsinore if even a single character dies. So in order to stop Fortinbras’ coup, you’ll need to keep everyone alive for the entire four days of the time loop. Which is hard, because, well … it’s Hamlet.

Actually keeping everyone alive is a goal that requires a lot of narrative experimentation, seeing radically different routes play out to their conclusion as a way of gaining information. And as you do this, you’re bound to witness some other events that suggest that a supernatural solution might be the more promising way out of your predicament. You’re introduced to King Hamlet’s ghost, who tells you to meet him again on the first night—revealing that he, too, is experiencing the time loop. The ghost requests that you find a diary to help restore his memory, and reading that diary (along with other documents) reveals that others in the royal court—including not only the king, but also the former queen Astrid’s servant Simona have previously experienced time loops. All of this suggests that you should track down a book from Simona, alternately called the Book of Fates or the Hand of Dionysus. Once it’s in your possession, the character Quince—the player who puts on the play that’s the thing that catches the conscience of the king—reveals that he’s an extra-dimensional being, who revels in the desperate machinations of tragic characters as his entertainment.

But you have the book he wants, and from here you can team up with the ghost and Hamlet and use the Act III play-within-a-play to reveal Quince’s treachery before the whole court, and then kill him. Which you do. You kill him. And for a moment, it seems to break the time loop, and create a version of events that has a happy ending. Except it’s a trick: Quince is not dead, he’s only messing with you. He permanently erases the ghost from existence, and reasserts his control.

Earlier I said that although Groundhog Day takes the form of a bunch of repetitious loops, it still conforms to the expected four-part structure of a typical Hollywood film. Maybe I’m imagining it—and you should certainly reserve the right to be skeptical—but Elsinore seems to me to a five act structure, which, if real, is pretty cool. Furthermore, unlike Save the Date, in which the large-scale parts are mostly delineated in terms of shifts in tactics toward the same basic goal, the segments of Elsinore are marked by genuine shifts in goals, making them feel more act-like. The way I see it is this: Act I is the game’s prologue and first run of the loop, where Ophelia’s only goal is to go about her life. Act II begins with the second run of the loop, and it introduces the player’s first two intertwined goals of saving Ophelia and saving Polonius. Act III is kicked off by the revelation that Fortinbras’ coup will always present a hard limit, so the avoidance of that seems like it will be the next important goal, although over the course of pursuing it players are likely to discover the ghost and the book. Act IV kicks off once you have the book in your possession, and get confronted by Quince, prompting Ophelia to form her alliance with the ghost and Hamlet, and pursue the new goal of defeating Quince once and for all. Act V begins with the revelation that there is no defeating Quince, and it continues for as long it takes for the player to choose an ending.

It is this final act where Elsinore gets really interesting to me, especially in comparison to Save the Date.

All of the endings in Save the Date are straightforwardly bad. Felicia dies, and your date has ended in disaster. At the top of this video, I defined a game’s off-ramp as as an ending that doesn’t feel like a fail-state—and these aren’t that. These endings do feel like fail-states. Save the Date interrogates the player’s desire for closure by straight-up denying an off-ramp, and then asking us point blank why we find that so annoying. It’s effective, but it’s rather blunt. And along the way, Save the Date breaks the fourth wall in a deflationary way. By the time we’re conversing with Felicia on the hilltop about the nature of narrative closure, she’s long since accepted that she’s a fictional character. And because of that, she’s sort of stopped being a character, and is now acting solely as the author’s mouthpiece.

Elsinore avoids this type of deflation. It keeps Ophelia front and center, and she gets to keep being a character. She doesn’t acknowledge that she’s in a videogame, the way Felicia does—though she does come to understand that she is trapped within the constraints of tragedy, and is being denied a fully happy ending by generic forces imposed on her without her consent.

The endings you ultimately have to choose from in Elsinore, after your plan to kill Quince fails, aren’t absurdly bad in the way that Save the Date’s are. Ophelia survives in most of them. And players might decide they like some of them! The “Sacrifice Predictability for Adventure” ending, where Ophelia becomes partner of and lover to a lady pirate, is pretty fun! And on the total flip side, the “Sacrifice Independence for Peace” ending, which sees Claudius punished for his crimes, everyone else surviving, and Denmark peacefully submitting to Norway, doesn’t feel particularly heroic but it is a pragmatic outcome for the Danish citizens, in a realpolitik sort of way. These endings aren’t totally happy, but they feel like acceptable compromises—so why not just pick one of them?

Well, there’s a very devious exchange between Ophelia and Quince. By the time you get the Book of Fates, and can actually sit around and peruse all of the possible endings you’ve unlocked, Quince has been revealed as the game’s final antagonist. And his goal is for Ophelia to choose an ending. Upon realizing this, Ophelia vows never to make a decision. And especially after the game has raised up your hopes that there might be a way to defeat Quince once and for all, only to cruelly dash those hopes, it feels like a betrayal of Ophelia’s spirit of defiance to pick an ending and close out the game.

This, to me, is a richer interrogation of closure than simply denying an off-ramp, as Save the Date does. Elsinore has off-ramps aplenty. But the game portrays the act of choosing an ending as an admittance of defeat by our player-character, a capitulation to the game’s antagonist. Rather than simply denying our desire for narrative closure, as Save the Date does, Elsinore gives us the power to indulge in it, but says that if we do, we will be complicit with the game’s antagonist. To take any of the game’s possible off-ramps would be an act of weakness.

I said early on in this video that I didn’t think it was incidental that both Save the Date and Elsinore are about time loops—and to sew up this point, I want to return to Groundhog Day one last time. The philosopher Stanley Cavell was a big fan of Groundhog Day, and he read the film as the film as the confluence of two different philosophical lineages.[ii] On the one hand, you have Nietzsche, and his idea of eternal return. But on the other hand, you have a strain of moral perfectionism, which Cavell attributes to the writings of Ralph Waldo Emerson. Phil Conners has to live the same day over an over again, but he can change his behavior as he re-lives that day. He can strive to better himself, to cultivate a new version of himself, to get closer to the perfect version of himself, one step at a time.

There is an oft-remarked-upon similarity between Groundhog Day and the experience of playing video games. When we play games, we too are trying and trying and trying again, and improving our run with every subsequent go. Games share this kind of perfectionist tendency with Groundhog Day—and this is especially true, I think, of games that are explicitly about time travel, or about time loops, or games in which the central mechanic is tweaking character decision points to see what happens downstream. The reason Save the Date and Elsinore dug their hooks into me so deeply the first time I played them was that pretty much every other game I had ever played that was about moving back and forward in time and manipulating events did have a “true” ending. These other games rewarded players for putting in the work—for exploring the narrative landscape to its fullest potential, seeing how every decision played out, exhausting every dialogue tree, tediously tracking down every dead end—by withholding their most revelatory and cathartic ending from everyone but those players that did the labor. These games had a final off-ramp, where you know you’re finally done because the game comes out and tells you that you’re special.

There’s a moment in Save the Date where Felicia analogizes the game that she’s in to Groundhog Day. She points out that, much like Bill Murray’s character undergoes a shift in Groundhog Day, from attempting to memorize his way into Rita’s pants to genuinely trying to improve himself, that maybe you also need to shift your attitude—that is, stop being the type of person who obsessively searches for a “good” ending, and put the game down. This is a clever moment. But even though it disrupts the player’s expectations, it still espouses that perfectionist self-improvement framework. Elsinore, I think, goes further—makes a cleaner break from the Groundhog Day tradition. It sloughs off that Emersonian influence, and goes straight to Nietzsche’s idea of eternal return, which is not about self-improvement in the slightest. In Nietzsche’s conception of the idea, there is no change over subsequent cycles, no development, no putting in the work to make everything perfect. There is only the awareness that this has happened before, and it will happen again, that nothing is ever new. Nietzsche’s point is that it would take enormous moral courage to embrace this awareness without shattering. And that is just what Ophelia is prepared to do in Elsinore. She would rather live in an eternally suspended state, repeating the same four days forever, than consent to a tragic ending for herself and her friends. It’s a much rarer conclusion to a time-loop narrative in any medium, and I respect Elsinore enormously for it.

That’s it for this episode of Shape Up. Stay tuned for the next episode, where I’m going to look at another cluster of games that do interesting—though somewhat different—things with repetition and off-ramps.

Alright … are the casual viewers gone? Did they leave during the credits? Can we actually talk about these games’ hidden true endings now?

So if you open up the directory of Save the Date, you can hack the game. Just like any good hacker, just look for the plain text file named “I_AM_A_HACKER.rpy,” and set the “hacker” value from “False” to “True.” If you do this, you’ll see the true happy ending of Save the Date, where you invite Felicia to an awesome dinner in your floating sky castle. And … hmm. This is definitely a happy ending, but … I dunno. Feels kinda hollow. Like the game is mocking me.

OK, so the hidden true ending for Elsinore. If you pick one of the eleven possible endings, and watch the end credits, but then re-open the game, and say you want to pick another ending, the ending you previously picked will be destroyed. Do this four times with four different endings, and finally Quince will approach you and begin to treat you as his equal. He gives you the “exeunt all” dialogue to force characters to leave the stage—that is, you can now kill all of your friends and family. Quince challenges you to kill everyone in Elsinore by Friday evening, and despite your newfound godlike powers it’s still a surprisingly tricky task—certain characters will rat you out or else kill you immediately if they see you using it, so you have to take advantage of moments when your prey is isolated. From here on in, the game becomes a very different sort of puzzle: whereas before you were learning people’s routines to spy on them and gain information, now you’re finding windows where you can murder without witnesses, even as the security around Elsinore grows tighter and tighter with each passing death. Ophelia changes, as well. In the early going, she’s horrified by her power, insists that it’s evil and that she shouldn’t use it. But the more you crack the puzzle, the more insistent she gets about finding a way to “win” the scenario. And by a certain point she’s cackling as she drains the life from people, having recognized them as nothing but puppets or playthings. If you successfully get Ophelia to kill everyone, she will join Quince in his inter-dimensional travel, and the two of you will wreak tragedy on thousands more. Only now is the game well and truly over.

OK—that ending is definitely not happy. And it also kind of feels like it’s mocking me for achieving it, right? Like, what sort of person would be so obsessed with “winning” that they’d pursue such an obviously demonic path? It’s like Ophelia has become obsessed or deranged like a … Shakespearean tragic hero. And her fatal flaw is that she doesn’t know when to end a game and leave things well enough alone. A flaw that she inherited from us.

That’s right, both the “hacker” hidden ending in Save the Date and the “Exeunt All” hidden ending in Elsinore are thematically consistent with the rest of their respective games. They don’t tell you you’re special, they don’t reward you for being clever, they’re still interrogating your motives, asking you why you thought it would be a good idea to do the thing you just did.

Elsinore also has one more additional hidden ending, the “Sacrifice Choice for Revenge” ending, where you burn the Book of Fates. This traps Ophelia in a hell in which she burns to death every four days for all eternity, but it successfully spites Quince in his quest for the book. I like this ending—it’s the one ending where Ophelia says true to her defiant spirit. It’s the one I chose the first time I played the game, and had to decide what my own personal off-ramp was going to be. But in hindsight, coming back to it a few years later … I wish the game hadn’t included it. It’s a little too have-you-cake-and-eat-it-too: you get to respect Ophelia’s wishes not to choose an ending, but then you still get to see an ending, get that sense of closure. The sense of satisfaction players get from seeing the credits roll feels anathema to Ophelia’s own predicament. In fact I kind wish that the game had

[i] Kristin Thompson, Storytelling in the New Hollywood: Understanding Classical Narrative Technique (Cambridge, MA: Harvard University Press, 1999). Thompson’s large-scale arguments about Hollywood cinema actually following a four-act structure are laid out in chapter one, “Modern Classicism.” The act breakdown of Groundhog Day that I am outlining here is from chapter five.

[ii] Stanley Cavell, “The Good of Film,” in Cavell on Film. Edited by William Rothman. (Albany, NY: SUNY Press, 2005). Cavell’s commentary on Groundhog Day in this essay is quite brief—it’s limited entirety to pg 345—and I’m synthesizing some of the ideas from the essay with an email Cavell wrote on the film, reproduced in the same volume.