Ueda Coulda Shoulda: The Quest as the Shadow

  • Reading time:23 mins read

by [name redacted]

A somewhat edited version of this piece was published by Game Career Guide, under the title “Rock in His Pocket: Reading Shadow of the Colossus“; here is the article as originally intended. This version is also available, largely intact, in The Gamer’s Quarter.

Going by his two big brain dumps – Ico and Shadow of the Colossus – Fumito Ueda is a complicated guy to put in charge of a videogame: an ivory tower idealist, with only a passive understanding of practical architecture. As a dreamer, his ideas are too organic, too personal to fit the cliches that most of us take as the building blocks of game design. Knowing that, he sidesteps convention whenever it gets in his way – which is often. The problem is in those conventions which, though they mean nothing to Ueda as narrator, just as frequently get in the player’s way. The result is meaningful, painfully gorgeous games that are a pain to actually play – making Ueda’s statements, in all their profundity, accessible only to the most devoted.

Paired with a more down-to-earth design team to translate his ideas – someone with a Valve mentality – Ueda could change the world. So far he’s been the master of the golden arrow: the idea so poignant yet so tediously executed that it creates a certain cognitive dissonance in the player, inspiring not so much awe as transcendence – a deep need to puzzle out what went wrong and how to better it. As far as human experiences go, that’s a pretty healthy place to be. So I guess Ueda still succeeds on a level, if not quite the level he’s shooting for.

Furthermore he is learning, even as his ambition balloons. For all its faults – and they are mostly superficial – Colossus is far more well implemented than Ico. Many of the problems seem like they were introduced at the last minute, more out of insecurity than inattention. The three things that we, as an audience, can in turn take from the game are the theory behind it, the problems Ueda still hit in applying that theory, and where those problems came from in the first place.

The Theory

The game’s premise is right there in its title – more so the Japanese than the English, though the English title tells its own story. Wander and the Colossus. There is the central conflict: between the wandering, and the Colossi. The game presents a huge, gorgeous, lonely world to explore, completely at the player’s discretion – punctuated by a handful of intense, scripted, violent boss encounters.

The game’s story is one of greed, sadness, obsession, and – more than anything – ambition: there’s a dead girlfriend, and the only way to revive her is to kill. Not just kill; kill huge and dangerous, beautiful and unique mythical creatures. For the sake of your love, you must go out of your way to destroy all. And with every life you take, you yourself die a little inside; become a little less human.

Violence in this game is horrible, nauseating, yet thrilling. It is not so much an act as a release. You do not so much plunge your sword into the beast’s skull as you allow it to sink, as you give yourself over to the inevitable. Then the gore shoots out at you, and you panic, stabbing again and again, grasping to the beast’s fur as it screams and thrashes in pain. And it dies, and you pant, and you just had the most exciting videogame moment in years. Then you try to run from the black threads that stream out from the carcass, but there is no escape from your actions. Sooner or later the darkness penetrates you. And you wake up haunted by yet another shadow.

The genius thing is, the game offers a choice. The player need not be violent. There is a whole world out there, full of nooks and crannies and grandeur. The game offers a loyal (and actually rather bright) companion in a horse, which the player may stroke at will. Granted, the player has to find his own meaning in that world; it offers no easy answers, and no promise of closure. It does not exist for the player’s benefit. It simply is what it is.

The Experience

What the world of course will be, to most players, is boring. In a videogame, meaning and purpose generally come from action; from fulfilling tasks, and progressing toward a preordained goal. Finding the best place to sit and look at the skyline is not the experience that most people expect out of a PlayStation; they want mindless catharsis. They want to know they have a place carved out just for them, and a definite series of tasks to fulfill, in order to win. That is meaning, right there. It’s religion, dating, education, employment; everything that we use to reassure ourselves of our own value – to judge that we are making something of our lives.

So the choice the game offers the player is not much of a choice at all: either be content with nothing but beauty and your own liberty, or do your work and get your reward – however horrible it might be. At least you won’t be bored, and then when the game is over you will have accomplished something.

If the outcome is inevitable, then what is the point of this conflict? Mostly, to enhance the resonance of the player’s choices – as whatever the player does, whatever violence he eventually undertakes, is of his own apparent volition. (And how much destruction has been begat of boredom, I ask of you.) Consequence and morality hold more weight when liberty, rather than simple fate, is at play.

The Method

Nearly every design element is geared either to convey a sense of scope or a sense of intimacy. The gameworld is supremely node-based (going by Brian Upton’s vocabulary), with an endless horizon and landmarks out the wazoo, enabling navigation. There are a few paths, primarily to pace the approach to a Colossus battle, and many districts – which again, in combination with the horizon and landmarks, aid the player with orientation. What edges exist mostly serve to encapsulate the node-districts – the tower area, the forest, the mountainous region – and give each its own sense of place.

The central tower is visible at nearly all times, from nearly any point in the gameworld. Save points are clearly visible in the distance, and can indeed be scaled (as they take the form of small watchtowers) for a view of the surrounding area, and maybe an glimpse of the next watchtower. Whenever a Colossus is defeated, it leaves a streak of light in the sky – a beacon to again help orient the player, wherever he might be.

The focus on height and distance provide a continual sense of perspective, of context for the player in relation to the gameworld – contributing to a sense of belonging or ownership. Being able to climb around on landmarks helps the player to claim them as personal discoveries rather than as sterile facts of the gameworld.

Although there is little purpose to most locations, and so no explicit reason to explore, the design focus on nodes puts a strong emphasis on downtime – on exploration. That said, the dominant landmarks always remind the player of his pending Quest, and always provide a tether back to the “game” – a goal, as such – the moment the player grows bored and wants to progress. The downtime paces out the action enough to heighten the drama and immediacy of the Colossus battles; the Colossus battles are overwhelming enough to suggest a recovery period.

The Purpose

More than that, the majesty of the world and the lack of a driving compulsion to progress lead the player to question the mission. When the player seeks out and encounters the Colossi, they are mostly minding their own business. Generally it is the player who must pick the fight – and go out of his way to do so. There is no rush. The dead girl is hardly going to grow more dead. The Colossi aren’t going anywhere. This mission is no one’s but the player’s own. After a certain point, the game starts to feel a bit like jousting at windmills.

Were the game instead a boss rush, with a linear route or sequence of events leading to each encounter, the player would feel a driving sense of purpose. Of course he is meant to fight the Colossi, as doing so, or preparing to do so, is the entire fabric of the world presented. With liberty, however, comes uncertainty. Clearly the player is meant to progress, if he should care about finishing the game. Yet does he care? Must he care? And if he does not, then why is he playing? What other meaning can the game hold?

Colossus is almost post-Campbellian, in that the Hero Quest – the framework that we normally think of as the meat of a videogame – is actively questioned. The player can subvert the Quest – turn down the call to arms – at any time. It’s all a matter of will. Yet the game does not abandon the Quest structure so much as approach it at a higher level. There is only so long the average player can live with “BUT THOU MUST!” before caving in. Unless he altogether stops playing, the player will, ultimately, surrender and accept his fate – much as the sword plunge is not so much an action as a release. For all the introspection it inspires, ultimately Shadow of the Colossus is about giving up. About doing anything to escape emptiness. Even being a false hero.

The Details

Without going too overboard, it’s worth noting that what makes the game work, as far as it does, is its wealth of all of the subtle details that so many other games lack, which humanize the world and its population – that make it all feel genuine enough to be worth taking to heart.

One of the biggest details that most productions (especially American ones – think of all those waggle-legged FPS characters) skimp on is animation – and here it forms the soul of the game. The player character – dubbed “Wanderer” by fans – is a gangly, awkward young man. He is uncomfortable wielding a sword. When he runs, he swings his arms and legs with abandon, occasionally tripping over his own feet. He has no idea how to jump, and nearly falls on his face each time he tries. The only place he is greater than inept is on his horse, with his bow in hand – although he can cling to things pretty darned well. He has a whole button devoted to clinging. See if you can spot the literature here, in light of the game’s themes.

The game’s main button, though, serves to interact with the horse. Call to it, spur it on. Cling to the horse, bond with it, and you can perform tricks like standing upright at a gallop. When you’re off the horse, you can pet it at will – though it gets freaked out by weapons. It will follow you to the grave, whatever weird situation you get yourself into. When you’re on the horse, you don’t have to touch a control; you barely have to steer. The horse is smart enough to keep you on-course, even down a winding path, without running into a wall or slowing down more than necessary. Again, stick this in your analysis machine.

Curiously, the player is given a different button to “use” Wanderer’s sword than to attack with it. Using the sword means holding it up the the sky. If there is sunlight, it will reflect off the blade and form a beacon, pointing the way toward the next Colossus. Think about what swords symbolize here, and how poorly Wanderer wields it as a weapon. The sword, which is clearly not a part of Wander’s normal life, as the primary tool to direct the player toward the next leg of the Quest – given all we’ve seen so far, I think it’s fair to say this is deliberate. At least as deliberate as the hold-and-release stabbing action. Or inaction, as it were. Reflex.

The Cans

As in Resident Evil 4 or that scene outside Minas Tirith in Return of the King, when the player rides his horse – which he will be doing at least half the time – the camera lists off to the side. Or rather, it points straight ahead as it frames Wanderer in a low corner. Slowly, slowly, game designers are learning the value of subjective yet functional camerawork. With them off to the left, you can see them ride – horse and boy alike – yet your attention is on the horizon, on what you’re riding toward. This is in place of staring at a horse’s rump for the half-dozens of hours you’ll be riding around, as you would with a more traditional tracking camera. Which, I ask, is a more significant part of the composition?

On a similar note, pressing the map button will cause the camera to swing up and rapidly zoom out, until the player’s location fades into a pinpoint on the (cartographic) world map. This serves a couple of purposes. It contextualizes the menu screen by tying it into the game world as an abstraction. At the same time, it helps to illustrate exactly where the player is in the gameworld. The eye, following Wanderer, is drawn to the correct point on the map – and from there can scout out the surrounding territory. Pressing the map button then becomes, in a sense, like climbing to the highest high ground possible. It is thematically tied to the game’s natural state of navigation, and it is stylistically tied as just another subjective viewpoint among many.

The game is so concerned with camerawork that it provides three whole camera buttons, each with its own subjective view, each with its own function. The idea, I assume, is to cut down on as much player confusion as possible during Colossus battles – as the player will be staring up at the beasts, running around them, dangling off of them, shaken around like a hankerchief on a fond departure. Perhaps they tried a few contextual camera systems, and found them all inadequate. This solution feels a bit unwieldy, I must say – especially since, in the heat of battle, often none of them do any good anyway; all one can do is wait for the camera to right itself. Ah well; these are the problems we face.

Then there are the little garnishes – the light bloom, the dust clouds, the way the camera blurs as it turns, to create a more organic feel. The sound of wind, and echoes where echoes belong. The fact that saving – which, given the trial-and-error nature of the quest, is important here – kind of makes sense. In Dead Rising, you save by sleeping or going to the toilet; here, you save by praying at and altar – all natural moments of pause, worked into the play narrative. It all serves to enhance the integrity of the game world.

The Snags

For all the game does right, there are two big areas that diminish that integrity: interface, and the Colossus battles themselves. Not only do they break immersion and irritate the player, either of which is forgivable to an extent; they can also interfere with the game’s themes – which is where we start to enter danger territory.

The interface issues are weird. They are just as clearly a last-minute change as they are clearly a concession to the fear that a traditionalist audience would fail to “get” the game. That gamers would be confused as to what they were supposed to do, and would quickly give up.

Some of the issues, like the HUD, are just cosmetic. When there are only two weapons the player can wield, and it is patently obvious which, if either, Wander is holding, is it really necessary to provide large, cartoony sword, bow, and fist icons? If there are prominent visual and aural cues when Wander is exhausted and about to lose his grip, do we really need a separate grip meter – which, to study, requires taking one’s eyes off the action for what might be a vital split-second? That said, the icons and meters do little harm; all they do is clutter the screen composition, and serve to remind the player that he is playing a videogame.

This sort of issue starts before even the game proper; between the long and wordless approach to the tower and the start of the player’s mission, someone almost-seamlessly wedged in a new bridge cutscene that yaks at length about the plot, about the player’s quest, his destiny, who he will be fighting and why and what will happen, and about how the game is played. At least it’s all in a made-up language, though even imaginary speech feels out of place here.

More significantly, if any game should trust the player to explore and discover the game’s story and nuances for himself, this is that game. If Ueda wanted to point the player toward the first goal, and to hint toward a method for finding the beasts, a better solution would have been simply to place the sword in the player’s way as he leaves the tower. Make it an obstruction. Picking up the sword for the first time, Wanderer would hold it to the sunlight – only for a beam to shoot out toward the horizon, pointing the way. The “use” button would appear in the corner of the screen to illustrate what to press, should the player wish to do this again, later. And that’s all the setup the game needs; the player, unless supremely unobservant, would then or eventually follow the beam out of curiosity.

The same kind of dumb padding is inserted after every Colossus battle: there’s the original, wordless cutscene in which we see a growing crowd of shadows around Wanderer’s prone body. He stands up, gets ready for the player to take control, and then the Greek chorus chimes in again (in a new DVD-only scene), telling you where to go next, what you will be fighting, dum de dum. Who cares. Stop presuming to waste my time for me.

Distraction and Direction

The plot dumps would be inane, if sort of harmless, on their own. Yet the game does not reserve its condescension for the margins between episodes. Some of its most obnoxious moments come when the player can least afford to be distracted, in the heat of battle with a foe way too big for comfort.

If the player takes more than a few moments to study a Colossus for patterns and formulate a strategy, the game begins to scream. An enormous text box will appear in the bottom-center of the screen, obscuring the player’s view. The hint it contains is often as edifying as “EASTMOST PENINSULA HOLDS THE SECRET”. The box stays on-screen for what seems like forever, with no way to get rid of it manually. Then if the player continues to fail at the precise action the game expects of him, the box will pop up again and again, at intervals of what, going by my subjective memories, can hardly be more than thirty seconds.

No! No! Never do this. Never do anything remotely like this. Navi is bad enough. At least the player can avoid Otis. This design decision on its own caused me to put the game on the shelf for over two years. However much I like a game’s ideas, I refuse to be a videogame’s monkey boy, especially if it chooses to scream at me. Especially if, in its concerted effort to patronize me, it causes me to fail at the very task it is demanding of me. Life is too short for that kind of frustration.

If the game must have a hint system, make it optional – as in Metroid Prime. Why so few games follow that model is beyond me. Frankly, though, if you feel a need to dictate to your audience – to ask “Hey, why aren’t you doing what I expect you to?” – then probably:

a)You are compensating for a basic failure in communication elsewhere.

b)Your expectations from the player are far too narrow.


The thing is, as a designer you never ever, under any circumstance, want a player to ask “What am I supposed to do?”. The question the player should be asking is “So how can I do this?”. The player should be focused on juggling the available options, and working out the best course of action – not on second-guessing the designer’s intentions.

If Shadow of the Colossus has a major problem, it is in direct communication of its ideas. Thematic hints and suggestions, sure. Want to explain something outright, though? Ueda’s only recourse is the text box. If he thinks of it. For instance, one of the first steps toward reaching the first Colossus is climbing an ivy-coated wall. The game gives no sign that Wanderer can climb, or that the wall is in any way remarkable beyond its inconspicuously green texture. Had Ueda noticed the possible hangup, he would have thrown in a text box, instructing the player to jump toward the wall and hold the grip button. What he should have done is illustrated the issue. Trick the player into grabbing the wall, by placing a ledge, or the end of an obvious vine or handhold, just out of jumping reach. Show something else climbing the ivy.

Where this problem really manifests is in the Colossus battles. Each is rigorously scripted, with a specific, correct course of action expected of the player. Occasionally the context is sufficient to suggest a technique – as with a flying beast that swoops when you shoot it and that flies in a set pattern past a raised platform. Or an aquatic beast, that surfaces in a set pattern. At other times, the player is left grasping at straws, trying to work out just what is expected of him. One Colossus, it turns out, will stare down into a barrow every time the player runs into it – should the player stay down there for a certain number of seconds. The trick is to exit through the back door, then circle around to grab the beast’s tail. Another Colossus is only assailable if the player backs under an overhang and waits for the beast to bend far enough over that its beard dangles close to the ground.

These are cute solutions, sure. The thing is, they are overly specific, not particularly intuitive, and not really signaled. Discovering them is not so much a triumph of mental or technical skill, or even persistence, as it is trial and error and a bit of luck. The time until the player happens to stumble across these solutions is completely wasted, as there is nothing else of importance to learn. Then typically the beast will shake Wander off once or twice, forcing the player to run through the same scripted motions over and over – making their falseness, their tediousness, all the more obvious. Learn one arbitrary note, and repeat. This is not growth; this is not reward. Granted, that the structure is so asinine and “gamey” certainly fits the game’s themes. Go and play the stupid videogame, or try to find meaning in nothingness. See which option people find more appealing.

Mixed Message

So all of this is annoying, and makes the game obnoxious to play – particularly to a less hardcore audience. It certainly put me off, and I admire the hell out of what Ueda is trying to do with the game. The real problem, though, is that it subtly affects the game’s thematic balance. On its own, the gameworld is idyllic, serene, while the game is hard and obtuse. With his ambition, the player unlocks what darkness is to be found. The player needs to put in a lot of effort to do evil; to discover his own opportunities. Even if you know what to do, the tasks are time-consuming and arduous to perform. Evil is a burden that the player must bear on his own accord – lending all the more poignancy to the game’s central dilemma.

When the game specifically tells you what to do and where to go and how to complete your missions, rushing you from place to place, that changes the tone pretty dramatically. Instead of action being the “difficult” option, finding a reason not to act is the task – and a task that will fail to even occur to many players, as they charge ahead to run through the motions, ride the roller coaster, as if this were any random action game.

Is the game less successful for its faults? Added up, I think so. The message is different. More pessimistic. A bit more obvious, I think. Without the blinking lights, the game would have been more about the subtle, narcotic appeal to evil (in the form of traditional videogame motivation), compelling the player to destroy the beauty around him. Now it’s more about whether the player is even aware enough to notice that he has an option other than evil, then how far and how long he takes that option before giving in to the inevitable. The new message is certainly valid. Perhaps it makes a more classically entertaining videogame. I just think Ueda missed the chance for something more unusual here.

The Devil of the Details

So where do the problems come from? Three things: insecurity, poor nonverbal communication, and ass-covering. The order I’m guessing is something like this:

1) Ueda and his team put together a ton of restrictive, overly clever setpieces, and failed to signal a number of key ideas.

2) During some late tests, players had a lot of trouble figuring out what Ueda’s team expected of them.

3) Instead of going back to clarify the issues, someone decided to slap explanations all over the trouble areas.

4) Now paranoid that people would have trouble understanding even the basic concepts, Ueda’s team stuck post-its on top of every possible object and command in the game.

I’ll just assume, based on the premise that these issues came up in late testing, that Ueda’s team was under time pressure by then – so there were solid arguments for not going back and fiddling too much. So why not make the condescension optional – much like the fully reconfigurable controls and optional 16:9 mode? Maybe it never occurred to anyone. Maybe it was quicker and easier to just slap in.

So there’s a good chance that, by the time the problems with Ueda’s basic approach – namely his tendency to be kind of out-of-touch with the player – became evident, it was already too late. And fair enough. He’s a bright guy, and they did seem to notice the issues – so maybe next time he will take them on board from the start. Still, I sure would like to see a “director’s cut” of this game. Maybe in another ten or fifteen years?