Guest Post: Gradius III and Impossible Difficulty as an Avant-Garde Texture

By LeeRoy Lewin

This latest guest post comes to us from game developer LeeRoy Lewin. Lewin is a member of  Washington-based dev collective VEXTRO and a regular contributor to JRPGs Are Dead. You can play his poem-game, Into the Mouth of Silence, on, and follow him at @wasnotwhynot.

When difficulty in videogames is considered, it’s usually in terms of how difficult a game is to finish, rather than the ideas or feelings it evokes. A game’s status as a consumer product is paramount. Challenge that impedes pleasure lessens the consumer value of a videogame, as it is viewed as leading to a less enjoyable experience, or an incomplete experience, or so on. Fast paced, moment-to-moment reactive shooters like Konami’s Gradius (1985), and especially the third game in the series, Gradius III (1989), are examples—at least at first glance—of ludic difficulty that gate progress around skill. These sorts of games require very physical and intimate understandings of their ludic constructs and are often designed to actively withhold information the player needs to succeed. This kind of negative reinforcement is not a requirement for difficult videogames, but it is a common design approach, and does actively encourage practice, meditation, and ultimately familiarization. Challenges are presented successively, so increases to a player’s abilities are easily measured. Performative difficulty, in this sense, is essentially understood by what it offers the player in terms of gratification for a job well done. 

This is opposed to, for lack of a better term, conceptually difficult videogames. A Mind Forever Voyaging, a piece of interactive fiction dated the same year as Gradius, is a work that takes mental labor to interpret in terms of imaging real political structures. In the game an early, settled America is rapidly adjusted and developed without regulation, ultimately illustrating the consequences of commonly unexamined American beliefs, and delivering a pointed criticism at the failures of Reaganism. (Additionally, now that command line interfaces are no longer dominant, parser-based fiction is a format with its own language and rules that can be intimidating to learn.)

Critiques for either ludic-performative or conceptual difficulty are similar at the root. Videogames that seem to withhold answers or guard understanding are often called pretentious or elitist. A game that necessitates time-consuming physical intimacy is thought of as unreasonable, or even cruel. At the base of these critiques is a feeling of being insulted, for time is precious. I would even say that videogames which center individual labor (when it’s not just an aesthetic of labor, a wholly different application) are considered worthwhile only when they’re able to reinforce a feeling of personal superiority. Time spent needs to be immediately justified or it feels wasted.

I’m going to be liberal with definitions, especially with work that I consider avant-garde or experimental. I believe work that is sufficiently unique, to such a degree that comparisons to other works no longer offer much grounding or understanding, constitutes art that explodes old boundaries while setting new ones. Experimental work is not just revolutionary, but also that which contains so much of a singular soul or style that it’s self-evidently ambitious in its field. (This doesn’t ordinarily pass as groundbreaking in videogames.)

If I were to say that, to me, Gradius III has a stronger and more evocative narrative than The Last of Us, or another darling like Final Fantasy VII, I would fully expect to not be treated as credible by both games writers and hobbyists alike. There are a great deal of ideas and themes presented in the aforementioned games, but they are conveyed rather bluntly. Any feelings of fuzziness that do persist are less a deliberately-applied form of dissonance and more a product of clumsy delivery and an inability to handle a vast amount of plotlines. Gradius III owes its complexity to ambiguity, but it’s laser-focused to build on fear and intimidation. Piecing together the sensations and ideas Gradius III evokes requires a personal meditation and investigation that more popular, so-called complex games do not.

Modernist poetry, especially in movements like futurism (e.g., Velimir Khlebnikov’s Incantation by Laughter), exhibit an aggressive difficulty that not only resists interpretation, but sometimes resists the poem being read at a fundamental level. The poems are vehicles for sensations and feelings; they are better judged by the experience of reading, instead of by any resulting meaning that can be derived from them. Gradius IIIa game that cannot be easily interpretedmay be referred to as a minimalist videogame due to its leanness and utter lack of exposition or effort to model realism. Difficulty is thus twofold in the game: it presents a ludic challenge that disallows progress without the required literacy, and a lasting struggle to make a meaningful connection with the minimalist expression.

Hope Mirrlees’ Paris: A Poem (1918) is another example of a modernist poem that deals more in sensation than in straightforward meaning. We can see in the following excerpt how Mirrlees displays the alienating feelings of a post-industrial city:

I want a holophrase
Black-figured vases in Etruscan tombs
Brekekekek coax coax we are passing under the Seine
The Scarlet Woman shouting BYRRH and deafening
St John at Patmos

If reading is a search for meaning, then Paris is difficult to read. Gestures to explicitly impart what it feels like in the city of Paris aren’t found. Instead, the poem conveys a series of sensations through a stream-of-consciousness that relates literal street signage to the ancient Greek playwright Aristophanes. The poem is not even conceptually difficult: we are told it is about Paris. Paris is challenging to understand because it resists being read, not because it disguises any of its meaning. Something that is difficult to read is a tool for meditation that can open impossible windows of thought. It’s writing that is naturally playful.

The premise of arcade shooter Gradius follows the player-controlled fightership Vic Viper through hostile space. A single spaceship warring against a horde is an ingrained videogame concept as old as the arcade game. A lone hero, or sufferer, was once romantic, the stuff of myths. In an action videogame it is the default, simplest role to render. “Action” here is defined as a state of play in which a player-controlled object will perish to a system-controlled threat if the videogame is left to run on its own, therefore requiring constant vigilance from the player. The system-controlled threats include every other fightership that moves in conjunction with surrounding deadly terrain. The avatar and threats are symbiotic; the objects are filled with meaning as they clash. Thus, the power of a minimalist action game is the microdimension it creates, a space of symbols narrativized only by their immediate interactions.

Extreme contrasts in stage progression derive from a desire to communicate interdimensional travel. The feeling of displacement and dissociation between stages was not (and still isn’t) common or even considered desirable in action videogames. A player travels over a planet’s surface, through an armada of biological weapons, and inside of a large creature, with no explanation of how or why. At the start of each of these very different stages is a refrain: a spot of deep space with no landmass, color, or complicated enemies. A refrain in any composition is a moment to meditate, reflect, and digest the totality of a work. In Gradius they represent both an accent and a place of stillness. They lead me to think of each Gradius stage as a stanza in a moving poem.

Most of Gradius is what you’d expect from a shooter of the time in terms of its interactions, but the final fight gives a distinctive pause. After playing as a machine, with a machine, against unending waves of machines, the final encounter is with the brain of the enemy leader. It’s only a brain, so it sits there, while the screen is fixed in place to frame a boss fight. There’s a moment of confusion as the brain continues to do nothing, and then the player reflexively starts firing. Shooting it does nothing, but a stray missile eventually destroys one of four power sources connected to the brain, each one located in the corners of the screen. Upon realizing this, the player will proceed to destroy the other nodes, and the still moment will end. A still moment is antithetical to the concept of an action game, especially an arcade action game. The distinct lack of a threat is unlike every other minute of the game. I felt bad for the floating, harmless brain. That guilt and stillness led me to think back on the death I’d left in my wake, and the destruction my fragility demanded of me for survival. Gradius ends with me confronting that fragility. The final death of the second living thing in this cyborgian performance closes the loop. Gradius starts again.

A majority of games in the Gradius series end with confronting one’s fragility as externally manifested. In an arcade shooter, a feeling of vulnerability is often the most constant and consistent element driving the player to shoot. Play in these games is really a kind of performance, a desperate memorization of the arcade machine’s demands in a pretense of strength and skill. Underlying all actions of a shooter is the knowledge that one slip will end everything. Vulnerability as a theme is seeped into Gradius and, aptly, vulnerability is the final enemy. A Gradius game is finished when it becomes simultaneously still and vulnerable, relenting to the player. Fatal weakness rendered in these final moments is to be exploited ludically. I can read this as destroying a manifestation of vulnerability itself in vainof course, Gradius will just loop back to the beginning. It could be that the most calculating and practiced individualone who has memorized all of the stepsno longer appears vulnerable, thus earning the right to destroy that which is unable to perform in such a way. It’s an uncomfortable interpretation, even as I believe it, in large part because the music in Gradius speaks to soaring feelings and a justified confidence, framing these actions guiltlessly. The game’s immediate sequels, Salamander (1986) and Gradius II (1988), became even more overt in flaunting their oppressive overtones. Their music sets a much more sunken tone and voice-acted clips quite literally tell the player to destroy.

Despite the healthy success of every other Gradius sequel, the release of Gradius III: From Legend to Myth (1989) was met with such a lackluster response that the arcade cabinets had to be recalled. It was considered too difficult for anyone to play, requiring superhuman dedication and reflexes to learn. After developers retooled it to be less difficult from the start, and added an adjusted beginner mode, the videogame was still not making enough money for operators, and cabinets began to disappear from Japanese game centers.

I found the team’s explanation for the game’s fate in a post-mortem interview sort of funny. Nobody playtested Gradius III to balance the difficulty for a complete playthrough, because they couldn’t. As long as a stage could be completed on its own and was free of bugs, it got a pass. My assumption is that their testers were very skilled at playing arcade shooters since the team expressed a fondness for Gradius and a confidence in their ability to play it. It’s telling that the first sentiment expressed within their general closing thoughts is a positive spin on a common exclamation: Gradius III is actually impossible. Rather by accident, Konami created an anti-commercial videogame in the extremely commercial climate of arcade games.

Because of the difficulty adjustments, Gradius III begins rather innocuously, with a tempo that would be deemed “appropriate” for a new iteration. There’s a tough balancing act inherent in videogame sequels; they demand a pace familiar to returning players, but still challenging enough to feel newly engaging, creating a certain sense of “flow”. Flow, a now-rhetorized and evangelized psychological concept deployed in many modern game design circles, is thought to happen when an activity is challenging, requiring a definite amount of skill, paired with clear goals, and a feeling of personal control. Familiarity enables that feeling of control, so difficulty must be incremental within a comfort zone to seduce players into being absorbed in a pleasurable task.

From various interviews I’ve read, it’s clear that flow wasn’t a game design goal in the ‘80s. Instead, designers aspired to a more nebulous idea of “fun”. They’re intensely related concepts, but to seek flow through design means exploiting feelings of fun in a precise, scientific way, while seeking fun itself suggests something more playful and intuitive. Regardless, flow and fun are often used to center a universality of experience that simply doesn’t exist. Elevated to an ideological level, they court an apparent purity of design drenched in a cloying simplicity that seems necessary to make the next Super Mario Bros.

It’s very fitting, then, that Gradius III becomes completely unhinged in a stage that is an homage to the original Gradius. Enemies in any Gradius game follow a set movement pattern, but always fire in the direction of the player. To compensate, it takes a moment for an enemy to fire, and most of the time they swarm in wide, open areas. Because of this, opposing ships are easy to hit, so it comes down to reacting in a dodge or predicting their patterns to preemptively strike. This isn’t just subverted in Gradius III, it’s ripped asunder. Turrets and carriers are placed in terrain that allows them to fire and spawn more ships, while staying protected from the player. Only capable of shooting forward, the Vic Viper has blind spots in the left corners which are mercilessly exploited by the enemy. To eliminate bases that will continue to spawn ships, open spots in the bullet spread are quickly scanned for and pursued by the player. When reactive strategy seems to succeed and an enemy base is destroyed, little buggers will creep into those blind spots and pelt bullets down. Extreme circumstances, like those surrounding the protected bases, can cause the player to panic and take the clearest route possible. Traversing the maze of flying bullets is no small feat in cramped quarters, but every time an easy or obvious way out is presented, it’s effectively a trap. This is masocore design philosophy, where deep formal knowledge of how a certain videogame works is exploited to subvert the player’s glib expectation.

This represents an awful step away from other Gradius games, in a way. Challenge like this isn’t confronting vulnerability. If an acceptable level of challenge in a videogame is one which is regulated to produce a flow state, and is rooted in understanding limits and presenting a clear win, then Gradius III isn’t challenging, but closed. It even feels cruel, like my fragility is being exploited and rejected. The game is so snappish and unpredictable that it provokes an acute feeling of intimidation in me. I can’t assert myself in this space; I can only stumble and grope around what is effectively a minefield. While I’m staring peripherally, weary and overwhelmed trying to keep up with the clutter of things on the screen, slowdown kicks in.

Takemasa Miyoshi, the lead designer, expressed embarrassment with how the game ended up playing, so I think slowdown was not desired. Choppiness, framerate drops, or anything other than a game running smoothly, speaks to broken software. Consumers of many stripes mock or become veritably angry with a videogame that doesn’t run properly, as if there is only one way a game should run. Even if those things are true, which I hope someday they aren’t, Gradius III owes its iconoclastic feel to its slowdown.

When the screen fills with too many sprites, the game slows to a crawl. Gradius III tends to become densely populated when it gets difficult, so slowdown occurs when a player would need it most. This sometimes leads to an aggravating slingshot effect. It can be hard to tell exactly how many things are causing the slowdown and which, when gone, will cause the game to return to its original speed. Because of the decreased frame rate, small spatial adjustments take proportionally longer directional input, but it can feel like any number of bullets or enemies disappearing reverts the speed, and a long input will suddenly send a player flying to defeat. Not being confident which mode of control is needed to play really compounds feelings of anxiety.

Towards the latter half of Gradius III, it shifts from incidentally having slowdown at the hard bits, to incorporating the slowdown itself as means of play. As the level design suddenly drops specific encounters and instead stuffs the screen full of as many things to dodge as possible, the amount of sprites on-screen is just enough for the game to avoid slowdown, meaning the player can manually cause slowdown by firing. In stage seven, this occurs because there are an extreme amount of meteors populating the screen (slowdown is referred to in-game as “wait” and can actually be turned off in subsequent ports). Shooting a meteor causes it to split into three much smaller shots: one continuing in the same forward trajectory, one going up at an angle, and one going down. Firing a shot and a missile will nearly always slow the game down. Weaving through the meteors, firing when it seems safe, manipulates the time of the game itself, and poses a choice to the player to endure either larger, but faster obstacles, or many, smaller and slower ones. Each shot lets the player breathe, then rebounds into a harder to manage state. Violence used to slow and seemingly manage the anxious state over time makes the stage unmanageable. But feeling out the game and finding opportunities to destroy, while also exhibiting restraint, creates the best chances of survival.

Few videogames test raw reaction speed like Gradius III. Unlike the exact and unchanging arrangements in stages, bosses act erratically, randomly selecting a behavior out of several, cutting down the effectiveness of practice and memorization. Attacks are not telegraphed, so defense can’t be predictive; dodging is only possible during the flash of an attack. Observing a few attacks teases out clearings, but this leads into the game’s masocore tendencies. Safe spots are bait, and bosses have attacks to explicitly punish availing oneself of them. Players must face death head-on, like a matador, in the hope they can sidestep destruction.

Arcade shooters are expected to bend with practice. Seeing the fruits of one’s labor materialize, even when it’s essentially memorization, is very affirmingand likely why Gradius III maintains a negative reception among shooter fans. There’s no fulfillment of labor or power because strategies are constantly being subverted, if not outright denied. The cube rush is an excessive, indulgent denial of the pleasures of a shooter. For over a minute, players must dodge glass cubes and gunfire in a cage match. Cubes randomly activate, dashing in the current direction of the player. Or just as likely, a cube will float in its still trajectory from right to left, sticking to the wall of glass and other cubes behind the player. There’s an active puzzle to guide cubes into specific positions to protect the player. Since it’s impossible to know which cubes will charge, building something relies on fervent hope. It can even backfire as misaligned cubes take up a player’s precious currency of space, giving them less room to maneuver. Memorizing which cubes fly at the player in order to deliberately stack them would take a great effortif it weren’t impossible. When the stage begins, one pattern of twenty is selected at random to determine which cubes are aggressive, thus blocking simple memorization. The cube rush is a set piece dedicated to its own pointlessness, indifference and denial of satisfaction. I think of the excessive depiction of signage in Mirrlees’ Paris: a recursive difficulty for its own sake.

Impossibility, or its near-equivalents, have no purpose in consumer culture. I confronted this when locked in a prison in The Beginner’s Guide. It’s a moment which is supposed to last four hours, but ends in an instant. Four hours would be excessive, narrator Davey Wreden says. The value of the idea would come across in a shorter amount of time. Wreden asks the audience what the point is of a game if it isn’t playable or accessible, or if a videogame even means anything if someone can’t get through it. I suspect this monologue was part of the inspiration for a teased piece by Brendan Keogh called “Videogames Without Players”:

“Here’s two ways to think about the player’s role in videogame play:


1) The player as the centre of the world. The videogame should provide the player with opportunities to make meaningful choices and to exert their agency in a meaningful way.


2) The player is the weakest, slowest, and most flawed component of an otherwise perfectly efficient cybernetic circuit. Designed without a human player, the system would work perfectly, without hiccups, and much faster.”

Gradius III is tragically beautiful because of its futile, cyborgian expectations. Though likely not the intention of the passage, I find it more useful to read “player” as an axiomatic thing. The player-construct that is rhetorized and strictly heeded to to make a successful videogame is characterized as a kind of ghastly everyman that substitutes human potential for a common human denominator. The player-construct is dogged by a sycophantic ideology, and it creates a culture that expects meanings and goals to only slightly fall outside its reach, if at all. Konami made a specific game, and at that time, pushing the envelope in anything other than technology was a hard sell in arcade culture.

A murky unknowability that envelops the player drives the tight knot of miraculous struggle in Gradius III. The sensations this game conjures are as difficult to grasp as in other challenging works, such as Yume Nikki, Problem Attic, and Pathologic. Difficulty is a transformative texture tapping into ambiguity. Unreasonability, open to being interrogated, makes room for a personal absorption of perspectives far removed from the norm. If difficulty is avant-garde, and if the avant-garde is transforming or even breaking accepted approaches for idiosyncratic expression, then Gradius III was once an avant-garde work.

Despite being panned at release, Gradius III has spawned a wonderful lineage. The density and application of the enemies became a motivating example for other shooter devs to push their limits. Gradius III’s failure inspired the intensity of the Raiden series and the horizontal shooter swan song of the R-Type series. These erratic shooters demand tight performances in ways that inspired the kaleidoscopic beauty of danmaku, or “bullet curtain” games, where torrents of vibrantly colored bullets fill the screen. Danmaku games can take the form of affecting dramas like Ikaruga or Touhou, the pure fractal abstraction of rRootage, or the esoteric mystery of Hellsinker. The recent popularity of Undertale and its heavily choreographed, danmaku-influenced dances, demonstrates the far reaching and unpredictable importance of seemingly “failed” avant-garde works.

It is true that ludic difficulty is more culturally acceptablewell, I should say subculturally acceptablethan other kinds of difficulties. What’s being celebrated is not the difficulty of the sensations provoked, or the destruction of the status quo, but the selfish ability to overcome difficulty through individual merit. That is an undeniable facet of difficult works; and there is a signaled egotism to understanding things or being able to do things others cannot. Understanding challenging videogames in a historical context undercuts that individualism, treating difficulty as a texture of experience and a tool for complex understandings, instead of as a recursive measure of one’s ability.

Videogames are so often designed intently to be flattering and comfortable, to be explicitly familiar and unchallenging, that there’s little appreciation for avant-garde works of any stroke. Even ludically difficult games are approached like this. I feel the reality is that an axiomatic player-construct is holding back the potential of difficult, even ostensibly impossible expressions.

Player-constructs are described boldly by Mattie Brice in KILL THE PLAYER:

“The existence of the player-construct enforces a product-consumer dynamic that wields power according to imperialistic and capitalistic systems of value over art, politics, and life.


The existence of the player-construct ruptures the connection between play and life through its dependence on the “magic circle,” which purposefully invalidates play that attempts to create meaning outside of preconstructed fantasies and within a subject’s lived experiences.


The existence of the player-construct allows only for experiences that require consent to play to exist, obscuring play experiences that humans are involved in unconsciously and non-consensually from recognition, analysis, and intervention.

The cultural impact of videogames is different and much broader than it was in the 1980s, yet games are still generally thought about in arcade terms. Metrics of immediacy are of utmost importance. Avant-garde approaches to game design have always been explored, but they’re not part of our canon for the most part. Instead, we tend to casually and flippantly declare what sort of game elements are good based on overtly pleasing and essentially unchallenging games of the past. Avant-garde videogames will nearly always fail based on those measures. It’s been decades and game culture as a whole is still failing experimental work, and is unsure of what to do with games that are broken, unreasonable, or even indifferent, like Gradius III.

Legitimizing fringe work is essential to creating a space accepting of diverse perspectives and experiences. Paris: A Poem was only retroactively considered an influential modernist poem by scholars many decades after its publication. Before the poem’s reappraisal, it went largely unread and misunderstood. No doubt its reception was a factor in Mirrlees later distancing herself from literature. There is somewhat of a parallel here to how Gradius III, and many other avant-garde videogames, have been received in the past. Difficulty, even in arcade videogames, is not a linear or binary concept, but a synthesis of expressions that are outside comfortable norms, and therefore take time to understand. Meeting resistance in a piece can be a sign of some deeper, intangible message, but current perceptions limit it to an egotism rooted in pretentiousness or pride. Even in a seemingly straightforward and egregious example of ludic difficulty like Gradius III, the challenge resonates as a tool of expression, a tonal texture, and creates an additional layer of conceptual difficulty when attempting to interpret the overarching themes. Viewing difficulty as additive to meaning, rather than merely as an obstacle to be overcome, helps us to build a bridge between the history of abstract and avant-garde art with the wide range of ludic experimentation taking place today.