on Thursday, December 5, 2013
Jeff Wheeldon recently published an article entitled, "Video Games, Texts, and Interpretation," over at Push Select. Its central thesis argues that the fundamental frameworks that we use for thinking about artistic works other than video games do not apply to video games because games cannot be treated as texts (where we are using the word "text" in the literary sense of a coherent set of signs original and inherent to an object of analysis). His argument wanders through Francis Watson's non-standard usage of the terms "work" and "text" and touches on the role of authorial intent before winding up at the main crux of the discussion. Wheeldon asserts that "a video game puts the player, the reader of the text, into the shared role of authorship." (Emphasis his.) Ultimately, he claims, this essentially renders the game non-static among players, and it thus cannot be treated as a text in the traditional sense.

This argument makes little sense. Indeed, the very act of aesthetic analysis implies a text that is the subject of the analysis. That statement, however, perhaps avoids Wheeldon's point, because one might better phrase his argument as saying that each game player is analyzing a different text due to their subjective experience arising from their "shared role of authorship." Such a claim, however, is really just an appeal to complexity. Wheeldon is mistaking the variable aspects of the subjective experience of a text, which exist when experiencing any sort of medium, for changes to the text itself, merely because there are a larger number of variables in the subjective experience of games than in certain other media.  This difference is perhaps easiest to see by way of examples in some of those other media.

Consider first linear, primarily narrative media such as feature films and novels. Wheeldon seems to easily identify these objects as texts when they are the subject of analysis. In these cases, the text is the synchronized and sequenced images and sounds in the case of the film and the sequenced characters and words in the case of the novel. All viewers and readers are attempting to identify meaning in these well-defined sets of symbols. Their experiences of these symbols, however, are not necessarily uniform. There are many obvious potential differences.  Viewing or reading locations may differ, changing the physical  and emotional state of the audience, or past knowledge and experiences may change what information can be gleaned from the text. Technical details like lighting conditions, font size, book bindings, screen size, page-turning speed, etc, will affect emotional impact. Though such experiential differences may affect any one audience member's ability to produce an analysis of the text or the emphasis of such an analysis, these are, for most purposes, clearly distinct from the text.

The scope of these experiential differences can be even larger for other media, though. Consider sculpture or architecture. The viewing patterns for these objects vary wildly with each viewing. In addition to many of the aformentioned complexities of the linear narrative forms (lighting, etc), we have 3-dimensional space with which to contend (note the similarity to many video games). Still, traditional criticism has had little difficulty analyzing such texts. We can enumerate endless other examples (music, poetry, performance art, etc) with their respective difficulties, but in all cases we have little difficulty identifying a text for discussion.

The same is true of video games. In this case, we have a system of rules, video assets, and audio assets as our text, and the player interacts with that text just as the audience does in the case of any other medium. The veneer of agency that most games offer is not far removed from the other experiential differences that we have discussed. One explores a 3-dimensional game world as one would wander around a sculpture (imagine, for example, walking through Sandy Skoglund's "Fox Games"); one chooses levels or scenarios as one might choose the order to view a series of paintings (like, perhaps, Monet's Haystacks series). In all cases, the rules and assets were put in place by a developer as a set of signs that convey meaning to the player -- the definition of a text. That text is static among players, and it is the subject of analysis that we discuss when writing about video games.

It is curious that Wheeldon has chosen to argue against this paradigm. He mentions several times The Stanley Parable, which (in both its incarnations, the 2011 mod and the 2013 game) takes the non-agency of game players, and by extension the textual nature of games, as one of its central themes. Campster, over at Errant Signal, recently described this theme eloquently:
Players in game systems aren’t free, they’re exploring a text with rules put there intentionally by an author.  The Window Ending, for example, references how players will eventually explore both paths to consume all of the content: [example] The Stanley Parable posits that thinking a player’s choices signify free will in a game is like thinking fictional characters have free will over their own actions when the whole text is already written out before them.  Your choices in games aren’t decisive in nature, but instead they’re exploratory; you’re not making lasting decisions but you’re choosing a branch of possibility space to examine.  You don’t have the freedom of choice, you have the walls of rules and mechanics that give shape and meaning to the game even as they confine you.
As in all novels, films, sculptures, and other art, this is true of all games. Even the simplest games are texts quite dense with information, so it is easy to mistake the variables that create experiential differences among players for an authorial agency. But it is indeed a mistake, and games are texts like any other. Games criticism must integrate the unique aspects of these texts in a way that has not been done for other media, but it is an extension of other aesthetic theory, not a break (an idea that I have tangentially touched on in the past).
on Thursday, August 15, 2013
It would be very peculiar indeed if a critic were to publish an academic essay on an abridged version of Gravity's Rainbow or a version of Eraserhead with explanatory voice-over added. Not only would this hypothetical critic's experience be far different from that of countless other readers or viewers and thus not generally applicable, but he or she would be writing of a totally different work of art with an additional author other than Pynchon or Lynch. Let's eliminate this complication and assume that Pynchon himself abridged the novel and Lynch, in his odd drawl, added the narration to Eraserhead, both artists doing so to aid the less experienced audiences in understanding the work (perhaps because their distributor insisted on such a version). We would still have a problem with these secondary works and their use in the critical discourse on the primary works. Part of this problem is just the aforementioned issue of disparate experiences, but there is a second issue related to the artists' decisions when creating the works. At least one of the two different (hypothetical) versions of each of these works must not be the maximally effective version for conveying the artist's themes. We tend to recognize this immediately with abridged versions of novels, because large swaths of content and style, which the author clearly intended to be a contributing factor to the interpretation of the novel, are missing. We could argue that something similar happens with the voice-over example, but in this case it is an issue of excess: every part of the work is not contributing maximally to the thematic content, and some pieces may now even be subverting or distracting from it.

Gravity's Rainbow and Eraserhead are, of course, quite difficult texts. They are not suitable for all readers and viewers because it takes a certain familiarity with the grammars of their respective mediums to make meaningful sense of them. As such, they are inaccessible to most of the populace. Few people take issue with this fact, though.  We accept that certain artistic achievements require some amount of training to apprehend. We work up to them, and we train our children with simpler works. At times we may use abridgement or adulterated videos as teaching aids, but we do not view these teaching aids as an end in themselves. We accept that an artists work, with each element chosen to be maximally effective in service of its themes, may naturally present a certain amount of difficulty, which may be large or small, to its audience.

Why, then, do we insist on multiple versions - that is, multiple difficulty settings - of each video game?

My comparison of video game difficulty to literature and film is certainly a highly flawed one. I am comparing apples to oranges in that video games alone offer relatively simple routes for the author to implement variable difficulty within the primary work. Further, difficulty is rarely used as a direct mechanic in any medium other than games; in other media, it is usually simply a by-product of a particular stylistic technique. This is a very significant difference because it suggests the possibility of variable difficulty being a device that can carry thematic meaning in and of itself. But my comparison to literature and film still has some suggestive relevance to video games because the majority of games, especially narrative-driven games, are not using variable difficulty as a direct mechanic, but, rather, include menu-based difficulty settings to maximize the appeal of the game to a variety of audiences. This use of difficulty settings implies that game designers are not typically using in-game difficulty with thematic purpose, but instead, they are injecting it for some tangential purpose, such as consumer demands or mere tradition.

In April and May of 1842, Edgar Allan Poe published a pair of reviews in Graham's Magazine of Nathanial Hawthorne's Twice-Told Tales. These reviews would go on to become quite well-known and influential pieces of literary criticism owing to their careful delineation of the significance of the short story. Poe celebrates the short "tale" as a genre, particularly because of its ability to sustain a "unity of effect" owing to its total consumption by the reader within a single sitting.  He notes that such a unity of effect requires of the author that "there should be no word written" which does not contribute to the "pre-established design." Most video games, of course, are not consumed in a single sitting and thus do not sustain a single unity of effect quite as Poe imagined for the short tale (though Poe might well have been impressed by our ability to serialize such effects as we do with modern TV and games). Nonetheless, his concept of the unity of effect is an instructive one.

At every point in the creation process of a work of media, the artist is making a representational choice with every piece that goes into the work. Likewise, the audience is making an interpretive choice at every moment in the experience of that piece. Each aspect of a work of art is a tool for creating meaning in some way, and something like Poe's unity of effect is achieved when all of these pieces are used together effectively in service of theme. If a game designer abdicates responsibility for feelings of frustration or achievement and tense focus or casual relaxation, then the designer is abandoning a portion of the thematic control the medium offers. No longer does the designer have any information about the emotional state of the audience at any given moment, and thematic meaning through game mechanics may be lost. There is no unity of effect, for at any point various aspects of the game may be working against each other emotionally. For these reasons, I suggest that it is not enough to simply take user-controlled difficulty settings as a given necessity and merely discuss various attitude toward them (as people have done, for example, here, here, and here). Instead, designers and critics must address this problem more fundamentally. They must ask what difficulty means in a given game, and seek to wield it as a tool for creating meaning. At times, this may mean removing control over such aspects from the audience.

I imagine that many players and designers would view the sacrifice of such user control as something that approaches sacrilege. To anticipate some of the objections, I note that I am not advocating a wholesale abandonment of variable difficulty - I am advocating its careful application with artistic purpose over haphazard adherence to convention, which at times, especially in some narrative-driven games, may mean abandoning difficulty settings. Perhaps it is no accident that many of the most artistically acclaimed games, such as Braid or Ico, lack the community-mandated difficulty options menu. Others might object to a lack of user-controlled difficulty settings from a more practical angle: customers expect them, so their removal might be a financial mistake. I suspect that this objection rests on faulty assumptions, because I doubt that any reliable statistics exist regarding how this would affect sales. Still, if it is an issue, it could be easily mitigated. In Valve's multiplayer shooter, Team Fortress 2, the default settings cause a warning to pop up if a player tries to join a server with more than 24 players. The message warns that the game has been designed for 24 or fewer players, and may not create the same experience with more than that. Because it is relatively simple to implement features like reduced or increased difficulty, designers could accompany alternate settings with a similar warning.

The demand for players to have endless control over their in-game experience in all cases, such as with difficulty settings, is a symptom of how much of the populace views video games. In demanding such features without regard for thematic meaning, we are treating games as consumer products rather than artistic media. Surely some aspects of consumer demand inform all media, but I would prefer that we treat games more like we treat film and literature and less like we treat microwave ovens and lawn mowers. If we are to move in this direction, we must view every game design decision in terms of its artistic meaning and its effect on the player. Re-evaluating how and when we utilize variable difficulty may be an important part of such a refinement.
on Friday, July 26, 2013
There is a peculiar, though perhaps not wholly unexpected, trend in the way many people view and describe the experience of video games: players are demanding "immersion." Though this word is often ill-defined, it usually refers to something like the player's ability to fully identify with their avatar to the point of "losing themselves" in the game.

In some discussion forums, this word pops up in nearly every discussion that makes value judgements on games. For example, as of 2013-06-26, about 5% (360 out of 7260) of posts on Reddit's TrueGaming subreddit, which purports to be dedicated relatively serious games discussion, take "immersion" as their subject, and the subject arises in thousands of other posts in the comments sections. Authors of games criticism often take immersion as an a priori desirable facet to any narrative game, as Stu Horvath did in a recent lamentation of his inability to feel the frailty of the characters of The Last of Us. This assumption seems faulty, and I suspect that games criticism is doing itself a disservice with its obsession with 'immersion.'

But before we get to that, it's perhaps best to clarify the concept of immersion a bit. Players are not referring to the "immersion" of typical academic discussions of art, which is characterized by maximal illusion of reality (i.e., attempts at virtual reality; see, e.g., Grau's Virtual Art). Instead, when players talk about immersion, they are really referring to an aspect of narrative distance, to use more standard literary terminology. In narratological studies of literature (and, by extension, other narrative texts like film and games), narrative distance refers to the 'separability' of the narrator's point of view from the story presented. Put another way, it is the extent to which the audience feels directly involved in the narrative - does the narrator place the audience's perspective directly into the story or is the perspective removed from the action? To a large extent, narrative distance is a stylistic choice of the author, but it is also impacted by the audience's approach to the text, as Film Crit Hulk illustrated in a recent article about audience reactions to spoilers. So far, there hasn't been all that much talk about these concepts as they relate to games in formal sense, although Matthew Schanuel almost hit upon upon it over at the Ontological Geek, albeit in a rather limited sense.

The issue with gamers' obsession with immersion is that it is formally limiting. It is a similar to the problematic nature of discussing game in terms of how 'fun' they are, as Campster over at Errant Signal discussed last year. Briefly, Campster takes issue with 'fun' as a critical term for games both because it is ill-defined and implicitly excludes a wide variety of modes of engagement with the player (e.g., can a game convey feelings of suffering, rage, or cautious reason if it strives only for an ill-defined 'fun'?). These same issues arise with the term 'immersion.'

To begin with, the term is presently far too ill-defined to be useful critically. A simplistic reading of 'immersion' suggests that it is increased by minimizing narrative distance. Taken most literally, this suggests that first-person games rendered realistically with no HUD, cutscenes, or menus should be most immersive and, if immersion is a priori desirable, most successful. This definition excludes much of the term's usage, though. Engrossing third-person RPGs and MMOs (and even, sometimes, 2D platformers like Limbo) are often held as pinnacles of immersion, as a quick Google search will readily demonstrate. And while many players disable HUDs to increase immersion (indeed, many PC games have modifications available that remove HUD elements), it is not self-evident that this tactic actually does increase immersion in the sense that many players mean. Roughly aiming a gun in the real world, for example, is relatively intuitive even if one is not looking down the sights, thanks to our natural spatial awareness of our body parts. In a first-person shooter, the crosshair usually acts as a crutch to simulate this awareness. Removing it may counter-intuitively decrease immersion by forcing the player to more consciously consider the aiming mechanic itself. When using the term 'immersion' critically, these alternate interpretations are not usually made explicit, so the term loses any explanatory force.

Perhaps the more worrying aspect of the immersion obsession is how it limits the formal options that we tend to assign to games. Let's return to considering literature and film, for a moment. Most bestsellers attempt relatively low narrative distance. The work of Dan Brown or romance authors typically places the reader close to the action, taking omniscient points of view while emphasizing the physical and sensory aspects of plot advancement. Similar approaches exist in blockbuster films. This approach is obviously quite successful in terms of interesting consumers, and it can create great empathy with characters. If we widen our net a bit, however, other successful options become apparent. In the canon of classic literature and film, many of our most revered texts are those rife with formal experimentation. (We will not mention other media, such as poetry and sculpture, that typically make no attempt whatsoever to create anything like immersion.) A work like Joyce's Ulysses, frequently considered among the greatest novels ever written, is hyper-allusive, carefully structured, and full of word-play. In short, it employs many techniques that require a large narrative distance to fully appreciate, and it thoroughly discourages long-term immersion despite the elaborately realized world that it creates. Even more extreme is something like Italo Calvino's If on a winter's night a traveler, which alternates between the second and third person points of view, telling the story of the reader's attempt to read various novels. This novel self-consciously plays with point of view and narrative distance to examine how the reader-author relationship informs the act of reading and how that, in turn, affects our lives. Immersion is acutely broken in these works in service of their themes, and the effects achieved by doing so could not have been obtained through other means.

By implicitly asserting that immersion is always a desirable quality, we are discouraging the analogous formal avenues in games.  While there will always be experimentation in the independent game development sphere, the obsession with immersion likely limits the ability of more mainstream development to use new and experimental devices in their game design. Perhaps more troubling, this tendency toward the praise of immersion limits the directions that we take in the critical discourse on games. We should not always be asking how a particular game could be made more immersive. We should not take immersion-breaking devices to be flaws in a particular work. Instead, we should be asking how different devices and narrative distances function within a given game. These are choices that game developers must consider when creating their game. If immersion is broken, we should not discard that as mistake but instead seek to understand how it alters the creation of meaning for the player by the game. If we encounter the game-equivalent of If on a winter's night a traveler, we must ask why the game is calling attention to our gameplay, not ask for a different game.

Games certainly have a unique ability to draw us into their worlds. That ability to create immersion opens up thematic pathways that are closed to other media, and that feature is often rightly praised. But surely games can also do other things. We already see this experimentation popping up in a range of recent titles, from the frenetic jump-cuts of Thirty Flights of Loving to the direct address of dys4ia. Let us not dsicourage the spread of such experimentation, and let us not do a disservice to games that do experiment with narrative distance by doing them a critical disservice.
 
--------

For additional interesting reading on the concept of immersion, see Jamie Madigan's article over at The Psychology of Video Games or the High Level Storytelling Design article at Frictional Games's blog.
on Thursday, September 27, 2012
Sexism in gaming is most frequently discussed in terms of objectified women in scant clothing that pervade action games. But let's not forget other, arguably more insidious forms. Like this deal from Amazon: The Viva Big Bundle of Games for Girls. Because girls certainly only care about Riding Academy 2 or the classic, My Boyfriend. Anyway, here's a screenshot of this wonderful collection in case the promo page gets taken down when the sale is over:


I find it pretty awful. Why not just call it something like the Kids Animals Pack or something and omit My Boyfriend?
on Tuesday, September 11, 2012
A recent paper by film critic Thomas Elsaesser (Elsaesser, T. 2011, New Review of Film and Television Studies 9, 247) offers an interesting take on how certain Hollywood  movies invoke interpretation in the spectators. (You may want to at least skim it before reading this post.) He posits that a new sort of auteurism ("post-auteur") has arisen, and he discusses it by analyzing the work of James Cameron, especially his Avatar. In short:
This (for lack of a better word) post-auteur authorship can usefully be discussed in the case of Cameron under several headings: auto-representation and personalized narrative, affective engagement with diverse publics, ambition to effect through technology a change of paradigm. The first I shall discuss as 'control through access for all', the second as 'control through switches of premise and double binds', and the third as 'control through performed self-contradiction'.
The control that is essential for any auteur theory is thus manifesting itself in a new way. Without reiterating the whole analysis, Elsaesser argues that Cameron carefully systematizes control of the audience's reactions by presenting mixed signals that induce cognitive dissonances. These dissonances "provoke the spectator into actively producing his or her own reading, in order to disambiguate the 'mixed messages' or to untie the knot of the double bind." Each spectator, then, arrives at a reading of the text that is at once at odds with the film and other readings but which results in a stronger "'ontological commitment' on the part of the viewer to his or her particular interpretation – a commitment that works in favour of the affective bond formed with a given film."

But what does this have to do with video games? It strikes me that this is essentially a description of how interactive media functions with its audience. The narrative contradictions that create this effect – dubbed "cognitive switches" by Elsaesser – manifest themselves in games as player choice. The dissonance that a filmic auteur like Cameron can choose to create in his audience is inherent to all interactive media by virtue of the changing experience from play-through to play-through. While Avatar induces different experiences in the spectators' minds, interactive media makes these differences literal in the text. While an author of interactive media may be able to achieve a level of control that fixes the number available readings, the default mode of creating meaning is one of these cognitive switches because the player is forced to make ontological commitments toward a particular reading with every interactive choice.

Game critics often compare interactive media to filmic media, and we often interpret games using cinematic modes of thinking. Game developers, too, clearly follow many cinematic conventions in structuring their games. Most commonly, this approach manifests itself as a straightforward discussion of narrative structure and visual presentation. These modes of interpretation lie squarely within the bounds of classical narratological arugments and types of spectatorship such as the voyeurism of feminist film theory. But these methods are clearly inadequate for games, which require ludic approaches as well. In games, the spectator is empowered, so we cannot, either in design or interpretation, use only passive approaches in our thinking. We cannot apply passive film theories to active games.

What film critics like Elsaesser make clear, however, is that filmic media, especially within certain recent trends, also create meaning within an active context analogous to that of games. We see this spectator empowerment emerging prominently with the rise in popularity of "puzzle films" or "mindgame films" (see, for example, Puzzle Films: Complex Storytelling in Contemporary Cinema, 2009, ed. W. Buckland, West Sussex, UK: Wiley-Blackwell for an extensive treatment or the extensive bibliography over at Film Studies for Free). Auteurs such as Christopher Nolan, David Fincher, Charlie Kaufman, M. Night Shyamalan, and others are disorienting spectators in ways that reward detailed pattern recognition, and they are crafting films with viewing experiences that are substantially different on subsequent viewings. To extensively quote Elsaesser from his chapter in the aforementioned Puzzle Film edition,
[T]he main effect of the mind-game film is to disorient the audience, and put up for discussion the spectator–screen relationship. The notable emergence (some would argue: reemergence) of mind-game films since the mid-1990s would be one sign of this "crisis," to which they are the solution at a meta-level.... [T]he mind-game films set out to aggravate the crisis, in that the switches between epistemological assumptions, narrational habits, and ontological premises draw attention to themselves, or rather, to the "rules of the game." These rules, in addition to what has already been said about them, favor pattern recognition (over identification of individual incidents), and require cinematic images to be read as picture puzzles, data-archives, or "rebus-pictures" (rather than as indexical, realistic representations).

Thus, what appears as ambiguity or "Gestalt-switch" at the level of perception, reception,and interpretation is merely confirmation of strategy at the level of production and marketing: with the mind-game film, the "institution cinema" is working on "access for all," and in particular, on crafting a multi-platform, adaptable cinema film, capable of combining the advantages of the "book" with the usefulness of the "video-game:" what I have called the DVD-enabled movie, whose theatrical release or presence on the international film festival circuit prepares for its culturally more durable and economically more profitable afterlife in another aggregate form. Which would lead one to conclude that the mind-game films make "mind-games" out of the very condition of their own (im)possibility: they teach their audiences the new rules of the game, at the same time as they are yet learning them themselves.
We thus see that recent films have activated the spectator by changing their artistic mode, and that this style is partially driven by the multi-platform, database-like way that we now consume media. Just as cinema has so informed how we structure games, there is little doubt that games have changed how we consume cinema.

What I hope to accomplish with this post is to demonstrate that filmic media, especially in the last few decades, activates its audiences in ways apart from narrative and cinematic images, which are the typical, but insufficient, points of comparison to games. These films empower the spectator to discover their rule-sets and make ontological choices, not unlike players do in games. The grammar of film studies is well-developed, but critics are just beginning to grapple with these new types of active interpretation and consumption. As game critics grappling with the same difficulties, we would do well to understand how an active audience informs our understanding of film, because that will certainly inform our understanding of interactive media as well.
on Monday, September 3, 2012
    2.5/10
    • Biart
    • Simulator, First Person Shooter
    • Release: Jan 18, 2012 (US)
    • Platforms: PC





    Some projects feel as if they were created without purpose or direction. They stumble along with no idea how to acheive their goals, because their goals never really existed in the first place. To my great disappointment, Depth Hunter is one such title. As a diving game (or is it a shooting game?), it had a unique niche of the gaming world to explore; it sadly left it empty.

    Depth Hunter touts itself as a "Spearfishing Simulator." This is sort-of true; you always have a spear gun with which you can shoot fish underwater, which you then attempt to haul in by managing the tension on the line. This is trivially easy if the fish is within about 8 meters and impossibly difficult outside of that. The problem is that there is no real reason to kill these fish. The game has a campaign mode of sorts, in which you are placed in the water and given tasks to complete. These tasks come one after another without context; you will be told to kill two fish of a certain king and then you will be told to kill as many fish as you can with a certain time frame. Sometimes, the tasks aren't fishing related at all. The player might have to find certain "treasure" objects on the sea floor or photograph a certain number of moray eels. The photography tasks are particularly bizarre; the player must find a correct angle that changes an on-screen number to 100% before the screen capture will count. All of these tasks take place over two of the game's three locales.

     As you might guess from my description, the campaign lacks direction, which makes it astoundingly uninteresting and painfully boring. It injects arbitrary tasks into the player's exploration of the game world, seemingly because it is undecided as to whether it wants to be a linear, goal-oriented game or an open-ended, environmental-exploration driven game. It therefore does neither well, despite having ample opportunity to do something unique with either. By injecting a goal-oriented gameplay mode, it suggests a narrative and character for the player. Indeed, the players boat sits on the surface of every level. I was profoundly disappointed that I could not get out of the water and into the boat. Aside from the potential as a fast-travel mechanism or a place to change your loadout like you would expect for a simulator, it would have allowed the game to create a real character motivation. The game could have, for example, given the player a back-story to motivate the tasks; that narrative could have uniquely explored any number of subjects, such as the economies of traditional fishing cultures, the impact of tourism, or even just a competitive sport-related theme. But the game does nothing, and the tasks appear pointlessly.

    One might note, however, that a narrative is not essential to a game. In a game-world, the environment itself might construct meaning and engagement. And it very well could have here. A fully realized ocean environment could provide more than enough content. It could act in an educational fashion, for example, elucidating the mysteries of reef environments or the perils that they currently face. Alternately, the game-world could simply create wonder in its audience with a truly beautiful, detailed, interactive world. Depth Hunter, however, is too poorly realized to pull off any of these things. The few types of fish in the world have even fewer behaviors. Big creatures (manta rays, sharks, moray eels) don't react at all; they swim in a fixed pattern or sit stationary. Small fish either swim back and forth or swim away if approached too close. It's good that they swim away if you get close, because the models and textures aren't particularly detailed or convincing. They are far better than the repetitive, plastic-looking ground textures, though, which are decorated by the same piece of coral thousands of times. The whole art design is just far too simplistic for the environment it tries to capture, so none of it works. Couple that with the total lack of interactivity, and there is simply no content in the environment to drive the game.

    Ultimately, I suspect that Depth Hunter was conceived as a something like a tech demo for its game engine, and that no one really thought about how the game should function. Its disparate elements (spearfishing, photography, explorable underwater worlds) never cohere, and none is done well enough to be anything of note on its own. I applaud Biart for taking an uncommon premise to production, but that is about the only praise I have to offer. I wish a game would really run with the underwater exploration conceit; sadly, Depth Hunter does nothing but botch it.




    Technical note: I hate having to discuss mundane software issues, but it's worth noting that Depth Hunter validates your serial number on every start up. About a quarter of the time, however, it seems to fail, at least on my machine. I'm not sure if this is some sort of broken, local check or if it is trying to connect to the developer for validation, but it is annoying nonetheless. Restarting the game usually corrects the problem, at least until the next time it goes awry.



    on Thursday, August 23, 2012
      5.5/10
      • E. McNeill
      • Real time strategy
      • Release: Unknown, out now.
      • Platforms: PC (version reviewed), Android



      Auralux is a study of sorts. It is an experiment, like Eufloria before it, in the most basic mechanics that define a real-time strategy game; it attempts to boil the genre down and show us how it works. There are no tech trees here, and there is certainly no grand historical context for a setting. There are no control groups, and you will not need to memorize scores of hotkeys. Auralux simplifies in the extreme: one unit, one command, plain primary colors against black.

      The one unit at your command is a simple dot, and the one command they accept is a move command with the cursor. These are produced at capturable, sun-like spheres (your production "buildings") laid out in geometric patterns to comprise a game map. Red, green, and the player's own blue each begin with one sphere on every map. When a dot encounters a dot of another color, they annihilate each other, and if many are tasked on an enemy or untaken sphere, they enter it to capture it. With more spheres, more dots are produced. Capturing all of the spheres on map means victory, and there are 24 maps of varying sizes and difficulties waiting to challenge the player. The complications are few and far between: Sometimes spheres can be upgraded to produce more dots, and the player has the option of playing in a high-speed mode. There also seems to be a population cap on the larger maps; I found this slightly disappointing because it is an extraneous rule in a highly simplified system. It is likely a necessary feature, though, to ensure smooth functioning on a variety of systems.

      It's easy to see how this setup relates to more standard RTSs. Micromanagement has been (mostly) eliminated, so all decisions are of the strategic, macromanagement variety. At any given moment, you have only the decision of where to commit your units. They can stay to defend or move to attack an enemy, or they can be invested in an unoccupied or upgradable sphere. The former actions comprise the combat aspect of RTSs, the latter is a simplified economy. The game becomes interesting because this reductive version of the RTS formula still leads to familiar, if smaller, versions of common RTS tactics, strategies, and battles. It calls attention to the types of balance that make such games successful, and it allows us to see how they function in great detail. For example, difficulty in Auralux is controlled largely by the starting positions of the three colors. In developing a successful tactic for a given map, the player must develop a way to balance the organization of the map before the other colors gain too strong a foothold. It quickly teaches the player the different ways that positioning can affect game balance.

      As much as this minimalist approach amplifies the strengths and subtleties of RTS games, it also, perhaps to Auralux's detriment, highlights weaknesses. People have always railed against strategy game AIs, and nowhere is AI weakness more apparent than here. In a way, perhaps by accident, this is an interesting commentary on the weakness of RTS games compared to other genres. Unfortunately, however, Auralux fails to adequately explore the impacts of AI opponents because it has only one type, and that type is not sophisticated enough. An AI that fails to implement or defend against flanking moves is hardly an AI worth discussing; an attack away from the front lines always catches the AI off-guard here. This severely undermines the emergent complexity of the game, and reduces the game's power to illuminate strategic mechanics. It means that difficulty and complexity of planning are functions only of uneven starting positions and army sizes - important aspects, but rather uninteresting. I would have loved to see some way of tweaking the AIs or even the ability to pit different AIs against each other.

      Auralux has another major omission with regards to opponents. The game is single-player only, despite the importance of multiplayer to games of this ilk. I'm sure this was a practical decision, arising out of the constraints of making a low-budget, one-developer game, but it nonetheless remains a glaring omission. Without it, the game is small, short, and limited in its ability to explore the genre. We can only wonder about the possibilities it could have offered.

      In a way, Auralux thus becomes a puzzle game. With such constancy in the opponents and their flaws, the player must sleuth out the correct set of moves that the AI can't deal with without destroying each other. Perhaps any sufficiently simplified game system becomes something that we would identify with a puzzle game.

      So where does this all leave us with Auralux? There is no doubt that it is a polished game with a very interesting take on familiar mechanics. And I'd be remiss if I failed to mention it's lovely sound design, with spheres pulsing with the music as a battle creates a randomized melody, all emphasizing the rhythm of the gameplay. As a whole, Auralux is enough of an exploration of the genre to leave me interested in further experimentation, but the game's inability to delve deeper because of its lone, primitive AI and lack of multiplayer means that it is inherently flawed. In a game that searches for how complex strategy emerges from simple gameplay rules, an inability to provide an opponent with which to create that complexity is an unforgivable limitation, even given the game's considerable strengths.

      Update: See the comments section of this post for a response from the developer of Auralux.