StarCraft

Front matter

Legacy of the Real-Time Strategy

Simon Dor

LANDMARK VIDEO GAMES

Contents

List of Figures vii

Acknowledgments ix

Introduction 1

Introduction

My experience with StarCraft could be mapped into three periods that represent the history of the game itself. I was introduced to real-time strategy (RTS) games when my cousin brought their copy of Warcraft: Orcs & Humans (Blizzard Entertainment 1994) at my grandparents’ house. My brother and I got the game quickly after and, as most players from the early 1990s, I was playing essentially in single-player mode. I would only have an Internet connection at home in 1999, during my teens, so I have fond memories of the few times I played Warcraft and Warcraft II: Tides of Darkness (Blizzard Entertainment 1995) in multiplayer through modem play, although I would not be remembered for my feats in these endeavors. Playing against humans and playing against computer opponents are two different universes, and I could only have a small glimpse of the former.

When StarCraft came out March 31, 1998, I knew I had to buy the game. Rather than two mirrored factions in a fantasy universe as in its predecessors, StarCraft brought three factions fighting for their survival in a sector of the Milky Way. The Terrans are humans using complex mechanical technologies, exiled from Earth a long time ago, and initially rallied under a Confederation that exploits colonies on different planets. Zergs are creatures taking various insect-like forms morphed from larvae and controlled by central brains, the Cerebrates, under the scrutiny of their Overmind. The Protoss are emaciated warrior aliens gaining their power from psychic capabilities and robotic technologies and regrouped in a caste-based society. These factions (most often called “races” to borrow the term from Warcraft) battle against each

MPARGRArT

EXPANSION SET =— LB 550 Wan j -—

Fig. 1. Main cover art for StarCraft and its expansion, Brood War other and struggle with internal conflicts through the narrative of each campaign. Blizzard Entertainment as a whole team is credited for the “Game Design,” but Chris Metzen and James Phinney are listed as “Lead Designers.” In November 1998, less than a year after the game release, Blizzard launched the Brood War expansion set that would add new game units and a new campaign for each race.!

Something from the game imprinted my imagination. I would design a website on StarCraft at lunchtime in high school, and I was ambitious (or more accurately naive) enough to have a section called “Strategies.” I felt the game was powerful, it unlocked in my mind quite a potential. Pierre Lévy (1994, 93) makes the distinction between “pouvoir” and “puissance” in French, two different words both usually translated in English as “power.” But pouvoir is more closely associated with political institutions and limits possibilities, while puissance (translated as “strength” in Lévy’s works) opens potentials. In that sense, StarCraft felt “strong” for me. It had a campaign filled with an interesting narrative

1. Brood War is often used to refer to StarCraft in general, the “Brood War era” referring to the e-sport period where StarCraft I was not released yet.

Copyright © 1998 Blizzard Entertainment. AN rights reserved) IZIAR Smackèr Video Technology. Copyright © 1994-1998 LAAN

RAD Game Tools, Ino. ENTERTAINMENT

Fig. 2. Original loading screen for StarCraft, showing typical characters respectively for Terrans (a Marine), Protoss (a Zealot), and Zerg (a Hydralisk)

where “decoding” skills, as I will call them here, were central. Yet, multiplayer mode felt like mind games where one had to “foresee” what their opponent is doing to choose their own course of actions. Moreover, my brother and I used the campaign editor that accompanied the game to create our own campaigns and scenarios adapted from our homemade tabletop role-playing games, with approximate English voice acting. I remember playing a few skirmishes against strangers on the Battle.net server, mostly in team fights with friends, but it was not a central aspect of the game for me.

A Narrative Around the Game

Icame back to StarCraftin 2007 when I decided to analyze a video game for my Master’s thesis. I wanted to compare StarCraft and Warcraft II: Reign of Chaos (Blizzard Entertainment 2002) in terms of narration. StarCraft have units called “heroes” who have a name and play a role in the story, and most campaign missions have a “your heroes must survive” winning condition, thus letting the narrative unfold without losing any important characters. Consequently, they are mostly left in the player’s base, as protected

Introduction + 3 as possible. In Warcraft II, heroes have different skillsets than normal units and can be revived if they die; they can be used in offensive strategies without too much risk. In StarCraft, their role is mainly narrative: heroes do not exist in multiplayer games. I quickly realized that, by focusing on the narrative in the game, I was missing the truly interesting narrative around the game.

This period corresponded roughly to a first attempt at internationalizing e-sports.? StarCraft competitions in South Korea were kind of a myth for me and mostly unknown in my social circles. The first tournament I watched was a “GOMTYV Classic” in 2008-2009 on the GOMTV video-on-demand platform, commentated by Nick “Tasteless” Plott, who would become one of the most renowed StarCraft caster, and Daniel “Super Daniel Man” Lee, whom I would later learn had been the manager of the StarCraft World champion Guillaume “Grrrr...” Patry (Eudes 2001, 8). The arrival of YouTube shoutcasters—the word used for e-sports commentators—was a way for e-sports to grow outside of South Korea. In 2007, e-sports “ambassadors” such as Tasteless and Dan “Artosis” Stemkoski moved to South Korea in the hope of launching their career as casters (Asia Pulse News 2012). They were among more amateur YouTubers who casted game tournaments using low-quality Korean footage from a tournament and dubbed their own commentary over the Korean one, widening its accessibility (T. L. Taylor 2018, 147).

A few game scholars were interested by StarCraft before that period, including Florence Chee (2005), Alexander R. Galloway (2007), and Gerald Voorhees (2008). T. L. Taylor was already doing fieldwork during the StarCraft tournament in Seattle at the World Cyber Games 2007 for her upcoming Raising the Stakes (2012). There were many contributions on the game in Gaming Cultures and Place in Asia-Pacific, edited by Larissa Hjorth and Dean Chan (200ga). But even to this day, RTS games in general are only marginally considered in most histories of video games (see Kent 2001, 2021; Wolf 2007; Donovan 2010), with a few exceptions very recently when they are included in e-sports segments (for example, Stanton 2015, 268-75). There was of course a “linguistic” and “cultural” gap between Korean e-sport and mostly English-speaking researchers, but this gap was redoubled by a “competitive” gap: it is not easy to know

2. Although e-sports are more officially composed as “esports” following the recommendations of the Associated Press Stylebook (Witkowski 2022), I will use “e-sports” throughout the book to respect the original wording of the StarCraft communities.

what happens on the screen and why it happens in any strategy game, yet alone one that is played at a professional level. Novice players were not commonplace on online servers 10 years after the game released; trying to win on the competitive Brood War third-party server iCCup was not an easy task, and I cannot say that I succeeded.

In that sense, the release of StarCraft I: Wings of Liberty (Blizzard Entertainment 2010) was an occasion for the rest of the world to jump on the RTS e-sports boat. Rather than having a fear of missing out of the first iteration, a lot of new fans could come and appreciate competitive strategy gaming. In 2011, I attended the first “Barcraft” event in Montreal: a night club opened its doors on a Sunday morning until the night to stream live on a giant screen the 2011 MLG Pro Circuit Orlando. It felt very strange: for the first time, I was not alone enjoying other people playing a video game. Since tournaments at that time were mostly streamed behind a paywall,* being there to see matches live and with friends and strangers was quite an event. The low-quality YouTube videos from a few years before seemed like ancient times. I watched live the victory of the Canadian Protoss player Chris “HuK” Loranger against Korean Protoss player Jang “MC” Min-chul, and the crowd cheered as they would in any sports bar while confetti dropped from the ceiling. The StarCraft II area would be a period of normalization for e-sports outside of South Korea.

Seeking Another Meaning

I briefly came back to playing StarCraft more seriously in 2017, when StarCraft: Remastered came out. The Remastered edition supports wide screen resolutions, includes more detailed 2D graphics and sound, restores a ladder function, and adds a matchmaking feature. The original game and its expansion are free-to-play since this Remastered edition.

StarCraft I had been released in 2010, and had two expansions: Heart of the Swarm, in 2013, and Legacy of the Void, in 2015. StarCraft IT changed drastically throughout the years: the expansions and major patches not only added new units, but also removed some important ones and changed the overall timings, thus altering the gameplay irremediably.

3. T. L. Taylor notes that MLG had a “pay per view” business model which they eventually dropped (T. L. Taylor 2018, 180). Some of the negative comments from gaming community are gathered in forum messages (see xrapture 2012).

Introduction + 5

StarCraft: Remastered, on the other hand, was meant to be as similar as Brood Waras it could. To quote Grand Davies, senior software engineer at Blizzard Entertainment, the developer’s “top priority was not to change the game but to keep it exactly as it always has been” (quoted in Korea Times 2017, my emphasis). Although it might make sense in terms of software engineering alone, this quote is the exact opposite of what I argue in this book: StarCraft has not always been what it is today. If StarCraft IT was arguably the first strategy game designed to be an e-sport, StarCraft is the first that became an e-sport. It has clearly not always been one.

It is a common bias to interpret a video game as a set of game rules implemented in hardware. Most game scholars agree that the cultural components of playing games are a fundamental part of them. Of course, it does not mean that players are free from any constraints when playing. Maude Bonenfant (2015, 39-40) explains that even if one could freely play without any constraints whatsoever by the game rules, they would still be constrainted by discourses, and by their own knowledge about games. Players’s “freedom” is still limited by one’s prejudices and cognitive schemata, and the discourses in which play makes sense. Play has conditioned our automated perception; the way signs are organized in a game gives a certain meaning, “which has as consequence that the player does not ‘seek for’ another possible meaning” (Bonenfant 2015, 42, my translation). StarCraft has been so closely associated with competitive play and Korean e-sports over the years that it is difficult to seek for another meaning; my goal is to do exactly that.

The influence of StarCraft on PC gaming, e-sports, and strategy games is undeniable. However, a game is not a monolith that delivers its secrets once we interpret it. Playing it means being part of a culturally embedded activity that is more complex than the cultural context of its production. As Melanie Swalwell warns us, as historians, we must go beyond the original experience of a game (2017, 220).* To some extent, the goal of this book is to follow Stephanie Boluk and Patrick LeMieux’s definition of “metagame,” which is “not the history of the game, but the history of play” (2017, 17). Rather than focusing on how the history of StarCraft was inevitably going to produce the first e-sport—which is of course not

4. The Remastered Edition of the game changed some things in the original engine that are difficult to keep track of, including effects like an “SD Graphics Filter” and “Real-Time Lighting” in the options. While I turned the “HD graphics” off to take screenshots for this book to get a better sense of what the game was like in 1998, some new functionalities might be shown in the UI or visual effects turned on.

true—I will trace back the diversity of usages of the game in gaming culture. Through a great timing and especially through a strong dedication and key decisions by its creators, its fans, and corporations seeking an opportunity, StarCraft shifted from a conservative but refined RTS game of the 1990s to an innovative e-sport in the 2000s. As we will see in the first chapter, it shifted from a decoding game to a foreseeing game. It is a landmark video game not for the history of its design, but for the history of its gameplay.

I cannot hide that I am a fan of StarCraft; I cannot see any reason why someone would want to write a whole book on a single video game without a certain dedication to it. The time required to play and understand a video game would be difficult to invest without having a genuine interest in play, even if it breaks the expectations that scientific investigation necessitates a pure “objectivity.” As Christian McCrea describes while writing on the same game, “a pretense of such objective distance is impossible” (2009, 179). At the same time, I will often refer to historical sources created by fans throughout this bobok—it has become a standard in game studies (Guins 2014, 85). While their view on the game might be “biased” by their love of it, most of gameplay history would be lost without their contribution.

Overview of Each Chapter

As we will see in the first chapter, StarCraft contributed its rock to the pyramid of what I call the decoding paradigm, while establishing the core fundations of the foreseeing paradigm in strategy games. The decoding paradigm has existed in strategy video game for a long time and implies the existence of artificial opponents that the player must think as game obstacles to be understood. Foreseeing, on the other hand, is similar to most board games where every player has predefined sets of tools and must anticipate everyone’s actions. These paradigms are fundamentally different ways of playing but coexist in most RTS games.

In that sense, StarCraft has a strange relationship with innovation and conservatism. Chapter 2 will tackle this paradox by underlining how the game is an intensification of common RTS tropes. Its narrative is not especially original, using science fiction stereotypes and usual mission briefings, while adding game characters with whom the player interacts and some narrative sequences during missions. It is conservative in terms of representation, following sexist and ethnocentric tropes, and it refused the “innovation” that 3D graphics represented. Its intensification

Introduction + 7 can really be felt by the interface and controls, which favors the swifter rather than the overthinker.

Chapter 3 will put in context StarCrafťs gameplay through gaming conventions and its technological conjuncture. The idea is to under stand the context from which a diverse set of practices could propagate, including but not exclusively e-sports. The problem of balancing a game through game patches will be explained, especially since multiplayer is renown for its infamous “Zerg rushes.” The material infrastructure necessary to play online—Battle.net—and its consequences on gameplay will be covered. In the end, the goal will be to underline how the conjuncture was perfect to foster foreseeing play.

Chapter 4 will cover how the sociocultural context of South Korea was a key part of how e-sports emerged. I will explain the role of governmental initiatives in spreading broadband connections and the importance of PC Bangs in changing the image of gaming in culture. I will also describe the golden age of the game and the immense social and cultural phenomenon that progaming was during the 2000s in South Korea. Ultimately, I will explain how Korean e-sport is structured around foreseeing.

The last chapter will cover the extent to which StarCraftwould expand its scope through the campaign editor. As with most games with an editor, players create their own maps to fight with their friends in their own custom skirmishes. But the editor went far beyond that: new game rules could be implemented and mini-games quickly emerged through this. While Blizzard gave the tools for players to experiment in their engine, players needed their own tools to extend this experience. That is how game hacks became common in the community, to the point where they are normalized and essential to organized progaming.

Each chapter will unfold one specific aspect that makes StarCraft more than a list of game rules, but a game that has an history of its own—which could, arguably, be true of each game. StarCraft will help us show the legacy of RTS games in gaming culture. First, let us explore how two different histories found a point of convergence in RTS games, and how StarCraft became the core example of this convergence. As we will see in this first chapter, most RTS games that emerged in the 1990s are both decoding and foreseeing games.

Chapter 1

... we looked back at our previous games and realized that our solo campaigns have never prepared anybody for an online experience at all. That never worked, right? We always sort of touted it that way — “It’s going to prepare you” — but it never really did. — DUSTIN BROWDER, LEAD DESIGNER OF STARCRAFT IT (QUOTED IN REMO 2009)

Two Conceptions of Game Time

If competitive play often leads to epic confrontations on the battlefield, sometimes the fight is quite short. During the second iteration of TeamLiquid StarLeague (TSL) in 2009-2010, the American Terran player Gregory “IdrA” Fields was matched against the American Protoss player Tyler “NonY” Wasieleski (nevake 2010a). IdrA had won the first game and started the second one by building a relatively fast second Command Center with one of its SCV—Terran’s worker unit—to have an economic advantage. Each player usually sends one of their first worker unit to scout, and that is what NonY did. But this worker unit also frequently annoys their opponent by attacking their counterparts while they perform their normal tasks—a tactic called “harassing.”

And thus, NonY harassed IdrA’s SCV building its Command Center to try to disrupt his attention. IdrA decided to switch the SCV building the structure: the normal maneuvre would be to click on the SCV performing the task, and to click ESC, then the next SCV could take when the first left. Unfortunately, IdrA misclicked: he clicked on the Command Center rather than on the SCV before hitting ESC, and thus cancelled the construction. He lost 100 minerals, but more importantly lost precious game time since he would need to start over the construction. To the surprise of the casters, IdrA then precipitously but irrevocably typed “go” (for “good game”), losing the second game of the match.

While in a lot of sports, every second counts and can be the one where a goal is scored, in StarCraft competitive games the first seconds of the game have repercussions on every second of the rest of the game, since its game economy is based on growth. StarCraft has a strong positive feedback loop regarding time: the beginning of the game is crucial to determine the endgame. If the experience of video games tends to be expressed in terms of space, competitive real-time strategy games should definitely—as their name suggests—be considered first and foremost as temporal experiences. The value of the first seconds is higher than that of the last seconds of a game.

This conclusion might seem counterintuitive, since a player can have a lot of time to recover from an earlier mistake. It is also not true of every StarCraft experience; it is mostly true in competitive or skirmish games. It is what makes the difference between what I will call the decoding paradigm and the foreseeing paradigm.’ The information management of StarCraft is wholly different when playing a solo battle or in the campaign—in the decoding paradigm—or when playing an online battle—in the /foreseeing paradigm. As we will see, each of these paradigms have a history of their own.

The very definition of strategy is different depending on the paradigm considered. The decoding paradigm is when the player must detect tendencies in the actions of the artificial intelligence opponent (called “computer” in the game) to anticipate future actions. They cannot be certain of the exact series of actions that will unfold or be possible; often, one can be sure that some actions, while theoretically possible, will not take place since it is not a tendency for the AI to do so. The decoding paradigm is the player “guessing” what the “rules” of the games are.

Most contemporary multiplayer strategy games rather fall under the foreseeing paradigm. Game actions are literally “foreseeable.” Every possible action in the game rules has its own prerequisites and players can easily manage to either read about them beforehand in the game manual

1. I used to call these two paradigms “decryption” and “prediction” to translate the French words “décryption” and “prévision” (see Dor 20144). But I sense these words do not induce any clarity. I must thank Bernard Perron for suggesting the new translations and using them in his work (Perron 2018, 113).

or try them ina custom game. Just as every player knows that a poker game has four Queens, they know that training Arbiters implies the construction of a Citadel of Adun, a Templar Archives, and an Arbiter Tribunal, and thus can anticipate Arbiters if they scout those buildings. The foreseeing paradigm starts when the player knows the rules of the game.

Real-time strategy (RTS) games before 1998 were often games solely inscribed in the decoding paradigm, and their legacy lived on in StarCraft. By being an accessible and competitive multiplayer game, StarCraft also built on games from the foreseeing paradigm: you can foresee your opponent’s actions by knowing their gaming possibilities and reading their strategy. Both these paradigms are important when considering the historical importance of StarCraft. The goal here is thus to see how the emergence of the foreseeing paradigm in the 1990s reaches a landmark with StarCraft. This chapter describes the historical background from which both paradigms of RTS gaming emerged.

The first section will describe the core rules of StarCraft, applying to both paradigms. The second section will describe the history of decoding, showing how Westwood Studios’ games since Dune II: The Building of a Dynasty (Westwood Studios 1992) have influenced the RTS genre. The third section will show, on the other hand, how multiplayer games such as those from Ozark Softscape played a significant role in the history of foreseeing, until modem play became more widespread. I will argue that StarCraft in the history of gaming is the main point of convergence between the decoding and foreseeing paradigms.

A Classical RTS

StarCraft is a classical RTS game and a common point of comparison for other games in the genre. The player must collect two types of resources (minerals and vespene gas), create and manage buildings and units to destroy every building of their opponents. The player clicks on their units to give them specific orders (move, attack, patrol, etc.) or use their special abilities.

Each unit is a “type” following what game designers Andrew Rollings and Ernest Adams called a tile-based aesthetics (2003, 340): 12 Marines will be represented as clones, with the same figure on the interface, the same voice when responding to orders, the same morphology on the game space. This aesthetics makes the game space easier to understand cognitively and strategically: the same unit types will have the exact same properties. Players also have an attributed color so that their units and buildings can be easily recognized. The game is set on a pre-created map that determines the topographical elements of the game: starting and resource locations, bridges, cliffs, ramps, etc. It is a quintessential example of why Pascal Garandel argues that video game space is first and foremost a means to an end, almost every element of it being geared towards play (2012, 131).

Each unit, building, technology, or upgrade costs a certain number of resources—minerals and gas—spent once. Worker units collect resources by moving back and forth from the main buildings to mineral patches and vespene geysers. By training more workers, the collection process will be faster and, eventually, the player needs to build a new main building near another resource location: an expansion.

Buildings have different functions: to collect resources, to train military units, to research upgrades and technologies, or to fight. For every faction, buildings are organized in a technology tree: creating a building will unlock other buildings, but also units, upgrades, and technologies. For instance, if a Zerg player wants to build Mutalisks, they will need a Spawning Pool, then to upgrade a Hatchery to a Lair, which will let them build a Spire needed for this flying unit (Fig. 3).

IQ [9 ie i i a ie

m N N —_

Fig. 3. Zerg building and unit dependencies, part of their “technology tree,” as annexed in the game box

StarCraft is a “real-time” strategy game since players perform their actions simultaneously. The units and buildings respond to player-issued orders through mouse clicks or hotkeys; the action then takes a certain time and the unit or the building accomplishes it autonomously. The player can then manage something else during that time. They must thus use their time efficiently, trying to time each action without having to wait for a certain task to finish. Units are either ground or air units. Their weapons can be melee or ranged, and attack ground, air, or both unit types.

The actions of the opponents are hidden from each player; they must explore and periodically scout their opponents’ bases to know what they are doing and where their units are. StarCraft uses the same kind of fog of war as popularized by Warcraft I. The game space is initially hidden to the player; they must scout to discover the topological elements of the map. Friendly units must be around enemy units to reveal them; otherwise, they remain hidden (in the “fog of war”) (Fig. 4).

The Decoding Paradigm

The StarCraft campaigns are a series of levels (or “maps”) to complete linearly. Each map is contextualized narratively and gives a set of predefined units, buildings, and topography. The player must fulfill specific objectives, stated in a mission briefing: in the “Norad II” scenario, they must bring a hero and two dropships at the site of a crash to rescue the

» 6 survivors. In other scenarios, they must “survive 30 minutes,” “protect a building,” or simply “destroy all enemy buildings.” Rather than having to create their buildings and units, computer players already have their own bases and will mainly react to the player’s actions and execute prescripted moves. In the “Norad II” example, the Zerg computer player already has plenty of Spore and Sunken colonies protecting the site of the crash.

The campaign is a clear and strong example of the decoding paradigm: the goal is to decrypt the opponent’s patterns or scripts and respond to these actions; the player “decodes” patterns in a figurative sense. The interest of decoding is to offer a challenge to a single player. They must somehow find an efficient strategy to overcome each obstacle or each map one after the other. In most maps, the equilibrium to reach is between having an efficient resource-collecting flow and defensive or versatile military units.

Fig. 4. The “fog of war” principle. Top, a hidden location. Middle, a location actively revealed by friendly units. Bottom, a previously explored location without active friendly units

In “Norad II,” the player’s buildings are already under attack, and they must be defended. Once the first blitz is countered with our initial military units, as Bart Farkas underlines, “there’s only a short period of time for you to get your defenses back up before the next onslaught” (Farkas 1998a, 102), while “six to ten SCVs to start gathering resources” should be enough (103). Building a too large amount of worker units could lead to insufficient defense.

In the decoding paradigm, the opponent acts in a precise way “encoded” in the game. In a successful game design, the player can anticipate their opponent and respond efficiently; this optimization characterizes this paradigm. By trial and error, or by intuition or habits, the player will know what to do and not to do. But in some cases it is not predictable at all:

Whatever you do, don’t attempt to build any units near Norad II, and don’t attempt to launch any attacks from this position. If you do, you ll bring the wrath of the Zerg down upon you, and the mission will be over.

Of course, there is no narrative nor strategic reason for the Zerg to not attack the crash site immediately. When the opponent starts with four bases and an army size at maximal capacity, they would quickly win, but the game would be too hard: the computer will never do that. The rules by which the AI plays must be decrypted.

In the decoding paradigm, a player never “knows” what an opponent can do; they must guess what it will do based on their previous experience of the game. Using the terminology of the pragmatic philosopher Charles S. Peirce, a player will forge a habit of mind, “[t]hat which determines [them], from given premisses, to draw one inference rather than another” ([{1877] 1991, 147). In strategy video games, one could talk of strategic habits: players forge inferences by observing their opponents acting or reacting similarly under similar circumstances.

In the decoding paradigm, these strategic habits are forged through experience with anterior enemy actions. For instance, players will know how to counter Hydralisks or Mutalisks, provided they have seen them before. They could also forge more general patterns of mind, observing for example that AI units tend to attack in small squads on base entrances. The player will play similarly when similar circumstances happen again: they are perceived as specific occurrences of a same phenomenon. In a case when they failed to predict the algorithm, or the opponent unleashed an attack larger or different than expected, they can simply reload the game from a previous state and try again to refine their habits by trial and error. Strategic habits of the player (how they will act predictably) are forged according to strategic habits of the AlI—what players observed and how they forged inferences.