Why Dota Sucks — 2. The Crude Birth of a Genre

Synopsis: In order to understand dota, you must first look at the circumstance of dota. You must look at the evolution of the genre within the Blizzard Entertainment real-time strategy ecosystem, through the third-party content creation for Warcraft II, StarCraft, and Warcraft III. In giving this matter extended scrutiny, you will find that the dota genre was not the matter of design genius, but of circumstance and impediment. Unlike the commercial game designers who have the talent and resources to shape the game experience, every Blizzard game offered a new set of restrictions for amateur content creators. Bound by these limitations, content creators used RTS game engines and RTS map editors to build a character-action game. You will find that these limitations—and the flaws that came with those limitations—have shaped the mechanical, aesthetic, philosophical, and economic conventions of its genre. This book will highlight those limitations as it is necessary to do so.


We should now discuss the rise of Defense of the Ancients as it emerged from the Blizzard Entertainment real-time strategy ecosystem. Now, why is that? While videogames are collaborative works that showcase the creativity of highly talented individuals, I believe that the history of videogames has been largely written by economics and circumstance. Videogames are art, but it’s a business of art. The great videogame designers have found the means to mold the business of videogames to their own interests and talents, whether they’re catering to the demands of the corporate board, working with others at an independent game developer, or creating free content for a commercial game engine in their spare time. This does not mean you cannot learn games through the simple act of playing them, but circumstance is a valuable tool for confirming your instincts, and the circumstance of dota is no exception.

A couple of unrelated events would usher in the popularity of third-party content for use with Blizzard software. Many games featuring elements we associate with RTS would find reverence and acclaim from their contemporary audiences.1 But with 1992’s Dune II: The Building of a Dynasty, Westwood Studios would codify the archetype for the RTS genre—build a base, build an army, destroy the other base—that is now commonly thought of as RTS.2 The following year, id Software would release Doom, and with help from the continuing and emerging adoption of the internet, the game would launch a voracious consumer appetite for player-created content. People wanted to make their own maps, and by using the internet, they could get complete strangers to play and enjoy them. In November of 1994, Blizzard Entertainment would launch their own RTS flavor with Warcraft: Orcs and Humans, setting the stage for what was to come.

While turn-based strategy games had already allowed players to customize, randomize, and create their own battlefields, it seemed fairly inevitable that RTS would use “Create your own maps!” as a selling point. And when Blizzard released their 1995 RTS classic Warcraft II: Tides of Darkness, they would provide players a functional, intuitive map editor. The nicest thing to be said about the editor is that it provided the user-friendly approach that would come to define the later editors for Blizzard RTS games. But when we compare it with the tools used to create content for today’s games, it was largely deficient. Players were given no access to the mission briefing system, the maps could not be linked together as part of a campaign, and those maps could not feature victory conditions beyond “kill everything”. In-effect, maps existed as individual entities within their own contained ecosystem and had little flavor beyond the arrangement of the units and terrain.

Third-party programs such as War2xEd and PUDDraft would give players more control over the design of their maps—allowing players to make rudimentary customizations for units and AI—and Wardraft would even allow players to customize and modify the game files. (In spite of this, “total conversions” would never catch on.3) Warcraft II ultimately lacked support for the scripting or programming that would allow units and buildings to act outside of their parameters.4 For example, a scenario where friendly knights run towards your base to “warn you of the impending danger” was simply impossible.5 These limitations would stifle the growth of what we now think of as “custom maps”, the maps which attempted to take Blizzard RTS games in unconventional directions.6 Puzzle maps and so-called “Olympic” maps usually asked the player to take a small number of units and get them to a destination on the map. But for the most part, these maps were simply exaggerations of the Warcraft II mechanics, instead of anything that players would consider its own “game”. Skirmish and singleplayer maps would dominate the landscape for content creation in Warcraft II, and the custom map would remain an unknown quantity.

Blizzard would provide the leap forward when they released their Campaign Editor alongside StarCraft in 1998. The new scripting techniques were demonstrated in the famous StarCraft singleplayer campaign, which featured a greater focus on narrative and character development than the genre had previously shown. The Campaign Editor gave players access to the scripting system that built the campaign, and with it, a flexible entry gate for anyone who wanted to scratch the game development itch. From there, the integration of StarCraft into Blizzard’s Battle.net online service—launched for use with 1996’s Diablo—assured mapmakers a larger audience than the trove of third-party tools and web sites.7 It is during this time that content creation would formally split into two separate affairs, with skirmish and custom maps being thought of as separate entities.

From this circumstance, the dota genre would make its entrance in the StarCraft custom map Aeon of Strife, as created by the mapmaker Aeon64.8 The map allowed players to select a powerful character—one individual in a team of players—to compete against an onslaught of computer-controlled enemies. While Aeon of Strife was conceived in an RTS map editor, its emphasis on powerful protagonists, large waves of fodder enemies, and straight-forward level design was less like StarCraft and more like the “lawnmower”9 hack-and-slash games (Dynasty Warriors, OneeChanbara, Ninety-Nine Nights) that would appear in the following decade.

As one small part of the StarCraft custom game scene, Aeon of Strife was a change of pace from the conventional game modes. StarCraft may be a classic within its genre, but like most early RTS games, it has lacked a genuine sense of scope and scale in war. As it has been described in third-party accounts, the gigantic wars on display in Aeon of Strife provided this sense of war for StarCraft players.10 Because of this, it would become a regular part of the custom game scene. Countless versions and variants of the map would appear as created by various mapmakers, even as Aeon64 was nowhere to be seen.11 And many of the concepts that define Defense of the Ancients would emerge in this era, including defensive control points and the sprawling, winding map layouts associated with the genre. Most crucially, some mapmakers would turn Aeon of Strife into a versus multiplayer game, with players attempting to drive their computer-controlled fodder against a competing army and their leaders.12

But even if the StarCraft custom game scene was thriving and popular, the maps remained incredibly basic. They co-opted the StarCraft art assets, game template, and user interface because they had to. The selectable characters in Aeon of Strife were simply powerful versions of the units that could already be used in the skirmish game modes.13 If you were a mapmaker and you wanted to make a new character, you gave an existing unit a cool name, increased their health, upped their damage output, and that was that. Those characters were simply more powerful clones of an existing unit.14 Meanwhile, total conversions of StarCraft faced the same issues as the ones built for Warcraft II.15 And when compared to the content creation for other games—including Doom, Total Annihilation, Quake, and Half-LifeStarCraft custom maps were limited in both scope and quality.

In spite of this, StarCraft custom maps became an incredibly valuable commodity. You see, the videogame publishing model was built on the same contract as the movie industry. Basically, you use the trash to fund the good stuff. Discerning videogame players may despise the “brodude” who buys FIFA every year, but the money he spends on Electronic Arts‘ sports games can be used to finance a Mass Effect, a Dead Space, or a Titanfall. However, such a contract can even exist within a single game, where you offer a wide range of game modes for a a wide range of players. The players who enjoyed StarCraft custom maps could act as a social and economic backbone for the smaller percentage of users who pursued the game as a top-notch RTS experience. As RTS games like Warcraft II, Total Annihilation, Command and Conquer: Red Alert, and Age of Empires II: The Age of Kings were quickly left to their die-hards, StarCraft maintained incredible longevity with the help of the third-party content.16 In the long run, it was worth every dime of Blizzard’s attention to encourage the creation of the custom maps.

The wild success of StarCraft (and 2000’s Diablo II) paid off handsomely for Blizzard. By the time Warcraft III: Reign of Chaos was marked for its July of 2002 release date, it was the most anticipated videogame in the history of the computer market. In spite of this, the end result was not a revolution. The game would be centered around small-scale tactics and it would share much in common with 1997’s Myth: The Fallen Lords and 2000’s Sacrifice. Warcraft III‘s defining concept was the powerful heroes that could level up, acquire items, and unlock powerful abilities over the course of a match. But unlike the previous Blizzard RTS games, the emphasis on base-building and fast economic management was brought to a minimum. Fans of Warcraft II and the other RTS games based on the Warcraft model had come to expect fast action and an emphasis on economic supremacy. And the focus on slower, more localized combat—with game-changing heroes at the center of the battle—would make Warcraft III one of the most polarizing games ever made.

On the other end, custom map creators fell in love with everything that the Warcraft III World Editor had to offer. They took advantage of the new maximum player limit, which bumped StarCraft‘s eight-player limit up to twelve. Since computer-controlled players are often used to fulfill essential map functions, the custom maps for Warcraft III were commonly contested on a five-on-five basis, providing the team sizes typically associated with the dota genre.17 On top of this, the editor allowed native access to “Just Another Scripting Syntax”, the scripting language that provides the backbone for Warcraft III maps. Players now had the power to program their game map, so long as the code functioned within the parameters outlined by Blizzard.

The most significant leap would come in the amount of content was now available to mapmakers. When Warcraft III was originally announced, Blizzard marketed the game under the moniker of “role-playing strategy”.18 And whether a narrative-driven singleplayer map or a versus multiplayer match, the player could slaughter “neutral” foes in order to gain resources (gold, items, experience) for their war effort. In the pursuit of this, Blizzard added hundreds of units and buildings that comprised every primary and secondary faction within the world of Azeroth. Where Blizzard custom map creators were once confined to the small range of units that were needed to flesh out the game’s playable factions, they could now choose from an entire world of creatures and characters. All of the available unit models, down to the most pathetic slime or troll, were given as much detail and attention as a Thrall or a Jaina Proudmoore.19 These creatures and buildings could be thoroughly customized for use in the new maps.

With little hesitation, players imported their favorite StarCraft custom maps into the new game engine. The premise of Warcraft III would prove a perfect match for Aeon of Strife, the story of powerful badasses going to battle with a sea of mooks, and the map was quickly ported over. However, the introduction of the hero model and the lower unit counts altered the appeal of the map.20 Much like the “Hero Arena” maps that dominated the early custom game scene in Warcraft III, Aeon of Strife was taking the Warcraft III hero model—a small portion of that game’s construct—and turning it into its own game. Aeon of Strife was giving players a single character that could learn spells, gain items, and level up, all performed through a point-and-click user interface. Very often, these maps thrived on a novelty appeal, with heroes and characters that could grow to comical levels of strength. And if this all sounds like it has some crossover appeal with the Diablo series, that’s because it absolutely does.

However, most of these maps would remain a novelty experience because the vaunted World Editor was limited in places that it needed to be flexible. The editor allowed mapmakers to customize nearly every aspect of a unit, but in a game where units are commonly defined by their range of spells and abilities, you could not customize the individual spells and abilities that were being used by those units. They were exact copies of the ones available for use in the skirmish and campaign modes.

Third-party programs were once again created in order to overcome these limitations, and it was a battle waged with limited success. During this period, the first Warcraft III maps with custom spells and abilities would emerge, and they were prominently featured in maps like Valley of Dissent and Keys of Sealing. In the midst of this development, a player by the name Eul created something called Defense of the Ancients. The early versions of DotA were grossly limited in quality, even in regard to basics like terrain design and unit placement. However, the map featured all the trappings we associate with today’s dota games: A large roster of playable characters, two bases connected by a three-lane map design, neutral monsters strewn about the map, and “hidden” merchants who sold players the most powerful items. But Battle.net content creation was a largely collaborative effort, with players borrowing and often stealing from each other. The lack of standardization for the third-party programs limited their reach and limited the progress of the dota model.

The community would get a big surprise when the Warcraft III expansion pack The Frozen Throne arrived in July of 2003. The upgraded World Editor offered access to every aspect of the Warcraft III game model, allowing players to create their own spells and even modify the base game rules. Nearly all of the third-party programs that had been created during the Reign of Chaos days were now obsolete. With this powerful editor, players could now do anything they wanted, so long as it functioned within the Warcraft III game engine, its interface, and its scripting system. With these powerful tools now standard across the entire user base, content creators had ease of ability to create and disseminate the concepts that we now see as synonymous with Defense of the Ancients. And just like Aeon of Strife, Defense of the Ancients had become a concept and a brand, and numerous players would create their own versions of Defense of the Ancients.

Early in the life of The Frozen Throne, a pair of content creators by the name of Meian and Ragn0r would create their own variant of DotA. They decided to take interesting and well-designed heroes in other versions of the map and place them under a single banner.21 Much as the best professional athletes compete in All-Star games, the map would be known as DotA All-Stars. However, the pair gave up on the project shortly after its launch and left it to the dustbin. Another mapmaker by the name of Steve “Guinsoo” Feak saw potential in the map and took over the project. His chief contribution was to facilitate interaction between the human participants, distinguishing it from versions of Defense of the Ancients which emphasized the destruction of the disposable soldiers.22 And through a long line of revisions, DotA All-Stars would acquire a reputation for its quality. Fans were now eager to contribute, and a team of mapmakers would work with Guinsoo in order to update the map.23

By 2004, serious multiplayer competition began to emerge around DotA All-Stars. To help organize the competition, Guinsoo’s compatriots in Clan TDA created a system of bots that was used to enforce conduct in DotA All-Stars matches played throughout Battle.net.24 TDA member Steve “Pendragon” Mescon would also create a community portal for DotA All-Stars and it would quickly grow in popularity. (At its high point, dota-allstars.com would have over a million registered users.25) By 2005, there was hardly a Warcraft III player on Battle.net who had not heard of the map, and many players did not take kindly to its existence. It was during this time that Guinsoo would hand off DotA All-Stars to fellow creators Neichus and IceFrog, a duo that possessed the programming skills necessary to continue increasing the complexity of the map.26

If the story had ended there, people wouldn’t be writing books about the topic. But at this point, even Blizzard had taken notice of the map, and the company organized a DotA All-Stars tournament for their first BlizzCon convention in October of 2005. (Blizzard’s support for the map would otherwise remain limited.) By 2006, other large tournaments were being organized and the best dota players were participating in them. DotA All-Stars had even begun to earn the ire of other custom game creators and players, because All-Stars was making it difficult for competing maps to get any exposure.27 And as the Warcraft III skirmish modes were being left to their most committed players, it finally dawned that DotA All-Stars was generating more interest—and acclaim—than Warcraft III itself. DotA All-Stars was becoming ubiquitous with Warcraft III in the way that Counter-Strike would become ubiquitous with Half-Life. This version of Defense of the Ancients would become such a success that competing versions of the map would be left to history, and the lineage of maps directly leading into DotA All-Stars is now singularly identified as “Defense of the Ancients“.

The map would have a profound impact on nearly every party associated with it. In 2006, a Singapore software company would launch “GGClient”—later renamed to Garena—and the program would be used throughout Southeast Asia to play Defense of the Ancients en masse with unauthorized copies of Warcraft III.28 The threat of piracy would become a major factor in Blizzard’s decision to launch 2010’s StarCraft II: Wings of Liberty without any formal support for offline play,29 and the new version of Battle.net allowed the company to remove a map from the service if it did not meet certain criteria for ethics and copyrighted material.30 31 Along with a less intuitive map editor, the StarCraft II custom scene would never enjoy the popularity and success of its predecessors. Interest in the game quickly waned without a robust custom scene to act as a backbone for the die-hards.

In response to this reality, Blizzard unveiled “Blizzard DOTA” at BlizzCon in October of 2010. But at this point, Guinsoo and Pendragon had taken up residence at Riot Games. Meanwhile, IceFrog was allegedly32 working with S2 Games on Heroes of Newerth and then later with Valve on Dota 2. In August of 2011, Valve would officially unveil that sequel, and both Eul and IceFrog were present on the development team. Later on, Blizzard would scuffle with Valve over the “DOTA” trademark, and Blizzard would rename their map to Blizzard All-Stars. After years of what an outsider would see as “development hell”, the map is now a commercial project known as Heroes of the Storm, which has an expected 2015 release date. League of Legends and Dota 2 have become two of the most popular games in the world, dota is now the hottest genre in videogame development, the market for RTS games has come to a crawl, and Blizzard is now looking from the outside-in on the commercially-lucrative genre that was created with their own editing tools.

Defense of the Ancients became so popular that it would not only come to dominate the game it was created in, but have a profound impact on the companies and individuals involved in its success. And yet, for all of this success, we need to maintain some perspective. Just as the puzzle maps were an alteration of the concepts in Warcraft II, and just as the early tower defense maps were an alteration of the concepts in StarCraft, Defense of the Ancients was an alteration of the concepts in Warcraft III. The Frozen Throne‘s World Editor may have been powerful, but it was not a programming language and it was not a videogame creation tool. It was a mapmaking tool, designed to make maps with the content and game rules outlined by Blizzard. It is crucial to understand this. When players manipulated the Warcraft III game engine towards more powerful ends, problems began to show.33

As we have already mentioned, modifications of the game and the game engine could not be disseminated on Battle.net. This meant that most custom maps used the same interface, control scheme, and game concepts in Warcraft III. Most damningly, the Warcraft III game engine does not provide native access for the “mouselook”34 mechanics now synonymous with first- and third-person action games on personal computers.35 But in addition, the peer-to-peer networking technology used for online play in most RTS games is designed for a game model with a large number of moving parts. Action games with fewer moving parts use a client-side prediction model to minimize the effects of networking latency, and not only could Battle.net not provide this benefit, the service (as designed for Warcraft III) uses a quarter-second hard-coded internet delay. This was intended to prevent players from gaining a substantial advantage with a better internet connection, but it discouraged the adoption of the concepts that are a mainstay in other excellent action games. And while the Warcraft III game engine has hosted concepts ranging from first-person shooters to Diablo-style action games, custom keyboard layouts (and the maps that used them) never caught on because they generate substantial input latency.36

So hopefully, you understand where I am going with this. It didn’t matter which game it was: Warcraft II, StarCraft, Warcraft III. Blizzard mapmakers were influenced by incredible restrictions at every step of the way, and not the kind we typically associate with the design of videogames, where designers attempted to overcome the technical limitations presented by limited hardware. Things that were common in other genres were impossible to distribute through Battle.net and impossible to perform in the various Blizzard RTS game engines. The result is that dota is a rarity and oddball within videogames. It is one of only two genres whose rules were foremost defined by software limitations—the limitations of Blizzard game engines and their content creation tools—instead of hardware limitations.37 Every time that Blizzard custom maps made their grand leap, there was simply another set of barriers that got in the way of “building the best game possible”.

What this should demonstrate is that dota was not a matter of genius and great design, but of circumstance and impediment. When developers were creating the blueprints for their genres—id Software and the FPS, Capcom and the fighting game, Westwood and the RTS—those companies had minimal notion of how these genres should play and what they should look like. But more importantly, they had the freedom, creative talent, and resources to define the experience. This is not to claim that all great videogames, whether amateur or commercial, are built on a clean canvas. But the games that thrived on borrowed assets were created by one of two groups: One, the designers whose vision was not compromised by the software limitations, where amateur projects like Team Fortress were built into a game engine specifically tailored to their genre. Or two, the companies with the resources to make those borrowed assets work. For instance, taking the Unreal 3 game engine—designed for use with FPS games and long the enemy of colorful art design—and turning it into a visually-stunning platforming game like Mirror’s Edge.

After spending a pathetic amount of time trying to understand videogames, there is one thing I can say with certainty: Great videogames are no fluke. You don’t just fuck around in a game editor for a couple thousand hours and end up with something worth playing. As manifested today and through significant revision, the history of the dota genre is the history of amateur game designers using their limited resources to manipulate a game engine which was never designed for formal modification, never designed for traditional action games, and was demonstrably terrible at doing many of the things that were commonplace in videogames. Or, in other words, the creators of dota used an RTS game engine, the RTS concepts laid out in Warcraft III, and the Warcraft III World Editor to create a character-action game. The same game model used to create Warcraft, Age of Empires, Command and Conquer, and Total Annihilation was used to create something that shares more in common with the Diablo and Dynasty Warriors series. And in order to distribute this new genre, the creators were bound to the limitations of Battle.net distribution.

These limitations have now defined the mechanical, aesthetic, economic, and philosophical conventions of the dota genre. Dota was an underground genre in the truest sense, operating entirely outside of commercial videogame development. The creators of Defense of the Ancients were given over half-a-decade of unopposed airtime to mold the expectations of the “hardcore dota player” and to mold their expectation of what a “great dota game” should look like. When commercial videogame companies finally decided to examine this genre, to take their resources and build a better game, they simply chose to whore the genre out. Both League of Legends and Dota 2 do little to deviate from Defense of the Ancients, and the other games which have attempted to capitalize on their popularity make the occasional step sideways instead of trying to bring the genre forward. Just as the Japanese role-playing game was largely the matter of sticking with what works when Final Fantasy and Dragon Quest made “press A to win” a commercial juggernaut, there has been little economic or critical impetus to reconsider the design of dota games.

So, here we are. There is no question in my mind that the people who turned these custom maps into their own genre were absolutely passionate about their work. These content creators spent countless hours trying to improve their formula simply because they thought it was a goal worth pursuing. But passion is not cause for celebration in itself. Throughout the rest of this book, I will point to this circumstance whenever it is necessary to explain the lessons of the genre. In spite of the commercial and critical praise that has been extended to dota, it is my opinion that the genre can be summed up in three words: “Eh, close enough.”

Continue to Chapter 3: The Camera

Discuss this chapter on the forums.