Why Dota Sucks — 6. The Chess Pieces

Synopsis: The characters in the dota genre have been hailed for their complexity and diversity. In order to debunk this, we should explain how the framework for the “chess pieces” came into place. The hero system is a direct transplant from Warcraft III, and technical limitations prevented content creators from making this system more complex. Commercial game developers have since adopted that framework. But in stripping the RTS concepts from Warcraft III and centering the action on the hero system, dota not only compares unfavorably to the RTS games where you control far more moving parts, but to comparable character-action games with more complex control schemes and skillsets. Ironically, it is the simplicity of the characters that provides the illusion of diversity, where the marginal differences become more important when the tools for victory are so similar. In the end, the simplicity is intended to create accessibility, where the range of simple characters assures you will not have to make grand leaps in order to learn them. Where other developers have looked to up to the complexity of a genre, the creators of the dota genre looked at one of the simpler RTS games and chose to make it simpler.

——

We should now focus our time and energy on the genre’s defining concept. Whether you think of them as heroes, champions, gods, or elsewise, every player selects a character for the battle, and they are designed to “support”, to “carry”, to “jungle”, or fulfill some aspect of a team’s overall strategy.1 While there are different selection formats that span a range of game modes, they follow the same ideal: Players want to choose characters that complement individual and team skill sets while creating difficult matchups for the opposing team. Many dota games—Defense of the Ancients, Dota 2, League of Legends, and Heroes of Newerth—are defined by their massive character rosters, all featuring over one-hundred participants. And because the genre has been standardized around a small number of maps and game modes, the presence of these characters as interchangeable but interesting “chess pieces” is crucial.

The creators of the various dota games have a number of considerations to make in the character construction process.2 From this range of choices, the dota genre has been hailed for the diversity and complexity of its character designs. Well, for starters, the chess pieces in the dota genre can only be as complex as the game systems that provide their foundation. In addition, diversity is a natural outgrowth of complexity. The more possible ways the complexity can be arranged, the more possible ways you can make things different from each other. So, in order to debunk the chess pieces, all we have to do is demonstrate that the framework for the characters is busted. We don’t have to debunk every single character when we can demonstrate the flaws in the production line. Let’s do this by diving back into the origins of the genre.

When Blizzard Entertainment released Warcraft III: Reign of Chaos in 2002, it marked the most complete exploration of “micromanagement”3 in Blizzard RTS games, with powerful hero units at the center of the battle. Each hero would pair their standard physical attack alongside four “hero abilities”. Unlike the single-click abilities commonly associated with RTS games, hero abilities could grow significantly more powerful over the course of a match. From here, most heroes were given two castable “active” spells, a “passive” ability that is permanently enabled for no mana cost, and an “ultimate” ability that is typically acquired in later parts of a match. To prevent the overuse and abuse of these spells, the castable abilities don’t merely cost mana, but are governed by a cooldown timer, and you must wait X number of seconds before casting a spell again. To round out this skillset, these heroes possess six item slots, and items can be acquired, purchased, and used over the course of the match.

It is important to understand that Blizzard designed this system with only their concepts in mind. The Warcraft III user interface is built for the maximum number of commands—Move, Stop, Attack, and so forth—that any one Warcraft III unit will acquire. As a result, the game’s “command grid” leaves no redundant space. Building more complex soldiers for an interface that won’t support them is a burden unto itself. But most critically, Blizzard coded their game around the limitations. The maximum number of hero abilities that can be used by any one hero is five.4 Likewise, a hero’s backpack can never hold more than six items. These are hard-capped limits and they cannot be changed. The workarounds for this system were flimsy and hardly anything that could be standardized across a wide range of maps. But most importantly, the fixes emerged well after Defense of the Ancients became a centerpiece of the Warcraft III custom game scene.5

If your goal was to create a more advanced template for heroes, you were shit out of luck. Your map was to be an advertisement and supplement for Warcraft III, and it was not to outdo Warcraft III.6 We can conclude that this system created considerable limitation for the custom game community. Mapmakers could not expand the framework for heroes or make it more complex. So, just as dota games borrowed their camera, their visual concepts, auto-attacking, and last-hitting from the Blizzard RTS, the defining concept in the dota genre is also a direct transplant from Warcraft III. Nearly every dota game uses a similar character template, ranging all the way from Defense of the Ancients to games (AirMech, Bloodline Champions, Monday Night Combat) which merely feature elements of dota.

But the Warcraft III hero system was intended for a game in which heroes are just one part of the machine. It was one concept to go alongside base construction, unit production, squad-based tactics, and economic supremacy. As part of a larger system, as one means to complexity, the hero system is adequate. But on its lonesome, it is not. The heroes in Warcraft III are not complex. Even Warcraft II: Tides of Darkness, which introduced many of the genre’s conventions and has been outclassed by the body of games which followed it, features spellcasters with five different abilities. Today, games like the Men of War series feature individual soldiers who are significantly more complex—and have significantly greater access to skills—than any one hero in Warcraft III. It is simply a mistake to hold the hero system as a high point for complexity, because it has been outclassed even within its own genre.

As a result, the heroes in the dota genre are complex and diverse…when compared with the heroes in Warcraft III, which were not only intended to be straightforward, but to complement a standing army. But once the genre became a subset of an existing game, and you were removing all of the other things that define the RTS, content creators needed to build new game systems, ones that would lead to more complex characters than the Warcraft III framework could ever consider. This would have required systems that allow for characters with more spells, more attacks, more options for mobility, a more complex range of user input, and possibly a more complex item system. This is precisely what did not happen because Warcraft III made sure that it could not happen. And once commercial game designers got a hold of the genre, there was no critical or consumer impetus to make it happen. Dota creators and dota fans were already happy with their “complex” and “diverse” heroes.

When it comes time to sing praises of character design, the most common namedrops are Meepo and the Invoker,7 two mainstays of the Dota series and two of the most difficult characters to play successfully. In reality, these characters demonstrate how limited the hero system actually is. Meepo’s defining trait is the ability to create clones of himself and use the clones to attack different parts of the battlefield. But this is a concept which compares unfavorably to, I don’t know, every RTS game which requires players to manage more than one front.8 The Invoker makes a more compelling case, because his four basic spells can be used to “Invoke” ten different advanced spells, making fast reflexes and proper spellcasting a must. But this shows that in order to achieve anything that would appeal to the veterans of more complex action games—or even veterans of the dota genre—its creators must entirely subvert the existing game rules. They must introduce additional complexity in the form of a bizarre and cumbersome gimmick, one that actually limits complexity because 1) “Invoke” is bound to a cooldown timer, and 2) you can only “invoke” (store) two of those ten spells at any time.9

“But the complexity of these characters is not just in their movesets! It’s the items! It’s about being able to choose your playstyle through the things that you purchase over the course of a match!” And what a mess this is. Even role-playing games figured out that items which give you functional complexity are far more interesting than items which give you numerical complexity. Most of dota’s early-game consumables are healing items that cannot be used in combat, and equipment with castable abilities is typically acquired late in the game. It is unlikely that a game will last long enough for players to create an entirely new skill set out of those purchasable items. In addition, the items in dota games are “character neutral”, meaning that any character can equip or use any item and gain benefit from it. The diversity that could be provided by a class-based equipment system has gone untouched.10

But even if dota presented players with better items, other games let you show off the loot. You wear it, you shoot it, you swing it. That’s right: Most items in the dota genre do not come with in-game graphics. To provide these graphics would expose how ridiculous it is that a floating ball of gas like Io can equip boots and swords like any human. This lack of visual intrigue may not stand out in Warcraft III, where items are a tiny portion of the game model, but it sticks out in a item-driven action game like Dota 2.11 A Divine Rapier may add several hundred attack points to the character’s statline, but you will never see a character wield it. So, it wouldn’t matter whether this system has “depth” or whether it is a means to more diverse and complex characters. Interesting mechanics should be backed with interesting visual outcomes, and the items in dota games aren’t even providing the outcomes that were present in the role-playing games of the nineties.

So, it’s quite simple: The unit properties associated with RTS games—movement speed, turn speed, collision size, spellcasting, all built for an RTS interface—were designed for a game in which you control dozens and even hundreds of soldiers at any time. These systems were designed to make armies feel distinct.12 This can be seen on full display in RTS games with asymmetrical factions—particularly StarCraft and Armies of Exigo—where armies feel bulky, or mobile, or fast, or offensive, or defensive, depending on the size and composition of that army. As a result, dota doesn’t just compare unfavorably to those RTS games, which feature a greater range and number of moving parts. They don’t even compare favorably to role-playing games like Planescape: Torment and Dragon Age, which require the player to control multiple “heroes”. By reducing Warcraft III to a game centrally played around one character, you are inviting comparisons to the countless character-action genres which were built to satisfy their most capable and enthusiastic players.

Even in the most immediate comparison to dota, we will find that the characters in an isometric brawler like Diablo II can draw from their own list of thirty different spells and abilities.13 Yes, thirty different choices. Place this alongside the robust Diablo equipment and loot generation system—where items provide a wide variety of interesting ideas—and it’s not even a contest. Whether or not this works in application—where players will upgrade and mercilessly abuse a small percentage of those skills—we can easily find that the chess pieces in Diablo are more complex than what you will find in any dota game.14

But when we begin to up the ante, dota looks worse and worse. Even the most basic 2D “belt-scroll” brawlers allow players to string together elaborate attack sequences, whereas “combos”15 have no place in dota because auto-attacking and the spellcasting system prohibit their existence. But when you work your way into 3D brawlers, you will find that fans of Devil May Cry 3, Devil May Cry 4, and Bayonetta are still finding new and novel approaches to the ridiculous range of weapons and moves available to them. The range of skills is best exemplifed by the so-called “combo videos” that demonstrate the most flashy and intricate knowledge of the movesets found in these games. I would argue that dota doesn’t compare well to a genre where the characters can carry more weapons—which often have their own movelists and can be freely switched during combat—than the number of spells and abilities that dota characters have.

And in many of these brawlers, complexity is not merely about what the player can do. In the Ninja Gaiden and Devil May Cry series, enemies are often as capable as the player himself. A wide range of dangerous enemies acts as a means to complexity, because new enemies will demand different tactics that will require a variety of skills and tools. But in the dota genre, computer-controlled opponents—whether the enemy army or the neutral enemies which dot the map—have one purpose: They are pinatas. They will never outwit the player and they will never beat the player to the punch. You whack them over the head and gold comes out. This could work in Warcraft III, a game with enough player interaction to make the computer-controlled participants a secondary attraction.16 It cannot work in the dota genre, where the farming phase is a primary attraction and players must fight the computer.

Then you can compare dota to the fighting game genre, a gold standard for fast action in videogames. In 1991’s Street Fighter II: The World Warrior, the player gets access to a six-button layout, providing six basic attacks with variations for standing, crouching, air, and close-up attacks. In addition, the player gets two or three special moves, each of which has three variations in strength.17 And that is the game which kicked off the genre proper! Fighting games such as Tekken feature gargantuan movelists and some of those moves—particularly “chain throws”—may require over a dozen inputs in order to perform successfully. In games such as Marvel vs. Capcom 2 and Skullgirls, long strings of combination attacks are the backbone of anything vaguely considered “standard play”. The demands of input are complex enough that even the best players will screw up from time to time. Whereas the input in the dota genre—select a spell and then click where you want to cast it—is so basic that it is almost entirely impossible to fail an input. You can merely misuse it.18

Perhaps most importantly, fighting games use the synchronous nature of movement and combat to make similar moves feel different.19 Take a look at the fireballs and projectiles that define the genre. Peacock’s Screwball Cannonball (Skullgirls) becomes a standard single-button move. Ryu’s Hadouken requires a “quarter-circle” input. Guile’s Sonic Boom uses a press-and-hold “charge” input. Ness’s PK Thunder (Super Smash Bros.)20 uses a two-button input and the projectile can be controlled after it is launched. Akuma’s Zankuu Hadouken and Morrigan’s Soul Fist (Darkstalkers) can be fired while in the air. Urien’s Metallic Sphere (Street Fighter III: Third Strike) combines a rotation input with a “release to fire” method. The result is that moves can feel wildly different in spite of their similar utility, and characters can feel innately offensive or defensive beyond their base skills and movement. Whereas characters in the dota genre feel the same because players are using single-press inputs in order to perform similar spells, abilities, and movement.

But it is not just that dota games have simplified input to the point of banality—failing to take advantage of the range of input provided by a keyboard—the entire genre swears by this simplification. Where RTS hotkeys were once about word association—mapping “Train Grunt” to the “G” key—Defense of the Ancients players took advantage of the hotkey customization in The Frozen Throne and moved all of the hotkeys to one end of the keyboard. Dota games now build on this trend, and the games played with the mouse and keyboard now standardize hero abilities to a row of keys. In Dota 2 and League of Legends, it’s the Q through R keys. In Demigod, it’s the 1 through 4 keys. The layout is so synonymous with the genre that players will often identify spells with the key that it is affixed to. For instance, noting that a character can be easily countered by another character’s “Q” or “W”.

This has created a unified theory of design when nearly every other genre features a diverse range of ideas and control schemes. This is how you end up a genre of brawlers where developers take wildly different approaches to things as simple as blocking, where the counter-attacks in the Arkham games, block button in the Ninja Gaiden series, dodge mechanics in the Devil May Cry series, and parrying in Metal Gear Rising: Revengeance lead to entirely different challenges and designs. This is how you end up with countless action game genres which all pursue different control schemes in order to make the happenings of their world work. The dota genre has standardized itself around the simplified character skillset, and sees similar concepts used in largely the same way with the same boring control scheme.

Ironically, it is this simplicity which provides the illusion that the characters are complex and diverse. In discussing the sense of sameness that can be found in the Killer Instinct and Mortal Kombat games of the nineties, world champion Street Fighter player Seth Killian argued that the best characters in those fighting games simply have better versions of the moves that everyone else is using. This forces players to extract victory from the marginal differences between these characters. As Killian notes, this is not unlike the actual game of chess, where in a game of symmetrical sides, white gains a huge advantage by making the first move.21 In a game where the differences between the tools are less significant, any deviation becomes more impactful and important.22 And in a game where players are using a wildly limited moveset, in a game where characters can be slaughtered in seconds, the minor differences between characters add up very quickly.

So you see, dota players have sold themselves a myth that their games feature some of the most complex and diverse character rosters in videogames…but are bound to one of the simplest control schemes in the entire medium, and are built around a character template that was merely a subset of an RTS game. This is not complex, and by extension, it cannot be diverse. Every faction in a decent RTS game is Meepo. Every character in a decent dungeon crawler is the Invoker. And because competing genres could be iterated and expanded through the punctuated leaps provided by sequels, they got more and more complex with every go.

And if you would argue that the complexity in many of the above games is not being used to compete against the adaptable intelligence of a human player: There is absolutely no reason that the systems in games like Ninja Gaiden or Diablo II could not be adapted to a versus multiplayer environment, just as the systems in Doom gave way to Quake and Unreal Tournament, and just as the systems in God Hand gave way to Anarchy Reigns. But do not sacrifice the quality of the game in the chase for competition. If it is purely human competition that you desire, then a friendly life lesson: There are better things to compete in than videogames. It does not matter whether primitive things like “skillshots” are being commandeered by players who play the game for no reason other than that it is their job. The complexity of the competition in the dota genre cannot hide the utter lack of complexity in its game systems, and to pursue that competition at the expense of the game is to end up with a significantly worse videogame experience.

Ultimately, the character designs in the genre are not about complexity or diversity. They are about simplification and accessibility. When you combine this simple control scheme with a massive roster of largely similar combatants, you are assuring that your player base does not have to make huge leaps in order to learn new concepts, or to even learn a new character. You don’t even have to learn a new set of controls. This system will allow beginners to choose what they are comfortable with, instead of having to learn uncomfortable concepts and master unfamiliar game systems.23 Or, as the creators of Dawngate have suggested, these systems will allow the player to “personalize your playstyle”. There are enough deviations (characters) within each “role” (support, carry, jungle, etc.) that you will eventually find something which you are comfortable with and enjoy playing. Or, to bring things full circle, you will not have to explore a wide range of diversity in order to play the game.

And let’s be clear: I am not arguing that if the characters in your game have more weapons and tools, it makes for a better game on principle. After all, execution of the concepts is paramount, and the counter-examples I have provided in this chapter have largely succeeded on this front.24 And you know why? Because these games use the right cameras. They provide the proper complexity in visual design. They demonstrate a competency in their repetitive elements. They provide the players with proper feedback mechanisms. And most importantly, their complexity is the result of game developers attempting to cater to their most enthusiastic players, and making the framework more and more complex with every go. But when it came time to build the dota genre, amateur mapmakers looked at Warcraft III and asked: “How can we take one of the simpler games in its genre and make it simpler?” And simpler they made it, turning one small part of that game into its own genre.

There’s only one question left to ask: If dota is using such a simple and basic character template, how did we end up with a perception that these simple characters lead to greater depth? Let’s answer that in the next chapter.

Continue to Chapter 7: Complexity Through Choice

Discuss this chapter on the forums.