Since the beginning of games as a medium, there was always a given uncertainty when it comes to authorial intent. Traditional media like film or music do not feature direct interactivity with systems in the same way games do; they are primarily tools of communication where a message is transmitted in code. They are a way to encode information, vague or specific, in a way that can be decoded by the recipient as long as the message is compatible with the key owned by them (our perception filters).
Games are in a peculiar place in this dynamic as they inherited this from traditional media, as they use their audio-visual tropes, but add another layer for communication that has not been thoroughly explored yet: Game Mechanics. Simplified, game mechanics are the formalized rules and goals under which a test is performed. In the context of video games, this is the code, the program executing input commands from the user. This interactivity is different than the decoding of information in traditional media. The rules are set by the designer, but a degree of freedom is awarded in utilizing the rules, therefore a variance is introduced in decoding intent. Sometimes rules can create interplay in ways the designer never could have foreseen.
Current games more often than not are obsessed in trying to treat mechanics as they would be in traditional media, trying to create games with a clear way of being played. Scripted sequences; meticulously planned item locations; “realistic” weaponry and movement, these all try to convey gameplay intent.
However, sometimes, something, somewhere doesn’t go as planned and a set of internal rules emerges that has not been taken into consideration by the designer. It is almost impossible to predict how a player will utilize certain elements, or how they will ultimately interplay with each other, and for better or worse these emergent rules and systems are then part of the game, providing the player with opportunities to exploit.
Take for example the simple oversight of leaving the acceleration- and speed-value unrestricted when moving backwards, which leads to this behavior:
At first glance, this seems like an egregious oversight with dire consequences, a game-breaking flaw of large magnitude. However, this judgment would be devoid of context as no mechanic is inherently good or bad. A concrete value judgment concerning an emergent mechanic can only be made in relation of how the game is played and what the ultimate goals set by the player and/or the game are.
In the above example, the infinite speed bug provides a unique way to speedrun Half Life 2. This doesn’t however mean that it is a good emergent mechanic for “regular” play. It all depends on what the challenge is in the environment set up by the rules. If the challenge is “finish the game as fast as possible”, going out of bounds is a valid tactic to succeed. However, if the challenge is to defeat all the enemies, going out of bounds would be an invalid way to bypass that challenge.
The assertion that you aren’t playing the game as intended are invalid in this case, as the game (rule-set) inherently does not care about an intended meta-goal like “experience the story.” Anyone that ever ran or wrote a Dungeons and Dragons campaign certainly knows what I’m talking about; players rarely play the adventure as you “intended” it to be played.
Ultimately, the value-judgment of a mechanic has to be made on a case-by-case basis. Bunnyhopping in Quake 3 Arena isn’t in itself good or bad, but it just so happens that it improves the gameplay by adding an emergent element of movement-skill to the game, increasing the skill-ceiling and enhancing tactical as well as strategic decisions available to the player. It can thus be deemed a beneficial emergent gameplay element.
In contrast, being able to glitch your way on a ledge in Dark Souls and cheese Queelag with a bow is not mechanically beneficial to the game. It does not introduce a new way of playing; it does not enhance the challenge by offering new tactical possibilities. To the contrary, it reduces the challenge of fighting Queelag; it removes risk by putting the player in a safe position and hence can be evaluated as a bad emergent element. This is not to be confused with, for example, making the Taurus Demon kill itself by making it fall down the ledge which is a skillful manipulation of the AI that requires timing and knowledge of animation and patterns from the player, although it can be performed by accident by a novice as well.
Of course, not all emergent gameplay elements are simply glitches. Animation Cancels for example are most well-known from fighting or hack-and-slash games like Street Fighter and Devil May Cry, but curiously, they also tend to feature in modern FPS games.
In Battlefield 2, the Dolphin Dive maneuver is an Animation Cancel in which the jump animation can be canceled mid-flight into a prone animation. This conveys the superior accuracy of the prone stance while air-borne while keeping the forward momentum of the fall, making for both a harder to hit target as well providing superior accuracy without impeding movement.
As a mechanic, this seems to go against the grain of the game where the stances and movement modes are supposed to convey different accuracy ratings that hinder or enhance the shooting experience. However, this tactic was quickly accepted and adopted by the community as it introduced a more varied and mobile combat experience than simply standing still while shooting.
In contrast stands the infinite medic resurrection exploit in Battlefield: Bad Company 2, in which there is a brief period of invulnerability to a revived player which can in turn revive another player, and so on, effectively making it impossible to kill those two players.
Besides Animation Cancels, there are other ways to create emergent gameplay mechanics. AI manipulation is another well known tactic in many games, but especially the Elder Scrolls series as its simulation leaves many doors open for NPC exploitation.
Possibly the most known exploit in Skyrim is the Bucket Exploit, in which you place an object on the NPC’s head to break line-of-sight with the player. This enables the player to steal or kill anyone in the vicinity without being “seen”, and therefore without penalty. As far as emergent elements go, this is one of the more creative ones and sharply on the edge of being considered a malicious oversight as it removes a penalty of failure.
Other emergent solutions go as far back as the old Ultima Series; the player could poison the food of Lord British to kill him, although the NPC was essentially invulnerable to regular attacks. Another similarly interesting tactic can be employed to kill the very powerful and famous NPC Drizzt Do’Urden in the game Baldur’s Gate. The player can surround and immobilize Drizzt with his party members, command them to leave the party and attack Drizzt with a ranged weapon or pole-arm staying well out of reach of his attacks. While Drizzt is extremely powerful and the player should not even be able to hit him, the rules of Advanced Dungeons and Dragons allow for critical strikes to be a guaranteed hit, making the exercise dependent largely on luck. In a recent playthrough of the Enhanced Edition, it took me close to three hours to defeat him with regular pole-arm attacks. Same goes for the Fallout series in which the player can kill extremely powerful NPCs in a myriad of ways, from “healing them to death” with stimpaks, which is considered a non-aggressive action, to pickpocketing TNT into their pants.
We can even go further back to the times of traditional tabletop games where parties would intentionally try to anger the most powerful creature accessible to them, usually a lesser god, and try to kill him by exploiting gameplay mechanics. Elaborate plans were created to deal with a specific deity and characters crafted for this singular purpose. I am confident in the belief that at this point in time all gods in the Dungeons and Dragons universe pantheon have been killed at least once by someone.
Designers vs Gameplay
Game designers often experience a strong resistance when it comes to emergent gameplay mechanics. The thought process is simple to understand: It’s their game, and by god, you are going to play it as they intended you to. Some of the time, this is a healthy outlook, as more often than not glitches and exploits are a hindrance to gameplay. Sometimes however, this is not the case.
Take for example K-Style GunZ. Emergent mechanics like the Butterfly Cancel opened up a variety of movement and attack options and a new layer of high-level play. When these “unintended” behaviors were removed from the game, the community clamored for their return as they enhanced the game and made it unique from every other third-person multiplayer shooter. Ultimately, these emergent mechanics were integrated into maps for learning and honing these skills. Chris Wagar has a far more detailed coverage of the mechanics in question in his More Than Mashing: Legendkiller 2 article.
Warframe found itself in a very similar situation recently when the developers tried to patch out movement-chaining from their game, essentially disabling repeated air-sliding/dashing. It didn’t last for longer than a week and the functionality was returned to the game via a hotfix prompted by complaints from the players. You can find out about the various movement and animation cancel techniques in Warframe on Shuki Warframe’s Youtube Channel.
Designers must realize that emergent mechanics can be part of their game and that not every exploit or glitch is in itself bad. Yes, more often than not, they can create imbalance, and especially in multiplayer games this is a large problem, but sometimes they actually add to the experience. Developers can’t simply reject an emergent mechanic. As demonstrated, they are not always evil in and out of themselves.
They must also realize that they do not hold the entire monopoly on intent, as the player can change the context in which the application of these quote-on-quote exploits might be a benefit to their design.
Designers must not be afraid to embrace the gameplay they created, by intent or by accident.