The Post-CAG Manifesto

The era of character action games (CAGs) is over. If Devil May Cry 5 was their final hurrah, Nier: Automata had set the framework for the post-CAG¹ era two years prior. The language of CAGs – launchers, stingers, devil trigger, combos and links – has transformed from a mantra to summon action fundamentals – positioning, tempo, resource management, and crowd control – to the mere suggestion of such. Their aesthetic trappings present the fantasy of excellence without a struggle for competency put forward by their mechanics. Juggles come for free, elaborate animations hide downtime, and enemies sit idle without challenging the player. The scene has finally morphed into shrouded action RPGs.

Now that we’re firmly in the era of the Soulslike, it may be tempting to look back on the CAG era and reminisce. However, to do so keeps us from recognizing poisonous elements lurking in the genre from the beginning. Those of us who got involved in the action gaming scene can take the post-CAG era as a chance to reevaluate our virtues and explore our desires. There are two post-CAG mindsets: one that devolves into action RPG apologia, and one that blossoms by sloughing the limitations of character action. To envelope ourselves in the latter we must steel ourselves with demands for the games we enjoy.

No More Progression

There will be no more leveling systems, no more gear grind, and no more shops. These all simulate the process of acquiring proficiency by scaling the character’s power as the player invests time. However, these often reward hollow time, where the player completes tasks that do not increase actual proficiency to speed up the process of their character gaining proficiency. There is no need for this. Games themselves allow for the player to practice, learn, lab, and improve on their own terms. The simulation solves a problem that didn’t exist. When a progression system mandates mutually exclusive decisions, then there may be merit, but too often CAGs implement them with an end goal of complete saturation of the character’s toolkit. This is especially apparent with difficulty systems designed to restrict new players from attempting higher difficulties that are tuned around more “complete” characters. Again, there is no need to introduce a progression system to do this. Restrict certain tools to specific difficulty levels if you absolutely must, otherwise let the player gain proficiency on their own terms.

This philosophy impacts the structure as well. The original Devil May Cry has adventure game concepts trapped in its DNA, where the player’s missions are contextualized around exploring areas and performing item fetch quests. These concepts deserve justice in a system that can accommodate them, not in an action game. Diversions for unrelated minigames should be dropped for the same reasons. Arcade beat-em-ups often kept to under an hour of playtime with just as much depth as a CAG by honing in on constant encounters without the same interruptions. The best non-combat moments from these games came when they leveraged parts of the original system, such as the glass-breaking minigame in Final Fight testing the player on their understanding of where their hitboxes fall on the Z-axis. Action games that want to break the pace should consider these methods instead of veering off into wasteful territory.

No More Combat

The CAG gameplan has put a crushing amount of weight upon enemy design as a form of stochasticity-driven dynamism. As the aesthetic goal of CAGs has been to present all-powerful characters who can handle any task, character movement tends to be linear with pivots that happen on a dime, and attacks tend to have ample interruptible-as-soon-as frames. Both of these traits limit the character’s ability to have complexity-driven dynamism, where the character’s handling, abilities, or outcomes change depending on their state. Characters may have enumerated coarse states that alter this (such as a jump state versus a grounded state), but CAGs have rarely developed past this. For the player to be threatened by an enemy, they must put the character in such a state that they will unavoidably be hit by an attack. Thus, enemies (which often feature similar or fewer features to that of characters, thus preempting them from delving into complex dynamism) must catch the player off-guard and attack during times when their character cannot escape. The dynamism required to do this ends up “looking random,” hence the “stochastic” tag, but the primitive implementation of these systems results in state-machine configurations that can be exploited. This crumbles the player’s perception that the enemy is dynamic in a random way; this is the downside of stochastic dynamism.

While overlapping many enemies at once (preferably of different types that potentially bolster complexity) aids this issue, it has been equally negated by the onslaught of instant defensive maneuvers that keep the player from ever being left helpless. In the post-CAG era, this latter concept has overtaken the entire mechanical sphere: all attacks are meant to be negated through timing. In its own cruel way, this fulfills the original CAG ideal by creating characters that are untouchable, and it does it without the need for dynamism for either the character or the enemies. Still, for those who wish to return to the consequential decision-making of older CAGs, this is not sufficient.

Combat will never dies as a gameplay concept, and there is room to improve the above character/enemy interplay on a dynamic front. But in the post-CAG era, why limit ourselves to this when it comes to action gameplay? If decision-making is the our virtue, then the goals of the game must be examined, as decisions are made relative to the goal they achieve. Combat often limits itself to complete slaughter as a win state², which not only calcifies the tie between the player and the enemies but also keeps ambiguous entities from entering the arena. In Jet Set Radio, the goal is instead spraying graffiti on pre-marked surfaces, where the character has minimal combat abilities and thus must avoid enemies instead of necessarily confronting them. In many racing games, such as Daytona USA, the goal is finishing before other opponents without eliminating them, and said opponents can passively assist the player with slipstreams just as often as they obstruct the player. In both of these cases, stochastic dynamism is retained for the enemies, but the depth of their behavior matters less thanks to goals that aren’t immediately oriented around them, along with the complex dynamism present in the handling for both games.³ Confrontation exists in these systems, but the lack of combat helps exemplify decision-making outside of the character/enemy relationship, and for these games in particular it lets them focus in on complex dynamism in a way that CAGs often lack.⁴

No More Characters

CAGs are character-oriented; most games in general are. However, this limits the spatial complexity of the player’s input, as they’re always rooted to an individual spot on screen. CAGs have continually tried to give their characters tools that allow them to transcend this, but given the restrictions of their combat systems and their dedication to exuberant power fantasies, solutions such as Final Fantasy XV’s warping end up removing commitment in the same way as other CAG defensive techniques. This also highlights an additional restriction of characters in most character action games: they lack any sort of self-adversarial ability. The need for combat in CAGs stems from the need for adversaries, as otherwise there is no fail state. Single-character games that don’t have this restriction, such as 3D platformers, tend to lose the stochastic dynamism that comes with enemy characters in favor of static terrain or periodic obstacles, and thus they add such enemies regardless. This criteria is important, or else we would want to look closer at post-CAG titles that actually do buck the single character trend, such as Astral Chain, NEO: The World Ends With You, or Slitterhead. While these titles gain spatial complexity from their use of multiple characters (which can sometimes be controlled simultaneously for additional complex dynamism), they’re still shackled to the other conventions of the genre in terms of progression and combat.

Falling-block puzzle games give us a view into what a non-character-oriented action game with self-adversarial techniques could look like. Variations of Tetris all have the player continually cycle what pieces they control, with their previous pieces forming a structure at the bottom of the playfield under the player’s guidance. Here, even though the pieces themselves control rather simply, the system manifests complex dynamism through the capability for any piece to interface cleanly with the structure below, where structures that cannot be cyclically cleared through filling every block in a line will eventually reach the top and end the game. This ties in with each game’s stochastic dynamism too, where piece drops are pseudo-randomly chosen by an algorithm. Here there are no adversarial entities, only ways for the player to end their game prematurely through poor play. The Tetris: The Grand Master series understands the strength of this “character switching” quite well, as the game’s difficulty is scaled via the gradual increase in drop speed and decrease in lock delay that pushes the player to switch contexts between newly arriving pieces faster and faster. Tetris also succeeds in this regard due to the complete ambiguity of each one of its separate piece archetypes. A game that delineates between its entities more is Wetrix, where good entities (suns that let you convert any pool of water you’ve made to points) and bad entities (bombs that destroy part of your underlying structure) have more clear-cut roles compared to the ambiguous entities (arrows that transform the terrain of your structure, water that converts to points, etc.). Still, even in this system, the player has utter control over each entity that appears, giving them the awful duty of eventually ending their own run through incompetence.