Commitment as a design axis
Every decision asks the player to live with it. In combat, recovery frames and no universal escape punish reaction and reward anticipation. In quest preparation, loadout and resource decisions made before entering a situation determine what is available under pressure. The skill is the same in both contexts: read ahead, commit, and absorb the outcome.
Anticipation rewarded defensively and offensively
Correctly reading an opponent's action carries a double payoff: avoiding damage and opening an offensive window. The skill loop becomes self-reinforcing. Better reads produce better output, which incentivizes further investment in reading. Commitment and defensive timing are not separate systems but two expressions of the same underlying principle: anticipation and technical mastery are the primary skills the game tests.
Distinct systems over varied implementations
A weapon roster built around different verbs rather than different numbers ensures each option teaches the player something new. A claymore doesn't simply swing slower than a sword and shield. It demands its own mental model, changing how the player reads space, timing, and risk.
Multiplicative playstyle design
Layering a modifier system on top of a base system generates combinatorial depth without proportional development cost. The legibility cost is real and requires mitigation at the UI and onboarding level, not in the systems themselves.
Adversarial AI as a readable system
Enemies with behavioral states that respond to player actions are a second language the player must learn. Damage changes behavior; behavior reveals new patterns; knowledge becomes a form of progression. The design obligation this creates: behavioral states must be legible before they are understood. Unreadable states punish rather than teach.
Tutorialization debt
When a game's depth exceeds what it teaches, mastery becomes self-directed by default. This is acceptable under three conditions: variables must be isolatable, failure must be consequence-free, and the full system must be reachable through play alone. Players who engage with that space progress; players who don't, plateau. The filter is on willingness to experiment, not willingness to seek external resources.
Core loop integrity
The core loop is reading enemy behavior, committing to a response, and converting a correct defensive read into an offensive window. That loop begins before combat: choosing which skills, items, and gear to bring is itself an act of reading and committing. Players who engage with both layers develop genuine mastery. Abilities that deal damage or clear content without requiring a read or a commitment let players bypass this entirely. In low-stakes content this goes unnoticed. Over time it obscures what the game is actually testing and delays mastery. Progression systems must reinforce the loop, not route around it.
Generative pacing
Intervals between action spikes must be engaging in their own right, not merely tolerated. Resource gathering, crafting, and side content are not filler between peaks. If they feel like dead time, they are dead time, regardless of intent. The design obligation is that downtime be productive, legible, and intrinsically motivated. It must also be optional: players with sufficient resources should be able to skip it entirely and stay on the action. Forced downtime is a tax; voluntary downtime is a feature.