using-systems-as-experience
Systems as Experience
Use this contract for the concrete task. Apply a short relevant check for a small change; expand investigation when the failure, risk or requested artifact warrants it. Resolve facts from the repository and runtime before imposing a process. Delegation is optional and should answer a bounded unresolved question.
Design contract
State the intended player experience, observed problem and design constraints. Preserve authored moments where useful; a systems-first philosophy is a choice, not a requirement for every game.
- Express a small interaction hypothesis: which mechanics combine, what players can observe and which choices/results should become possible.
- Use an interaction matrix or concrete scenario to expose missing affordances, dominant strategies and unreadable outcomes. More combinations do not automatically mean more meaningful depth.
- Make cause/effect, feedback and recovery visible enough for players to form and revise hypotheses. Hidden depth needs discoverable evidence, not arbitrary obscurity.
- For optimization play, expose useful measurements and tradeoffs without reducing all play to one mandatory solution.
- For emergent narrative, connect state/event consequences and player interpretation. Event variety alone does not establish memorable stories.
- For community/mod systems, define compatibility, permissions, determinism/content identity and failure isolation. Do not trade save compatibility or security for an assumed community benefit.
- Prototype the smallest interaction that can test the hypothesis, then use relevant human playtests. Synthetic inspection can identify structural problems; it cannot establish enjoyment, discovery or community response.
Evidence and output
Produce a bounded design change, interaction/prototype spec or review containing intended experience, mechanism, competing explanation, observable signal and playtest plan/results. Distinguish design intent from observed player behavior. Stop or simplify when the prototype does not support the hypothesis.