Separate preference from comprehension
A player can understand a system perfectly and still dislike it. Another can enjoy a dramatic moment without understanding what caused it. Treat these as different findings. If a test concerns combat readability, ask whether someone recognised the threat and understood their response before asking whether the encounter was exciting. Neither answer replaces the other. Clear rules do not guarantee a compelling game, but enthusiasm alone can conceal confusion. Define the question so that the observations actually address the decision waiting at the end of the session.
Choose a representative task
Give the tester a situation that contains the relevant problem without surrounding it with unnecessary work. To study inventory comparison, provide several genuinely different items and a reason to choose between them. A single obviously superior item cannot reveal whether the interface explains trade-offs. Conversely, a large collection with unfamiliar statistics may test patience rather than comparison. Set up the task carefully and preserve the same starting conditions between participants. Consistent context makes different reactions easier to understand without pretending that every player is interchangeable.
Keep instructions neutral
Explain the task without giving away the solution. 'Find a way across the room' invites observation; 'Use the glowing switch on the left' removes the very uncertainty the test may need to examine. If a participant becomes stuck, allow a reasonable pause before helping, then record exactly what assistance changed the situation. The point is not to make someone uncomfortable. It is to distinguish what the game communicated from what the facilitator supplied. That distinction is essential when deciding whether a cue or interaction needs revision.
Look for explanations, not verdicts
Notes such as 'bad menu' or 'fun fight' capture a reaction but leave the design problem unresolved. Ask the participant to describe what they thought would happen and which information led to that expectation. Combine their explanation with the observed sequence of actions. A long pause may indicate careful planning, unreadable labels or a distracting notification; the duration alone cannot decide between them. Preserve contradictory observations rather than forcing an early consensus. Differences between players can identify the boundary conditions of an otherwise promising design.
Change one meaningful condition
After a test, choose a change that follows from the evidence and makes the next session informative. If players missed an exit because its shape blended into the wall, adjust that relationship before adding a new tutorial. Keep a record of the previous condition so the team can compare outcomes. Do not describe a small test as proof about all players. Its contribution is a clearer question, a better-supported decision and a new version worth examining. Testing becomes a habit of learning when each round leaves a reason for the next.
Explore Shadow & Ember
