When Should a Detective Game Show a Hint? A Three-Step Ladder for One Clue
For a small mystery game, show a hint when the player asks for help or reaches a clear, tested point of friction—not simply because a timer ran out. Start with a nudge toward the evidence, then name the relationship to notice, and reveal the deduction only as a final, explicit step. This gives a designer tuning one clue a practical way to offer help without taking the inference away.
Build hints around what the player still needs to infer
Write the intended deduction in one sentence before drafting hints. Then separate the steps a player must take to reach it: find the relevant evidence, connect it to another fact, and draw the conclusion. Each hint should bridge one step and stop before doing the next step for the player.
Consider this original example. A fictional museum guard says she heard the east gallery clock strike nine during a theft. A maintenance log shows the clock had stopped at eight the day before, and a conservator’s note says it was restarted at ten on the night in question. The intended deduction is that the guard’s account cannot establish when the theft happened: the clock was not keeping time then.
A three-depth ladder for that clue might read:
**Where to look:** “Review the maintenance log and the conservator’s note about the east gallery clock.” This points to the relevant evidence without explaining why it matters.
**What relationship to notice:** “Compare the clock’s stopped and restarted times with the guard’s account.” This identifies the comparison but leaves the player to judge its consequence.
**Solution confirmation:** “The clock was stopped when the guard says it struck nine, so her account cannot confirm the theft time.” This states the deduction.
Keep these depths distinct. If the first hint already says the clock was stopped during the theft, the later steps add little. If the second merely repeats the clue locations, it may not help a player who has already found them. Each level should contribute one new piece of reasoning.
Let the player control how much help appears
A player-triggered hint is a reliable default for an optional ladder: the player chooses to open help, and can choose again to reveal the next depth. Labeling the controls plainly—such as “Hint 1,” “More guidance,” and “Show deduction”—helps communicate how much will be revealed before the player commits. Keep the first step low-detail so the player can stop as soon as they have enough.
The Acolyte lead designer describes a natural-language game in which players could ask questions or explicitly request help, with guidance adapting to the question. The same account also identifies a risk: capable players could uncover information before the story was ready, weakening its intended pacing. These are one designer’s observations about one game, not universal proof, but they point to a useful pair of considerations: make help easy to request, and consider when information is allowed to surface. [Jamin Smith’s Acolyte design account](https://www.gamedeveloper.com/design/deep-dive-designing-for-spontaneity-and-non-linearity-in-alternate-reality-games-with-i-acolyte-i-)
Automatic prompts can serve players who do not realize help is available, but they are also the main route to an unrequested spoiler. If you use them, offer a low-detail prompt first—“Need a nudge?”—rather than displaying clue content outright. A player should be able to dismiss it, and the prompt should not advance to a stronger hint without another clear action.
Use clue state first; use elapsed time cautiously
An eligibility rule answers when a hint may be offered or unlocked. Prefer game state that reflects the actual clue task: for example, make the relationship hint available after the player has inspected both the log and the conservator’s note. If the player has not found either, the first hint can point to them. If both have been found and the player has made no progress, the second can become available. State gates make the help relevant to what the player has already done.
Elapsed time can be a secondary signal, but standing still does not necessarily mean being stuck. A player may be reading, taking notes, or thinking. If you use time, treat it as permission to offer an optional prompt rather than permission to reveal the answer. Base the delay on playtesting and the pace of the scene; there is no universal number of seconds that guarantees the right moment.
A simple rule for this one clue could be: make the first hint available at any time on request; unlock the relationship hint once both documents have been examined; allow the confirmation only after a second explicit request. If you also want a nudge after a long pause, offer it only while the deduction remains unresolved, make it dismissible, and log whether players accept it. These are design choices to test, not established thresholds.
Keep help consistent with the game’s world
A hint can interrupt the mystery if it arrives through a voice or interface that does not fit the scene. In its Lacuna devlog, developer Julian Colbus describes concerns with conventional hints that appear in an out-of-world menu or seem implausible coming from a character. The team’s approaches included optional Highlight Mode, which outlined people and objects with new information, and redundant clue channels that could confirm important details. The devlog also gives a specific use for highlights: signaling when an object becomes investigable and when no highlighted leads remain. These are examples from Lacuna’s design, not proof that one format suits every game. [Lacuna developer devlog](https://www.gamedeveloper.com/design/detective-game-design-puzzles-vs-story-lacuna-devlog-)
For a single clue, that might mean placing the first hint in a case notebook, highlighting the relevant documents, or making a character’s optional remark available after the player has asked about the clock. Choose a presentation that fits your fiction and preserves the player’s role in making the deduction. A redundant clue can help someone who missed a detail, but avoid making every channel spell out the whole answer.
Playtest for unwanted reveals and unproductive stalls
Test the ladder with people who have not seen the solution. For this clue, record four things: whether a player asks for a hint, whether an automatic prompt appears before they want one, whether they find the two records, and whether they make the clock connection before seeing the confirmation. Also note wrong turns: for example, a player might decide the guard deliberately lied when the evidence only shows that her time estimate is unreliable.
Define two useful failure signals before the session. An **unrequested spoiler** occurs when a player sees a hint they did not choose and says it gave away the reasoning or solution. A **dead end** occurs when a player cannot identify a next action, even after inspecting available evidence, and does not make progress. A pause alone is not enough to classify someone as stuck; ask what they are considering after the session or at a neutral checkpoint, without feeding them the answer.
After each playtest, adjust the smallest relevant part. If players miss the documents, improve their discoverability or the where-to-look hint. If they find both but do not compare the times, sharpen the relationship hint. If the confirmation is repeatedly needed because the evidence feels ambiguous, review the clue itself before adding more forceful prompts. If an automatic prompt produces unwanted reveals, delay it, lower its detail, or require the player to opt in.
This process gives one clue a measured hint ladder: each step responds to a different obstacle, and each reveal has a clear trigger. The playtest tells you whether the timing and wording help players resume deduction while leaving the final connection theirs to make.