A Clue Proves Only Part of the Story: How to Keep Players from Treating Partial Facts as Complete Conclusions
An intact seal is a useful clue, but it cannot establish on its own that no one entered a warehouse. Use an inference boundary table to distinguish observations and rules, required premises, supported conclusions, and unresolved gaps—and show players what to check next.
Start by giving “the seal is intact” a limited scope
An intact warehouse seal initially supports this statement: “At the time of inspection, this seal showed no visible damage.” To reach “this door has not been opened,” you need to know whether the seal remained in its original position and whether opening the door would necessarily damage it. To reach “no one entered the warehouse,” you must also check other entrances and define the period under discussion. The truth of a clue and the scope of a conclusion are two different things.
When writing, do more than label evidence “reliable.” State both the conclusion it supports and the extensions it does not permit: “Under these premises, this evidence supports a particular judgment; it does not yet cover these subjects or periods.” If a premise has not been checked, keep the conclusion conditional.
The following story, The Rainy-Night Warehouse Receipt, is a fictional teaching example. Its characters, setting, dialogue, and rules were written for explanation. It is not a real user case, a tested result, or a DramaFork product feature.
Establish the scene: what does one seal actually cover?
Warehouse keeper Su He seals the warehouse’s main door when the evening bell rings. After the morning bell the next day, the bookkeeper discovers that a crate of dye is missing. The player examines the main door: the paper strip spanning the gap lies flat, its mark is recognizable, and it has no visible tears. Su He says, “The seal is perfectly intact. No one could have gone in last night.” The period under investigation runs from the sealing of the door at that evening bell to the following morning bell.
When the scene first appears, establish that there is a small loading door in the west wall. Before the evening bell, porters were still taking inventory inside the warehouse. Whether the small door was shut overnight and when the porters left remain unanswered. These questions are visible from the outset, and the loading door is not an entrance introduced only at the ending.
This example requires one fictional rule to be stated in advance: opening the main door in the normal manner will tear the paper strip spanning the gap. However, whether the strip now in view is the one placed there last night has not been checked. The known rule about opening the door and the unanswered question of whether this is the original strip must not be merged into an established fact.
When the dye left, whether anyone entered, and who moved the dye are separate questions. Even if entry during the period under investigation is ruled out, that does not establish that the dye was not removed, much less clear any particular person of responsibility.
A reusable inference boundary table
First label observations, known rules, and known events separately. Then enter the narrowest judgment each supports. “Cannot infer” means the evidence is still insufficient, not that those conclusions are necessarily false.
| Information obtained and its type | What can be inferred | Premises that must be established | What cannot be inferred | Next check |
|---|---|---|---|---|
| Observation: after the morning bell, the paper strip on the main door shows no visible damage | The strip looks intact at the time of inspection | The inspection is sufficient to detect the damage being discussed | No one entered the warehouse during the period under investigation | Compare with the previous evening’s sealing record |
| Known rule: opening the main door in the normal manner will tear the strip | If the original strip remained in place and intact throughout the period under investigation, this method of opening the door can be ruled out during that period | The original strip was not replaced, never left its original position, and remained intact throughout | None of the entrances were used | Check whether it is the original strip and whether it remained in place and intact |
| Observation: a small loading door exists in the west wall | The main door is not the only entrance that needs checking | No additional premise is needed to establish its existence | Someone passed through the small door | Examine the small door and its opening and closing records |
| Known event: porters were taking inventory inside before the evening bell | There were people inside before the door was sealed | The information concerns the same inventory check | The warehouse was empty after the door was sealed | Check when inventory ended and when people left |
When adapting the table, label the information type in the first column: observations state what was seen, rules state their conditions of application, and events state what has been confirmed to have happened. Do not mix in interpretations such as “suspicious” or “innocent.” The second column limits the subject, action, and time; the third lists conditions still requiring confirmation; the fourth records judgments prone to overextension; and the fifth turns the gaps into investigative actions.
This table is an author’s working document for checking narration, choices, and case-closing text. It need not be shown in full. However, players should have an opportunity to discover gaps that affect a fair judgment before making the crucial choice.
Work through an example: new evidence fills only its corresponding gap
In the first round, the player has only the observation of the intact strip at the scene. An investigation note could read: “The main-door seal currently shows no visible damage. Whether it is the original strip from last night remains to be checked.” Immediately offering a choice that says “This proves no one entered the warehouse” would package an observation as a conclusion endorsed by the author.
In the second round, the player obtains a visual record made when the door was sealed and compares the distinctive ink blot and number. This supports “the strip now in view matches last night’s record,” but does not independently establish that the strip remained in place all night. An additional piece of evidence does not let you skip an unchecked premise.
In the third round, the player asks about the small door and the inventory check. Suppose the story provides verifiable evidence establishing that the small door was opened during the period under investigation. The player can now say, “This entrance was opened during that period and needs further investigation.” They still cannot say “someone entered through it” or “the dye was carried out this way.” An opened door, people entering or leaving, and goods passing through each require corresponding evidence.
This round could be written as a short exchange:
Su He: “The seal on the main door isn’t torn. How could anyone have gone in?” Player: “That still doesn’t account for the small door in the west wall. It was opened during this period, but we still need to find out whether anyone passed through, who they were, and which direction they went.”
The player has not immediately learned the truth, but can decide what to do next. A progress report could read: “Entry during the period under investigation cannot yet be ruled out. The small door is confirmed to have been opened; whether people or goods passed through remains unresolved.” A limited conclusion can still move the task forward.
Three approaches, and where they tend to fail
For explicit teaching, notes can display “Currently supported” alongside “Still to be confirmed.” To leave more room for deduction, show the seal’s position, the small door, and the inventory timing, let a character make an overly broad claim, and give the player an opportunity to question it. In a very short piece, the question can be limited to “Was the main door opened in the normal manner?” Even then, the premise that the original strip remained in place and intact throughout the period under investigation must be stated.
These three approaches differ in how much guidance they provide and how broad a question they ask. Do not omit a crucial premise and then expect players to supply it themselves at the ending. Nor should you repeatedly remind them that “we cannot be sure” without giving them actions that reduce the uncertainty.
A common failure is to put an overreaching conclusion in the reward text: the dialogue is careful, but the completion message announces, “Entry by anyone has been ruled out.” Another is to pile up evidence with the same scope: a photograph of the seal, a description, and a character repeating what the photograph shows still provide no check of the small door. More evidence does not necessarily mean broader coverage.
Do not turn “we cannot prove that no one entered” into “therefore, someone definitely entered,” either. The dye might have been removed before the door was sealed, or the wrong items might have been counted. If you use these alternative explanations, the work should also provide corresponding avenues of investigation. There is no need to enumerate every possibility.
When reviewing the draft, circle words such as “all,” “all night,” “necessarily,” and “only” in crucial conclusions, and identify evidence that covers each one. Then construct a situation that satisfies the established facts but makes the conclusion false. In this example, you could hypothesize: “Between the evening sealing and the following morning bell, someone entered through the small door in the west wall while the strip on the main door still showed no visible damage.” This is compatible with the current observation, the known rule, and the evidence showing that the small door was opened, yet it would make “no one entered during the period under investigation” false. This is a possible situation to be tested, not an established finding that someone entered. Merely saying “the small door was used” is not enough, because someone might only have left. Once you find such a situation, add evidence, narrow the conclusion, or leave the issue unresolved. Do not let a confident tone do the reasoning for you.
