TTRPG Witness Credibility Tracker: A Fair Mystery Template

Handwritten investigation notes on a desk, representing a TTRPG witness credibility tracker

A TTRPG witness credibility tracker helps a Game Master manage an investigation where witnesses remember different details without deciding that every contradiction is a lie. The goal is to keep testimony consistent, offer players fair ways to test it, and let social scenes change their understanding of the mystery. This system-neutral template works for D&D, investigative games, and mysteries in nearly any fantasy setting.

Credibility is not the same as honesty

A truthful witness can be wrong about time, distance, clothing, or motivation. Someone who lies about one thing might tell the truth about another. Treat the information itself as the thing players investigate, rather than asking them to roll once to detect whether the entire character is trustworthy. A credible account needs evidence, not a number displayed over the NPC’s head.

Before play, record three separate elements: what the witness actually saw, what they believe happened, and what they are concealing or mistaken about. This distinction prevents a GM from changing the facts mid-session just to outwit an unexpectedly clever theory.

Use five columns for a one-page witness record

FieldGM question
ClaimWhat would the witness say in plain language?
AccessWhat could they see, hear, or plausibly know?
PressureWhat might make them omit or distort details?
CheckWhat evidence could confirm or challenge the claim?
ChangeWhat happens when players present that evidence?

Write a sentence for each field, not a full biography. The tracker should fit beside your session notes. A claim can be reliable despite weak access if corroborated; an apparently confident statement can be wrong because the witness heard a bell from another street. For interview question prompts, pair this with our TTRPG witness interview guide.

Worked example: the missing guild ledger

A warehouse account book disappears before the guild election. Mara, a dock clerk, says she saw the brewer’s apprentice leave the warehouse after the ninth bell. In the GM notes, the bell actually rang late because the tower keeper stopped to clear a jam. Mara saw a figure with the apprentice’s red cloak, but not the face. She also owes the brewer money and fears losing her job.

Her claim is that the apprentice left late. Her access is a partial view from across the quay. Her pressure is a debt, not proof of guilt. Players can inspect the bell mechanism, check a delivery record, or speak to the apprentice. When shown that the bell ran late, Mara updates her timing honestly. If confronted about the debt, she initially becomes defensive; a respectful offer of confidentiality opens the conversation instead.

Provide at least two ways to verify important claims

Do not require the party to ask the one perfect question. For each essential fact, prepare an independent physical check and a social or logistical check. The warehouse time can be tested through a bell keeper’s account and a baker’s oven schedule. The cloak can be checked against a laundry receipt and the apprentice’s testimony. These are different routes to the same conclusion, not separate passwords.

On a failed roll, provide the observable fact plus a complication: inspection takes longer, the clerk asks for a favor, or the evidence is incomplete. That preserves forward movement. Our failed investigation check advice explores this principle in detail.

Track changes after questioning, not during every sentence

A dynamic interview should never require the GM to update ten tiny modifiers. Use only three narrative states: unexamined, challenged, and resolved. A challenged claim has a specific competing piece of evidence. A resolved claim has either been corroborated, corrected, or shown to be deliberately deceptive. A statement can also remain unresolved, and that is fine.

Keep a date or scene number beside significant changes. If the witness heard a rumor yesterday but witnessed the event three days ago, distinguish those two sources. Our mystery timeline template makes that chronology easier to protect.

Let pressure create character rather than a hidden truth meter

A witness’s reluctance should have a human explanation. Fear of a superior, shame about missing work, a promise to a friend, or loyalty to a neighborhood can all shape an interview. Do not assign automatic deception to anxious, neurodivergent, or socially awkward mannerisms. Describe what an NPC actually does—refuses to name a source, asks for a private conversation, consults a written note—rather than making body language a universal lie detector.

Also give the players meaningful choices. They can insist on a public confrontation and risk frightening the source, or build trust through another lead. Their choice affects cooperation but does not retroactively change the underlying event.

Copyable witness credibility tracker

  • Witness: __________________
  • Claim in their words: __________________
  • What actually happened: __________________
  • What they could observe: __________________
  • Reason for error or omission: __________________
  • Physical verification: __________________
  • Second verification: __________________
  • After evidence is shown: __________________
  • State / last updated: __________________

Common questions

Should players see the tracker?

Give them a player-facing clue list, but keep secrets about the underlying truth behind the screen. Players should not have to guess how your hidden credibility rating works.

How many witnesses should a short mystery have?

Three distinct viewpoints are often enough for one session: someone who saw a detail, someone who understands a routine, and someone personally affected. More interviews are useful only when they change the available decisions.

A fair witness mystery rewards checking sources and revising theories. If the players leave saying, ‘We were wrong about the time, but the clues were there,’ the tracker is doing its job.

Similar Posts

One Comment

Leave a Reply

Your email address will not be published. Required fields are marked *