How to Play: Rare Answers, Scoring, Timers & History
Understand open answers, real timers, insight scoring, estimated familiarity, valid-submission counts, retry, reveal and archive rules.
Start a run
Daily play rotates 7 Bible-only questions from 15 existing Bible topics; new unlimited runs use seven Bible prompts, packs use their actual length, and new friend rooms default to five Bible prompts. Other religions require selecting a topic pack; older rooms retain their original set.
Submit one person, book, place, event or object within the prompt's scope. Current timers range from 16 to 20 seconds. Pack and unlimited practice offer an in-run untimed toggle; daily and friend duels remain timed.
Accepted translations and aliases identify one entity. Similar spelling alone does not make an answer correct; numbered books and distinguishing personal identities matter.
Unrecognized input, learning and end reasons
Unrecognized means the current bank and matcher cannot confirm the input, not that the answer cannot exist. Keep and edit the input to retry, and check the prompt and study guide.
Skipping and timing out are recorded separately from a correct answer. Reveal/help and untimed conditions are labelled rather than silently compared as unaided timed results.
The learning reveal and complete accepted-answer spectrum remain available, followed by the next question, references or a replay.
Scoring and estimated familiarity
A valid answer earns 100 minus its smoothed familiarity, between 1 and 99 insight points. The seven-question theoretical ceiling is 693 points; actual attainable scores depend on the listed answers and frozen run values. Insight is a game score, not a measure of faith. Historical scores retain their original points; the former display multiplied these by 10 metres per point.
Initial familiarity from 1–99 is an editorial/AI estimate, not a player-choice percentage or a probability distribution that sums to 100. Its starting weight is 40.
Scoring familiarity = (40 × initial estimate + 100 × valid hits for this answer) ÷ (40 + all valid hits for this question), rounded and clamped to 1–99. Observations belong to this instance and content version, mix modes and count submissions rather than distinct players or the whole internet.
A run pins its questions, matching/scoring versions and initial output. Later edits or statistics do not silently rewrite the saved result.
Dates, archives and old results
The chosen date and full question snapshot stay with a daily run. Questions are reused; a new order is not a claim of newly authored questions.
A new Bible daily or replay never overwrites an existing original daily result. Previous mixed daily snapshots stay under their original rules; different daily-selection versions are compared separately. Replays are separate; unversioned old results are labelled and restored in their recorded question order.
Where no recorded historic snapshot exists, a retained bank can generate an explicitly labelled reconstruction. It is not claimed as the edition originally published on that date.
Storage corruption or save failure is surfaced rather than silently replacing the original data. Clearing browser storage can remove local records.
Friend collisions and comparison
Room participants answer the same question set. The same entity can collide across different translations or aliases. The room selects void or halve for collisions.
Shares retain mode, date, content/matching/scoring versions, untimed and help conditions. A result card alone does not establish equal conditions across all runs.
The nickname leaderboard is manually submitted and belongs to this instance; it is not an identity-verified global competition.