How we collect and count

Lucky Sign publishes anonymized aggregates from voluntary daily rituals. This page defines what our numbers mean — and what they do not.

What participants do

Each day, participants receive one sign (number or symbol). They live the day, then optionally report whether the sign "found" them, tag a feeling, and keep a short sentence. No rerolls. No predictions.

What we measure

Participant-days — one person carrying one sign on one local calendar day. This is not the same as unique people unless a report says otherwise.
Noticed rate — share of moments with noticed_level yes or a_little.
Emotion and place tags — optional self-labels from a fixed ontology.

Indexing thresholds

Public SEO pages index only when data density is sufficient:

Sign pages: 200 participant-days, 20 public stories, 14 days coverage
Sign + emotion pages: 20 stories, 15 people, 10 days
Story themes: 30 stories, 20 people
Research reports: 500 unique participants (monthly)

Limitations

Self-reported. Not a controlled study. Selection bias applies — participants opt in. Late-arriving data may appear in monthly research recompute but not retroactively change finalized yesterday snapshots.

Citation and license

Aggregates may be cited with URL and retrieval date. CSV exports where available are licensed CC BY 4.0 with attribution to Lucky Sign.

Markdown mirror: /research/method.md