One tap a day for 8 to 12 weeks. Then an honest answer to one narrow question: did your daily ratings rise and fall with your Chord more than with 199 decoy charts that share your birthday and your days?
This is a test, not a promise. Most people will see no link, and we will tell you so.
The chord values, the correlation and the decoy comparison are computed [REAL]. The daily score, the decoy design and the thresholds are chosen conventions, fixed in a locked protocol before any code was written [CHOSEN]. What a day felt like belongs to you [INTERP]. A portrait of days already lived, never a forecast.
Birthday season: every year around your birthday the transiting Sun, Mercury and Venus come back near their birth places, so the chord rises with the calendar. Your 199 decoy charts share your birthday, your birth time and place, and the exact days you log, so that seasonal rise is the same for them. It cannot make your real chart win.
1 · Your birth data
Birth time matters only for the Moon. If you don’t know it, use 12:00. The time-zone offset is checked on the server from the place and date, using historical rules.
On this device only. Your birth data, daily ratings and notes are kept in this browser’s storage. There is no account. Export a file if you want a backup.
When you tap “See my result”, your birth data and your ratings (dates and 1–5 numbers only) are sent to our server to be computed. Nothing is saved there.
Your notes never leave this device and are never part of the test.
Honest either way. The verdict rule is locked in advance. If there is no link, the page says so. With one person and 8–12 weeks, only a fairly strong link can be detected.
About 1 person in 20 with no real link sees “tracked” by chance. That is what a 0.05 threshold means.
Stays on this device. It is never sent anywhere and never part of the test.
Progress
Calendar days to the minimum (56) and the recommended length (84)
Logged days (minimum 40)
Calendar day
0
Logged
0
Streak
0
Status
–
Try it with demo data
Two synthetic users on the founder’s public chart, 84 days each. Pure noise: ratings are random and unrelated to any chart, so the honest answer is no link. Planted signal: ratings were built to follow the real chord at a stated strength (b = 0.5) plus noise, so a working test should catch it. This is the positive control: it shows a “No link found” would mean something.
Verdict
Your chart against its 199 decoys
Each grey bar counts decoy charts by how well their chord lined up with your daily scores (Spearman ρ). The gold line is your real chart. The dashed teal line is the typical ρ that a planted signal of strength b = 0.5 produced on these same days.
decoy chartsyour real charttypical planted signal (b = 0.5)
Exploratory comparisons (do not drive the verdict)
Your days and your chord
Your lived experience is real data in its own register. The verdict above tests one narrow question: whether the numbers line up across days. Tap a day to see your note, your score and that day’s chord.
chord κ1 (left scale)your daily score 1–5 (right scale)a day with a note
Tap a day on the chart or in the list.
The locked protocol (v1.0). Daily score = the mean of energy, creativity and mood. Exposure = your hub chord κ1 (natal + progressed + transit around your natal node) at 12:00 UT each logged day. Statistic = the Spearman correlation ρ between them. The same ρ is computed for 199 decoy charts with your birth month, day, clock time and place, and birth years drawn (seeded, so repeatable) from 100 years either side of yours. p = (1 + decoys with ρ at least yours) / 200, one-sided, so the smallest possible p is 0.005. Minimum: 56 calendar days and 40 logged days. “Your Chord tracked your days” needs p < 0.05 and ρ ≥ 0.15. Anything else is “No link found”. The quintile chord κ5 and the three single ratings are exploratory only. Every result also comes with a positive control: how often a planted signal would have been caught on your own days. Pooled analysis, for people who opted in before seeing a result: the mean rank across people, as in the pre-registered population study. No public claim is made before 200 contributors.