Track record · Open log
Every range we published, every day.
One row per day since June 2026. The range was published at 06:15 UTC, before the 24-hour window opened, and scored automatically against the real close. Nothing is recomputed, nothing is removed. Count it yourself.
01 · Two counters, never one
A wide range hits more often. So the coverage figure only means something next to the width of the ranges behind it, and next to what the lean of each reading did. All three are shown, on the same days.
Range coverage, full series
–
Average range width
–
Lean held, strict reading
–
This says whether our range and lean, stated in advance, held. It is not a return and not advice. The lean is scored as a distribution under three stated rules on the methodology page; the strict reading is shown here.
02 · Rules, fixed before the outcome
Horizon: 24 hours from the morning read (06:15 UTC).
Range: low and high named in the read, scored against the CoinGecko close at the end of the window. Inside = low ≤ close ≤ high, inclusive.
Width: (high − low) / baseline, in percent, as published.
One row per day: the first read of the day. A later manual run never replaces it.
History is append-only. A miss stays a miss.
Rule version: 0052 (24 August 2026) · scoreboard v1 (5 September 2026) · open log (19 September 2026)
Range: low and high named in the read, scored against the CoinGecko close at the end of the window. Inside = low ≤ close ≤ high, inclusive.
Width: (high − low) / baseline, in percent, as published.
One row per day: the first read of the day. A later manual run never replaces it.
History is append-only. A miss stays a miss.
Rule version: 0052 (24 August 2026) · scoreboard v1 (5 September 2026) · open log (19 September 2026)
03 · The log
| Day | Baseline | Published range | Width | Real close | Inside |
|---|
Loading the log from app.xlumience.com…