Methodology

Conviction is expressed as an integer from 1 to 4, rendered as filled dots (●○○○ low → ●●●● high). Each stage caps the conviction of the stages below it: a mixed reading upstream holds downstream conviction one notch below maximum, and Bitcoin cannot carry full conviction while macro liquidity and risk appetite disagree. This is the same upstream-conditions-cap rule the FX framework uses. The model produces a directional reading, never a buy/sell call or position sizing.

New to the labels — the gate chips (Open / Partly Open / Closed), the cohort states, the conviction dots? The cross-surface glossary maps them to the equivalent words on the FX, CbRates, and Asian Stress monitors.

The Altseason Conditions Score (APS) is the synthesis at the bottom of the cascade — a 0–100 composite of six current market conditions, scored daily, given the state of all five gates.

It is not a calibrated probability. No forecast event and no horizon are defined, and no out-of-sample record is published. The band boundaries were assigned, not fitted to observed rotation frequencies.

It is also not a count of conditions met. The six factors do not carry equal weight, and each scores in tiers rather than simply on or off: a condition can register as partly met and contribute less than the same condition fully met. A higher reading means the conditions are scoring higher on this scale — not that more of them are present.

And the scale does not currently reach 100. One factor is capped below its own maximum because a confirming input for it is not yet wired, so the highest reading currently attainable is short of 100. The gauge is still drawn to 100, which is why this is said here rather than left for a reader to infer.

The Trigger Watchlist under the score names the specific market conditions that would move the call. Each trigger is labeled with the stage it belongs to and split into upgrades (conditions that would raise the reading) and downgrades (conditions that would lower it) — for example, Bitcoin dominance closing below its eight-week average, or stablecoin float flipping to contraction. Triggers describe observable market thresholds; the model’s internal weights are not published.

Framework parameters — thresholds, lookbacks, and cohort definitions — are held fixed for twelve weeks at a time. Only the live inputs move between freezes; the machinery does not. Freezing the parameters stops them being re-tuned inside a window, so a reading cannot be adjusted after the fact. It does not by itself establish that the parameters generalize, and it does not remove any selection that happened before the freeze; no out-of-sample evaluation is published. Any parameter change happens at a scheduled freeze boundary, never mid-window.

Source cadence. NFCI and its sub-indices update weekly (Wednesday, covering the prior Friday) via FRED. VIX uses the prior daily close (FRED VIXCLS). Bitcoin and cohort prices refresh hourly; stablecoin float and macro series refresh daily — all from CoinGecko and FRED.

Derivation. A derive step recomputes the rotation signal after every fetch, so the monitor always reflects the latest inputs, gated by the slowest series in each stage.

Publication cadence. The Altseason Conditions score and its triggers are scored and published daily, at 14:00 UTC. Between publications the monitor state is indicative; the published score is the reading of record.