One hierarchy, re-read every day.
Sir Elliott builds the complete weekly-to-15-minute tree in one reading, then reads it again — whole, once a day — against the count it accepted the day before.
Keep four purposeful windows
The store retains roughly ten years of weekly data, three years of daily data, one month of hourly data and one week of 15-minute data. Every scale is collected at its own candle close, and only bars that can no longer change are kept.
Turn candles into structural pivots
A volatility-scaled ZigZag with a fixed threshold per scale removes market noise while keeping every supplied endpoint traceable to a real candle. The threshold never adapts to the data, so a quiet week and a wild one are read with the same ruler.
Read from context to detail
Weekly pivots establish Cycle and Primary structure. Daily, hourly and 15-minute pivots may only subdivide the active parent branch.
Build one integrated hierarchy
Every reading takes all four ZigZags together and returns one coherent tree rather than four competing counts. The first reading of a market and every later one are the same reading.
Maintain two genuine scenarios
Primary and Alternative share the same evidence but express materially different interpretations, each with probability, targets and invalidation.
Structured before presentation
The model returns explicit wave ids, parent links, pivot endpoints, invalidations and forecast zones instead of unstructured chart prose. An answer that cannot be drawn is refused and the previous count stands.
Re-read on a timetable
Once a day each market is read again in full, with its accepted count in front of the model. Confirmed structure is kept where it still fits and revised only where the evidence moved.
Keep every accepted history
The complete structured JSON and its reason are stored as an immutable revision before Atlas renders it.
When does the model read?
On a timetable, and for no other reason. Every market with the model switched on is read once a day, senior scale first, with yesterday's accepted count in front of it. Nothing in the tape brings a reading forward, and nothing in the last answer pushes one back.
What is deterministic
Binance and Yahoo Finance candles, database persistence, ZigZag pivots, structured parsing, validation, scale isolation, rendering and revision history are handled by code. The model supplies the Elliott interpretation inside that bounded contract, once a day; free-form prose is never used to draw the chart.
What the method cannot promise
Elliott Wave interpretation is probabilistic and multiple valid counts can coexist. Scenario probability is a relative model assessment, not a win rate. Forecast dates are intentionally represented as broad horizons and target areas. Sir Elliott is analytical software, not personalized investment advice or an automated trading system.