The engine journals what the market did.
Fetch and cache historical candles, replay them deterministically, and write structure events to a journal your strategies can trust.
From candles to structures in two commands
The A1 → B path: HistoricalDataSource feeds the StructureEngine, which journals deterministic events for swings, equal highs/lows, liquidity breaches, sweeps, order blocks, and FVG events.
$ wickd fetch --market BTC_USDT_PERP --timeframe 4h \
--from 2026-07-01T00:00:00Z --to 2026-08-02T08:00:00Z \
--alias jul-btc
cached candles → data/cache/.../candles.jsonl
saved alias jul-btc
$ wickd backtest --dataset jul-btc --run-id jul-btc-smoke
replayed candles · emitted structure events
wrote runs/jul-btc-smoke/structures.jsonl
Honest about status
Implemented
- fetch, backtest, manage commands
- Dataset aliases
- Structure journaling
- Binance USD-M via CCXT
Placeholder
- analyze — parses --run-id, returns not-implemented
Not yet
- Setup & trade engines
- Live execution
- AI analyst agent
- Dashboards