Signal → direction

The signal-to-direction lifecycle is how Kraite progressively evaluates seven active indicators at one timeframe at a time. It runs hourly at :30 via kraite:cron-conclude-symbols-direction, starts each Binance symbol at the first configured timeframe, advances only when inconclusive, and exposes the resulting LONG / SHORT / no-direction state to token selection.

This is the lifecycle lens view of the indicator pipeline. For the indicator inputs themselves, see Indicators.

End-to-end flow

   :30 tick
      │
      ▼
   kraite:cron-conclude-symbols-direction
      │
      ▼
   for each tradeable Binance symbol:
      │
      ▼ (rate-limited via TAAPI throttler)
   QuerySymbolIndicatorsJob at 1h
   fetch 7 active indicators: 5 directional + 2 validation
      │
      ▼
   ConcludeSymbolDirectionAtTimeframeJob
      │
      ├─ conclusive → persist LONG / SHORT + timeframe
      ├─ inconclusive → repeat at 4h → 12h → 1d
      └─ all exhausted → clear direction and mark invalid
      │
      ▼
   ConfirmPriceAlignmentWithDirectionJob
      │
      ▼
   CopyDirectionToOtherExchangesJob + support/resistance persistence
      │
      ▼
   Bybit / KuCoin / Bitget exchange_symbols rows updated to match Binance

Per-timeframe decision

Production has five active directional indicators (candle-comparison, three EMA checks, and emas-same-direction) plus two validation indicators (chop and pivotpoints). There is no majority weighting: every available directional conclusion must agree on the same LONG or SHORT, and both validation gates must pass. Mixed directions, no directional result, a failed validation, missing constructs, or mixed-run data makes that timeframe inconclusive.

Progressive timeframe path

   1h: conclusive LONG  ─► stop at 1h, persist LONG

   1h: inconclusive
   4h: conclusive SHORT ─► stop at 4h, persist SHORT if change rules pass

   1h: inconclusive
   4h: inconclusive
   12h: inconclusive
   1d: inconclusive     ─► clear direction, mark invalid

The configured list is an ordered fallback path, not a multi-timeframe vote. A first-time or unchanged direction can conclude immediately. A direction change must reach the configured minimum timeframe index and keep a consistent path: earlier evaluated timeframes may be inconclusive or agree with the new direction, but cannot contradict it.

Known timeframe configuration drift

Production currently walks 1h → 4h → 12h → 1d. A migration intended 4h → 6h → 12h → 1d, while the seeder still restores the production list. The runtime conflict is documented; the intended product list remains unresolved.

Cross-exchange propagation

Architectural decision

Direction is concluded once on Binance and copied to every other exchange's matching exchange_symbols row. The reasoning: directionality is BTC-anchored, BTC moves the same way across exchanges to <0.1 % drift on liquid tokens, and TAAPI's rate-limit budget is finite. Duplicating the indicator query per exchange would burn the budget without changing the answer. CopyDirectionToOtherExchangesJob does the propagation as the last step of the conclude pipeline. Symbol-name divergences (BTC→XBT on KuCoin, 1000SATS→10000SATS on KuCoin / Bybit) are handled by the token_mappers table.

Freshness

indicators_synced_at means the pipeline completed an end-to-end attempt, not necessarily that it produced a direction. A successful conclude stamps it, but so do unchanged indicator data, all-timeframe exhaustion, and a path-invalidated direction change. Transport or step failures that abort before those terminal paths remain visible as stale pipeline activity.

What this lifecycle does NOT decide

  • Which symbol to trade — that's token selection's job, downstream.
  • Position size, leverage, entry price — the position lifecycle's open phase decides these.
  • Whether to open at all — gated by the market regime's shouldBlockOpens().

This pipeline answers "if this symbol were to be opened right now, which way?". Nothing more.

  • Indicators — the seven active checks evaluated at each visited timeframe
  • Token selection — the next layer that picks which symbol to actually open
  • Position lifecycle — the consumer of symbol_direction
  • Market regime — the circuit breaker that can veto an open even with a clear direction