Output D — Laggard Candidates

Event-driven surfacing of coins that historically follow a hub coin's moves with measurable lag.

Cadence

Event-driven; emitted when a hub coin's 5-minute return crosses the configured breakout threshold (default 4%) and the regime confidence multiplier permits.

Schema

The pydantic model is memenet.schemas.laggard.LaggardAlert.

Example payload:

{
  "alert_id": "lag_20260502_140523",
  "timestamp": "2026-05-02T14:05:23Z",
  "type": "laggard",
  "hub_coin": "BONK",
  "hub_move_pct_5m": 4.2,
  "candidates": [
    {
      "coin": "WIF",
      "expected_lag_minutes": 30,
      "historical_correlation": 0.78,
      "historical_hit_rate": 0.62,
      "confidence": 0.71
    }
  ]
}

Confidence

historical_hit_rate is the empirical fraction of past hub-breakout episodes after which the candidate moved in the same direction within its lag window. The confidence field is the product of the base statistical confidence and the Output E regime confidence multiplier.

How to consume

Alerts are appended to outputs/signals/laggard.parquet and optionally forwarded to a configured webhook (Phase 5).