Flow Persistence Score

Quick Reference

Property	Value
Dimension	signal
Category	order_flow
Version	v0.9.0 (Beta)
Output Column	flow_persistence

Flow persistence score: autocorr(ofi, window) - measures temporal persistence of directional order flow

Formula

rolling_corr(ofi, ofi, window)

CDM Inputs

Column	CDM Table	Description
ofi	cdm_*	CDM source table

Parameters

Parameter	Type	Default	Description
window	integer [5, 1000]	100	Rolling window for autocorrelation calculation

Output

Column: flow_persistence

Autocorrelation of order flow imbalance over the rolling window

Market Intuition & Trading Rationale

Flow persistence score measures the temporal consistency of OFI direction: rolling_corr(ofi[t], ofi[t-1]). It answers: is the current order flow direction similar to recent flow, or is it flipping randomly? High persistence means sustained directional pressure — the hallmark of informed flow. Low persistence means chaotic flow — the hallmark of noise trading.

This is the signal-dimension counterpart to order_flow_autocorrelation (stability dimension). The signal feature captures the magnitude of persistence (how strongly OFI predicts itself). The stability feature captures the reliability of that persistence estimate. Together they provide a complete picture of flow predictability.

Persistence tends to be higher during institutional trading hours (when large programs are being worked) and lower during retail-dominated periods. It also varies by instrument — more liquid instruments tend to have higher flow persistence because institutional participation is higher.

Usage Cases

• Informed vs noise classification: High flow_persistence_score (> 0.5) = informed flow (trade with it). Low score (< 0.2) = noise (fade extremes or avoid trading). This classifier is more robust than OFI magnitude alone.
• Execution program lifecycle: Monitor persistence through a trade. Rising persistence during your entry = you're catching an institutional program mid-execution (good). Falling persistence = the program is winding down (prepare to exit).
• Regime filter: Only trade OFI-based signals when flow_persistence_score exceeds a threshold (e.g., 0.3). Below the threshold, OFI is too noisy to trade reliably — stay flat.

YAML Definition

name: flow_persistence_score
description: 'Flow persistence score: autocorr(ofi, window) - measures temporal persistence
  of directional order flow'
category: order_flow
version: v0.9.0 (Beta)
dimension: signal
status: Pre-release
required_inputs:
  - ofi
output_column: flow_persistence
output_description: Autocorrelation of order flow imbalance over the rolling window
parameters:
  window:
    type: integer
    description: Rolling window for autocorrelation calculation
    required: false
    default: 100
    constraints:
      min: 5
      max: 1000
formula: rolling_corr(ofi, ofi, window)