Iceberg Likelihood Score

Quick Reference

Property	Value
Dimension	signal
Category	order_flow
Version	v1.0
Output Column	iceberg_score

Likelihood of hidden iceberg orders based on order book depth pattern anomalies

Formula

(order_book_depth - rolling_mean(order_book_depth, zscore_window)) / (rolling_std(order_book_depth, zscore_window) + 1e-8)

CDM Inputs

Column	CDM Table	Description
order_book_depth	cdm_lob_snapshot	Order book snapshot data 鈥?depth levels, bid/ask prices and sizes

Parameters

Parameter	Type	Default	Description
levels	integer	10	Number of order book levels to analyze
refresh_rate	integer	100	Order book snapshot interval in milliseconds
zscore_window	integer	50	Rolling window for depth anomaly z-score

Output

Column: iceberg_score

Iceberg likelihood score (0 to 1)

Market Intuition & Trading Rationale

Iceberg likelihood score detects hidden order patterns via depth anomalies: (depth - 渭) / (蟽 + 蔚). When a large hidden order is being worked through an iceberg algorithm, the visible depth at a level replenishes immediately after each fill 鈥?creating a detectable pattern of unusually persistent depth. The z-score flags when depth at a level is statistically unusual relative to its own history.

Usage Cases

• Hidden order detection: High iceberg_score at a specific price level 鈫?likely iceberg order at that level. Trade in the opposite direction of the iceberg (the hidden order will push price toward you).
• Execution avoidance: If iceberg_score is high on the side opposite your order, a large hidden order is absorbing liquidity there 鈥?your execution will face additional slippage.
• context: Primary signal in iceberg_detection pack 鈥?iceberg_score is paired with refill detection and depth stability metrics.

YAML Definition

name: iceberg_likelihood_score
description: Likelihood of hidden iceberg orders based on order book depth pattern
  anomalies
category: order_flow
dimension: signal
version: v0.9.0 (Beta)
required_inputs:
  - order_book_depth
output_column: iceberg_score
output_description: Iceberg likelihood score (0 to 1)
tags:
  - microstructure
  - iceberg
  - signal
parameters:
  levels:
    type: integer
    description: Number of order book levels to analyze
    required: false
    default: 10
  refresh_rate:
    type: integer
    description: Order book snapshot interval in milliseconds
    required: false
    default: 100
  zscore_window:
    type: integer
    description: Rolling window for depth anomaly z-score
    required: false
    default: 50
formula: (order_book_depth - rolling_mean(order_book_depth, zscore_window)) / (rolling_std(order_book_depth,
  zscore_window) + 1e-8)