LiquidMind Academy

ICT/SMC concept library with interactive charts

Points of Interest (POI)/

POI Cluster

SMCOrder BlockFVGConfluence

TL;DR

A POI Cluster is formed when two or more independent Point of Interest types — such as a Bearish Order Block, an unmitigated FVG, and a Previous Day High — occupy the same price zone simultaneously. The engine automatically merges them into a single composite zone with elevated priority. Clusters carry significantly more institutional weight than any single POI in isolation, because multiple independent algorithmic processes converged on the exact same price area.

Interactive ChartBTCUSDT · 4H
lightweight-charts

How It Works

  1. 1

    The engine continuously scans all active POIs across every type (OB, FVG, EQL/EQH, PDL/PDH, Breaker Block, BPR, FLIP). When two or more POI zones share an overlapping or immediately adjacent price range, the system flags a Cluster.

  2. 2

    Cluster detection uses a configurable proximity threshold (default: 0.15% of price). Any two active POI zones whose boundaries fall within this threshold are treated as part of the same cluster region.

  3. 3

    The composite zone is defined by the union of all contributing boundaries — from the widest bottom to the widest top. The intersection (the overlap only) is used as the high-conviction entry zone.

  4. 4

    Cluster score is calculated additively: each contributing POI type adds its individual score, and a multi-type confluence multiplier (default ×1.4 per additional POI type) is applied to the composite total. A three-way cluster easily clears the READY threshold that a single OB cannot reach alone.

  5. 5

    In the signal pipeline, entry confirmation is gated at the intersection zone — the system waits for price to reach the narrowest overlap before counting the setup as READY. This prevents premature entries on the outer fringe.

  6. 6

    Clusters are dynamic: if one component POI is invalidated (e.g., the FVG is fully mitigated by price), the cluster is re-evaluated in real time. If only one POI remains, the cluster tag is removed and the surviving POI continues as a standard single-type zone.

LiquidMind AI Context

In LiquidMind's scoring model, Clusters are the highest-conviction POI category by construction. A three-component Cluster (e.g., Bearish OB + unmitigated FVG + PDH converging within 0.15%) can push the composite POI score above the READY threshold immediately — bypassing the incremental accumulation required for single-type POIs. The Reasoning Agent explicitly annotates Clusters in its trace output, listing each contributing POI type, its individual score, and the multiplier applied. Traders can inspect this breakdown in the Mind Trace panel. If the Cluster's intersection zone is narrower than the minimum zone width parameter, the system falls back to the union zone for SL calculation to avoid unrealistically tight stops.

System monitors this pattern in real-time