LiquidMind Academy

ICT/SMC concept library with interactive charts

Points of Interest (POI)/

Previous Day Low / High (PDL / PDH)

LiquidityDailyTime-Based

TL;DR

Previous Day Low (PDL) and Previous Day High (PDH) are the lowest and highest price levels reached during the prior daily candle. These levels are natural magnets for Smart Money — stop losses cluster just beyond them, forming pools of Sell Side Liquidity below PDL and Buy Side Liquidity above PDH. Sweeping these levels is a key manipulation tactic before institutional reversals.

Interactive ChartBTCUSDT · 4H
lightweight-charts

How It Works

  1. 1

    At the close of each daily candle, the system records the high (PDH) and low (PDL) of that completed day.

  2. 2

    PDL acts as a support-side liquidity pool — retail traders place stop losses just below this level, creating Sell Side Liquidity (SSL).

  3. 3

    PDH acts as a resistance-side liquidity pool — retail traders place stop losses just above this level, creating Buy Side Liquidity (BSL).

  4. 4

    When price sweeps below PDL (PDL sweep), it collects SSL — Smart Money fills long positions by triggering retail stops. A bullish reversal typically follows.

  5. 5

    When price sweeps above PDH (PDH sweep), it collects BSL — Smart Money fills short positions by triggering retail stops. A bearish reversal typically follows.

  6. 6

    The significance increases when PDL/PDH sweeps coincide with other POIs (EQL, Breaker Block, EFVG) in the same zone — confluence amplifies conviction.

LiquidMind AI Context

LiquidMind automatically tracks PDL and PDH for every configured trading pair. At each new daily candle, the system creates active POIs marking the previous day's high and low. PDL is treated as side=1 (long bias on sweep) and PDH as side=-1 (short bias on sweep). These POIs are assigned a 2-day TTL and replaced when a new day begins. During POI clustering, PDH/PDL levels that overlap with other POI types receive elevated priority in the scoring pipeline.

System monitors this pattern in real-time