PD Array Confluence Selection (Secret To Selecting Algorithmic PD Arrays)
Also: secret to selecting PD arrays, PD array confluence, loaded deal setup, time-anchored PD array confluence, algorithmic PD array selection
Visual Context Required
This concept requires chart visuals for full understanding.
The "Secret To Selecting Algorithmic PD Arrays" is ICT's framework for identifying the highest-probability PD array levels out of the many that can exist on a chart. The secret is CONFLUENCE combined with TIME ANCHORING. Not all PD arrays are equal. With IPDA data ranges of 20, 40, and 60 calendar days of valid PD array lookback, many levels will exist on a chart at any given time. The arrays that deliver the most reliable reactions are those where multiple elements align simultaneously: CONFLUENCE: Two or more PD arrays converging at the same price level or in very close proximity. This includes: - An inefficiency (FVG, suspension block, volume imbalance) overlapping with a quadrant level of a wick - An order block aligning with a fair value gap - A breakaway gap level coinciding with a premium/discount wick quadrant - A suspension block's CE or quadrant aligning with another PD array TIME ANCHORING: The PD array is approached at a time when the algorithm is expected to react — specifically: - At a session open (NY open, London open, Asian open) - During the first 30 minutes of the opening range (9:30–10:00 AM ET for equities/futures, or the equivalent session open window) - Inside a macro time window (ICT macro times: typically on-the-hour or specific minute marks where algorithmic reactions are expected) When ALL THREE elements align — (1) a clear PD array, (2) confluence with another supporting PD array at the same level, AND (3) time anchoring at a session open, opening range, or macro time — the result is what ICT calls a "loaded deal": the highest-probability trade setup in the methodology. ICT explicitly frames this as the product of 20 years of experience: "what I'm about to divulge to you is 20 years...of experience that led to this observation and realization." Open inefficiency rule (supporting principle): When a bullish FVG or open imbalance is NOT filled by subsequent price action, this is a sign of bullish strength — the market is so eager to go higher it will not offer a discount. Conversely, when bearish, if a bearish FVG is not traded into from below, that is bearish strength. Open (unfilled) inefficiencies indicate the algorithm's directional commitment. This supports identifying which PD arrays are most likely to be respected. IPDA data range rule: PD arrays are valid for review up to 60 calendar days back. The three standard lookback periods are 20, 40, and 60 calendar days. Within those ranges, arrays at confluence with other arrays and at time-anchored moments are the highest-priority candidates.
Identification6
- Step 1: Apply IPDA data ranges — mark significant PD arrays from the past 20, 40, and 60 calendar days.
- Step 2: Look for price levels where two or more PD arrays CONVERGE in very close proximity (within a few ticks or a small percentage of ATR).
- Step 3: Grade each PD array using quadrant levels (25%, 50%/CE, 75%). Confluence at a quadrant level of one array AND the body/boundary of another array at the same price is a highest-priority zone.
- Step 4: Note the time context when price is approaching that confluence zone. Is it at a session open? In the first 30 minutes of the opening range? Inside a macro time window?
- Step 5: If all three conditions (clear PD array + confluence + time anchoring) align — this is a loaded deal setup.
- Open inefficiency indicator: if price has NOT filled a nearby FVG or suspension block in the direction of the expected trade, treat that as confirmation of directional strength and add weight to the confluence trade.
Entry3
- Enter at the confluence PD array level during the time-anchored window using any valid ICT entry trigger (FVG on lower timeframe, displacement, structure shift).
- The entry itself should be at the most precise level within the confluence zone: prioritize the CE of an FVG or suspension block, or the quadrant level of the wick where confluence is occurring.
- Do not enter simply because a PD array is present — wait for the time anchor AND at least one supporting array at the same level.
Stop2
- Below the lowest boundary of the confluence zone (bullish) or above the highest boundary (bearish).
- If entering at the CE of a suspension block or FVG within the confluence zone: stop below the low of the entire confluence zone.
Target2
- The draw on liquidity identified from the next higher reference level — old highs (buy-side), old lows (sell-side), or the next premium/discount array in the direction of the trade.
- When the market has open inefficiencies above (bullish) that have not been filled, those open FVGs become the minimum target: price is expected to trade into them.
Invalidation4
- Only one PD array present with no confluence from another array — reduces probability; look for confirmation from time anchoring at minimum.
- The PD array is approached outside of a session open, opening range, or macro time — no time anchor weakens the setup.
- The open inefficiencies in the bullish direction have been filled (price traded down into them and closed them) before the trade — suggests reduced bullish strength; reassess bias.
- PD arrays older than 60 calendar days — outside the IPDA valid lookback range; discount their significance.
Inferred Conditions (Unvalidated)
- The 'loaded deal' is a convergence of: (1) clear PD array, (2) second supporting PD array at same level, (3) time anchor. Three-element alignment dramatically increases probability over any single element alone.
- The most powerful confluence combinations are: suspension block CE + FVG boundary; wick quadrant level + order block; breakaway gap + session open FVG. Any two-array convergence at a time-anchored moment qualifies.
- ICT's reference to 'lead pipe sense trade' (Larry Williams) indicates this is the highest-conviction trade type in the methodology — one where the probability is so overwhelming that missing it or losing it requires exceptional circumstances.
ICT Quotes
""the ones that are so obvious and they converge and layer top of or in very, very close proximity to another supporting PD array, whether it be an inefficiency or a gap like the Wix, if it's laying on top of a quadrant level like that, and it's also time anchored, nor if it's time for it to move, it's at a session open, it's during the first 30 minutes of the opening range, or it is inside of a macro time. Man, you have a loaded deal.""
""what I'm about to divulge to you is 20 years...of experience that led to this observation and realization.""
""I have my PD arrays back as far as 60 days.""
""leaving inefficiencies open like this is exactly what you want to see if you're bullish...the market is so overzealous to move higher, it need not even concern itself with coming back down in here.""
Timeframes
Version History3 versions
ICT YT - 2025-10-12 - Advanced ICT Liquidity Concepts 10-11-2025.srt
""I have my PD arrays back as far as 60 days.""
Establishes the IPDA data range rule: PD arrays are valid up to 60 calendar days (three ranges: 20, 40, 60 days). Also establishes the open inefficiency rule: FVGs not filled during a directional move signal strength in that direction. Breakaway gap definition clarified: specifically forms between two quadrant levels during a trending breakout move. Measuring gap (halfway-point gap) distinguished from breakaway gap.
ICT YT - 2025-10-14 - ICT Secret To Selecting Algorithmic PD Arrays 10-14-2025.srt
""if it's laying on top of a quadrant level like that, and it's also time anchored...it's at a session open, it's during the first 30 minutes of the opening range, or it is inside of a macro time. Man,…"
Full articulation of the three-element confluence selection framework: (1) PD array present, (2) second supporting PD array convergence at same level, (3) time anchored to session open / first 30 minutes of opening range / macro time. This is presented as the product of 20 years of experience and the definitive answer to "how does ICT know which PD array to use." Live demonstration on Gold (XAUUSD) and NQ futures. Also demonstrates suspension block as the confluence anchor during a retracement that held above the upper quadrant before launching higher.
03 - ICT 2026 Market Commentary ⧹ March 21, 2026.en.srt
"We have the arm wrestling match between two PD arrays. So right away does this indicate high probability? No, because both sides could be justified. When it is one-sided and the analysis using the con…"
CONFLICTING PD ARRAYS = LOW PROBABILITY. ICT explicitly defines the inverse case: when two PD arrays on the same chart argue for opposite directions (e.g., an inversion FVG arguing bearish vs. a SIBI arguing bullish), the setup is NOT high probability — "both sides could be justified." High probability requires the analysis to be ONE-SIDED with no credible alternative interpretation. This is the first explicit articulation of the negative filter: when PD arrays conflict, probability drops and the trader should NOT force a trade. Demonstrated on Euro dollar daily chart where a bearish SIBI and a potential bullish inversion FVG were in direct conflict.
Notes
This concept provides the meta-framework for all PD array usage in the ICT methodology. Individual PD array concepts (FVG, order block, suspension block, etc.) each have their own YAML entries. This entry documents the SELECTION criteria — the filter that distinguishes high-probability arrays from noise. The three IPDA lookback windows (20, 40, 60 days) correspond to roughly one month, two months, and three months of calendar days. ICT notes he maintains PD arrays back 60 days, which some students find excessive — his response: experience teaches which levels within those 60 days are most relevant (the confluence + time-anchor filter is exactly that experience codified). The "open inefficiency = strength" principle is a corollary: if you are bullish and a prior FVG has NOT been filled, that is STRONG confirmation — the market does not want to go down. Conversely, a fully-filled FVG on the way up (while bullish) reduces conviction and requires a new confirmation structure before re-entry. Larry Williams reference: ICT attributes "lead pipe sense trade" terminology to Larry Williams' 1970s trading books, meaning a trade so obvious it is almost certain. This is the quality ICT is aiming for with the three-element confluence. See also: htf-pd-array-hierarchy.yaml, suspension-block.yaml, fair-value-gap.yaml, macro-time.yaml, ipda-data-range.yaml, breakaway-gap.yaml
Asymmetry Notes
The framework is symmetrical for both bullish and bearish setups. In bearish context: look for confluence PD arrays in premium territory (overhead resistance); open inefficiencies that have NOT been traded into from below indicate bearish strength (market will not offer a discount). The same three-element rule applies: confluence + time anchor = loaded deal.