The Reference the Algorithm Delivers Price Against
Three lookback ranges in daily candles — and the one question they answer: are you buying expensive or cheap?
Before you start: Only needs a daily chart and the idea of premium and discount — nothing before it.
Part of the My Edge Official framework — refined through thousands of hours of chart study.
IPDA — the Interbank Price Delivery Algorithm — is ICT's name for what actually moves price between pools of liquidity. It does not read your setup. It delivers price from where liquidity has been taken to where liquidity still sits, and it references how price has been delivered recently. The 20, 40 and 60 trading-day ranges are how you see that reference on your own chart.
- Core Idea
- Why It Exists
- How Professionals Read It
- Practical Application
- Common Mistakes
Core Idea
Three lookback ranges, all counted in daily candles and all counted backward from the same place: the first trading day of the previous month. Twenty days back, forty days back, sixty days back.
Each range gives you three levels and nothing more: its high, its low, and the equilibrium exactly halfway between them. Above equilibrium is premium. Below equilibrium is discount. That is the entire construction.
What you do with it is one question: where is price right now, relative to each of those three ranges?
Why It Exists
ICT's argument is that price is delivered algorithmically, and that the algorithm works from a reference — not from your opinion about trend. The data ranges are that reference made visible. They answer two things no single candle can:
Where price has already traded. Sixty days of delivery, framed. The high and the low of that stretch are where liquidity was taken, which is what the algorithm was reaching for.
Whether you are buying expensive or cheap. Discount and premium are not sentiment. They are a measured position inside a measured range. In discount you are looking to buy; in premium you are looking to sell. A setup that ignores which half it formed in is a setup without context.
How Professionals Read It
Read one range at a time. Three overlapping boxes on one chart teaches nothing — mark the 20, read it, then the 40, then the 60.
The three will often disagree, and that disagreement is the information. Price sitting in premium of the 20-day range while still in discount of the 60-day one is a specific, common and very readable condition: expensive against the last month, cheap against the last quarter. The reverse — discount of the short range, premium of the long one — is its mirror.
When they agree, the read is simply stronger. Deep discount of all three, at a level the algorithm has reason to reach for, is as clean as this framework gets.
Practical Application
Inside each range, mark which PD arrays sit in its discount and which sit in its premium. An Order Block in the discount half of the 60-day range is a different proposition from the same Order Block sitting in its premium half, and the array itself looks identical in both cases. The range is what tells them apart.
💡 The count is in trading days, not calendar days. Weekends and holidays produce no daily candle, so they are not counted. Counting calendar days is the single most common way this gets marked wrong.
Equilibrium is a level in its own right, not just a dividing line. Price delivered back to the 50% of a range it has left is one of the most frequent things this framework shows you in advance.
Common Mistakes
- Starting the count on the wrong day. Not today, and not the first of the current month — the first trading day of the previous month. Start it a month late and all three ranges are wrong together, which is worse than not marking them at all.
- Marking all three and then reading none of them. Three boxes drawn once and never referenced again is decoration. The ranges are a question you ask at every significant delivery, not furniture on the chart.
- Treating the longest range as the most important. Sixty days carries the most history, which is not the same as carrying the most relevance today. Which range matters depends on what you are trading and when.
- Forgetting that the ranges move. They are fixed for the month, and then they are not. A range that framed last month's delivery perfectly may frame this month's poorly.
How to Use It
In practice, on the daily chart:
- Find the first trading day of the previous month and mark it. Everything else is counted from this candle.
- Count back 20 daily candles from it. Mark that stretch's high and low, then the equilibrium halfway between them.
- Do the same for 40, then for 60. Keep each one on its own — turn the others off while you read it.
- For each range, answer one question: is price in premium or in discount of this range right now?
- Inside each range, mark which PD arrays sit in the discount and which in the premium. That is your shopping list, and which half it came from is part of the entry.
- Re-check the answers at every significant delivery. The levels stay put for the month; where price sits relative to them does not.
There is also a rolling version of the same reading — counted from today rather than from a calendar anchor, over the last 20 daily candles and the last 3. It moves forward with every new day, which makes it the shorter-term companion to these ranges rather than a replacement for them: see Dynamic IPDA 3/20.