ICT Weekly Range Expansion — The Tuesday Open Model and Three PD Arrays

Most intraday traders look at the weekly candle once — on Sunday evening, out of curiosity. Yet that candle is exactly what one of the simplest ICT models is played on: the weekly range expansion. The idea is surprisingly concrete: if the week is going to expand in one direction, Monday usually prints its extreme, Tuesday starts the real move, and Thursday finishes it. Instead of chasing every hourly wiggle, you enter once — at the Tuesday open — and let the weekly candle work for you. In this article we take the model apart: the bias, the three PD Arrays, the exact entry time in New York time, and the mistakes that break the whole thing.

What the weekly range expansion model is

The weekly range expansion is a model from ICT's charter material, built on a single premise: a weekly candle that has a bias doesn't grow evenly across five days — it unfolds in waves, and the most predictable wave runs from Tuesday to Thursday. The model has three stages:

1. STAGE — the direction of the week. On the W1 chart you determine where the weekly candle wants to go: after liquidity or into an unfilled FVG.
2. SETUP — an expansion getting under way. The week has to be an expansion week, not a consolidation week. Monday builds the extreme from which the move will launch.
3. PATTERN — the execution. An entry at the open of Tuesday's candle at 4:00 AM New York time, targeting the opposite PD Array.

All the navigation happens on just three types of PD Array — and that is this model's greatest strength. Instead of the entire PD Array Matrix, you use only:

• a Fair Value Gap — an unfilled gap on W1/D1 that price wants to come back to,
• an old high or low — an extreme with orders resting above or below it,
• a liquidity pool — a cluster of stop-losses at equal highs/lows.

These three elements play a double role: once as the justification for the bias (where price came from), and once as the target (where it's heading).

Step 1: the weekly bias

Without a correct bias, the model doesn't exist — the Tuesday open is then just a random hour. You read the bias off the weekly chart by asking two questions:

Bullish bias: did the previous W1 candle collect sell-side liquidity (i.e. break an old low)? Or is price sitting in the discount zone, the lower half of the current range? If so — the natural magnet is buy-side liquidity and PD Arrays in the premium zone.

Bearish bias: the mirror image — the candle collected buy-side (broke an old high) or is sitting in the premium zone of the upper half of the range. The magnet: sell-side and PD Arrays in discount. How exactly to mark those zones is covered in the article on premium and discount.

An extra confirmation is a Market Structure Shift on the daily timeframe in the direction of the bias. It's not mandatory, but it filters out the most false weeks. If after this analysis you still can't say in one sentence where the week is going — it's not a week for this model.

Step 2: the role of Monday

In this model Monday is not a trading day — it's an observation day. Statistically, Monday is the day most likely to form the low of the week in a bullish week and the high of the week in a bearish one — the market first manipulates against the direction, collects liquidity, and only then moves the right way. It's the same logic that Power of 3 describes at the scale of a single day, and that the weekly profiles describe at the scale of the week: the weekly range expansion is, in effect, a play on the most common of them.

Practically: on Monday you only note where the extreme forms and check whether the price behavior agrees with the bias. The Monday low (in the long scenario) later becomes the natural reference level for the stop.

Step 3: the entry at the Tuesday open

The actual execution falls at the open of Tuesday's candle at 4:00 AM New York time — and that means New York time specifically, not your broker's server clock (if you're converting from another time zone, see our guide to market session hours). That's the middle of the London part of the day: the market already has liquidity, and the weekly expansion usually hasn't picked up speed yet.

The entry rules are simple:

• Long (bullish week): you buy at the opening price of Tuesday's 4:00 AM ET candle or lower. Every dip below the open is a better price in the direction of the bias — not an obstacle, but an opportunity.
• Short (bearish week): you sell at the opening price or higher.
• Stop: in the ICT original — 50 pips from entry, calibrated for the major forex pairs. On other instruments, convert that into a percentage of the weekly range; on BTC or indices a rigid "50 pips" makes no sense — a sensible measure is a fraction of the average weekly range or the ATR.
• Target: the PD Array aligned with the bias — the liquidity pool above an old high and premium zones (long), or the liquidity below an old low and discount zones (short).

Step 4: holding into Thursday

The expansion window is Tuesday through Thursday. You hold the position at least until the New York session open on Thursday (the 8:30–9:30 AM ET window) — that's the stretch in which the weekly candle usually delivers most of its range. Friday is a different story: the market more often takes profit and retraces back into the weekly range (the profile ICT calls TGIF), so holding a position "because it'll keep going" exposes your profit to unnecessary risk. The model has a clear ending — and that too is one of its virtues: one entry decision, one exit decision, zero micromanagement along the way.

Example: a textbook bullish week, step by step

Here's how it looks in practice on a hypothetical EUR/USD week. Over the weekend you do the W1 analysis: the previous weekly candle broke an old low (collected sell-side) and closed with a long lower wick in the discount zone. One floor up, on D1, you see an unfilled FVG a few dozen pips above price and equal highs even higher — you have a bullish bias and two targets. On Monday price indeed slides lower still — you don't panic, because that's exactly what the model expects: Monday builds the low of the week. You note its level.

On Tuesday the 4:00 AM ET candle opens. Price sits slightly below Monday's close — you go long at the opening price, place the stop below the Monday low (it fits within the model's limit), and set the target at the daily FVG. Wednesday delivers the first leg of the expansion; on Thursday, ahead of the NY open, price taps the gap — you close the position according to plan. On Friday price retraces a third of the weekly range, and had you "held a little longer", you'd have given back a big chunk of the profit. One analysis, one entry, one exit — that's an entire week's work in this model.

Weekly range expansion on crypto and indices

The model was born on currency pairs, but indices and crypto respect the weekly cycle too — with two adjustments. On indices (NQ, ES) a strong execution window is also the Tuesday cash open at 9:30 AM ET — that's where real exchange volume comes in. On BTC and ETH the week technically starts Monday 00:00 UTC, but the institutional rhythm is still set by the New York clock, so the entry is played in the same Tuesday-morning NY window; you derive the stop from the weekly ATR instead of pips, and you treat the thin weekend price action as noise, not signal. The bias requirement stays unchanged — and on crypto it filters out even more weeks, because consolidations tend to run longer.

The most common mistakes

• Entering on Monday. The most expensive mistake in this model. Monday prints the extreme of the week — whoever enters too early gets stopped out by the very move that was supposed to fuel the position.
• The wrong time zone. 4:00 AM means New York time, not your broker's server time or your local clock. Mixing up zones shifts the entry by hours and turns the model into a lottery.
• Trading every week. The model works only in expansion weeks. An inside week or a consolidation won't generate a Tuesday–Thursday move — and a bias forced into existence is not a bias.
• Holding into Friday. The expansion has its own timetable. After the Thursday NY open, the probabilities are already working against you.
• A rigid stop on every instrument. 50 pips on GBP/USD and "50 pips" on BTC are two different worlds. Scale the stop to the instrument's volatility, don't copy the number.
• A model with no bias. If W1 hasn't collected liquidity and isn't sitting in clear premium/discount — there's no stage, so there's no trade. For help with the day-to-day directional read, see the article on daily bias.

The weekly range expansion is a good first "higher timeframe" model for a trader who has so far lived exclusively on M5: it forces you to read the weekly chart, teaches the patience of a single entry, and shows that time — not just level — is half the edge. Before you play it with real money, test on your own market how many weeks per quarter actually meet the model's conditions. The result will surprise you: fewer than you think — and that's exactly why it works.

FAQ

Rafał — Krypto Bez Ściemy

Trader and founder of Strefa Tradingu. He’s been breaking crypto down on YouTube for years — no hype, no signals, with a focus on market structure and risk management.

kryptobezsciemy.plInstagramYouTubeKlub Zero