ICT Bread & Butter Buy Setup — Bullish 4-Session Intraday Model

Bread & Butter is the most "workmanlike" model in the ICT catalogue. No hunting for the trade of the week: instead, a map of an ordinary bullish day, laid out across four sessions, each with its own script, its own hour and a target of somewhere between a dozen and a few dozen pips. It's a model for the trader who wants to understand the whole day, not just one window. In this part we cover the bullish variant (the buy setup); the mirror-image bearish variant is covered in the article on the Bread & Butter sell setup. As always: with the times in ET, and without pretending that 2–3 setups a day equals 2–3 wins a day.

What is the Bread & Butter buy setup

Bread & Butter buy is the template for an up day. When higher-timeframe institutional flow (order flow) points to higher prices, the typical day unfolds like this:

• the day opens at or near the low of the daily range; a small dip below the opening price is possible,
• London prints the first intraday leg up and forms the low of the day,
• New York continues the rally until around 10:00 AM ET,
• between 10:00 AM and 12:00 PM ET the high of the day forms, after which price usually corrects lower.

Where does the fuel for the rally come from? According to ICT theory, the price-delivery algorithm (IPDA) runs one of two price engines in a bullish program — mechanisms that manufacture discount for buyers before the real move launches:

Offset-accumulation — the market is marked down below an old low to trigger the sell stops of existing long holders. This manipulation "manufactures" sellers at a deep discount — and their supply is paired against smart money's buy orders. On the chart: a sweep of the low and a dynamic recovery.

Re-accumulation — the market pulls back into a fair value zone (an FVG, an order block) without breaking the low. The pullback hurts current longs just enough for some of them to sell — and again: their supply is the liquidity underneath institutional entries.

Recognising which engine is running tells you where to wait: offset-accumulation — below the old low; re-accumulation — at the discount PD Array.

The model's parameters are deliberately modest: a position usually lasts 1–2 hours, the target is 15–30 pips (50 at most), the execution timeframe is the M5, risk-reward around 1:1, risk 0.5–1% per trade. There tend to be 2–3 setups a day, one per session. It's a model of repetition volume, not one-off fireworks — and that's exactly why risk discipline matters more here than in models that fire once a week.

The four sessions — conditions step by step

Session 1 — London (opens 12:00 AM ET, entry 5:00–7:00 AM ET). With bullish order flow, expect London to form the low of the day. After the open, a protraction lower is possible — the classic London Judas Swing. The proper entry window is the pullback between 5:00 and 7:00 AM ET: when price taps a discount PD Array, you look for a long on the M5.

Session 2 — New York (8:20 AM ET). If London confirmed institutional sponsorship on the long side and set the low of the day, NY usually continues the rally — unless a higher-timeframe premium array has already been reached intraday. At 8:20 AM ET (the CME open), watch for the New York Judas — a dip into intraday discount — and fade it. Target: the nearest premium array from the H1/H4/D1.

Session 3 — London Close (10:30 AM–1:00 PM ET) — note: this is a trade AGAINST the trend of the day. When London and NY have moved up in agreement, a higher-timeframe premium array has been reached and it's at least 10:30 AM ET — expect a correction from the high of the day. The signal: a failure swing on the M5 at the high plus a bearish order block. Risk: only ~10 pips above the high of the day; target: 20–30% of the daily range. This is the hardest setup of the model — rare, and easy to confuse with continuation.

Session 4 — Asia (opens ~7:00 PM ET). With bullish order flow, a long at or just below the 0 GMT opening price, expecting a 15–20 pip expansion higher while the Asian Range forms. Always target only those 15–20 pips — the Asian range is limited by nature. For a trader in Europe this is the middle of the night; realistically — one to skip, or to automate with alerts.

The common skeleton of every entry: confirmed bullish order flow on D1/H4 → a discount zone below price → an IPDA engine (a sweep of the low or a pullback into the zone) → an MSS on the M5 → entry on the retest → stop beyond the discount extreme with a buffer → 15–30 pips and done.

A worked example

GBP/USD, bullish order flow (the daily above a breaker, the H4 building higher lows). The day starts quietly; after the London open, price slides below the opening price and at 3:40 AM ET breaks the low of yesterday's Asian session by a few pips — offset-accumulation in action. The push down leaves an FVG on the M5.

Between 5:00 and 6:00 AM ET price pulls back into that gap. The M5 prints an MSS higher — long entry on the retest, stop 4 pips below the bottom of the sweep. After 70 minutes the position collects 25 pips at the equal highs from the night — closed as per the model, without waiting for it "to really run."

At 8:20 AM ET... the same thing, only higher: the NY Judas drags price into intraday discount, the M5 shows an MSS, the second trade of the day rides to a premium FVG from the H1 — another 20 pips. The high of the day forms at 10:45 AM ET; the London Close setup doesn't produce a clean failure swing this time, so there is no third trade. Two trades, +45 pips in total, day closed before 11:00 AM ET. That's what a model day looks like — and alongside it exist days when London doesn't set the low, NY doesn't continue and the only win is a small stop.

Note the proportions: both trades together lasted just over two hours, and the decisions fell inside windows known in advance — the rest of the day is observation, not clicking. That rhythm, not any single result, is the essence of the model.

What to trade it on and how to do the risk math

The model was born on forex and that's where it feels best: GBP/USD, EUR/USD, USD/JPY, USD/CAD plus gold — instruments that respect the four session windows. On the NQ and ES indices the sequence occurs as well, with a particularly clean New York window around 8:20 AM ET and the 9:50–10:10 AM ET macro slot. On crypto things are slipperier: BTC and ETH have a clear US-session rhythm, but an "Asian session" or a "London Close" on a 24/7 market is a template you first need to verify against your own instrument's data — without that verification you're importing statistics nobody has ever computed for crypto.

The model's math forgives little. At 1:1 risk-reward, the breakeven point is a win rate comfortably above 50% after costs — and the spread and slippage on the M5 can eat 1–2 pips per trade, which against a 15–30 pip target is a real percentage of the result. That's why two parameters are non-negotiable: 0.5–1% risk per trade (because 2–3 entries a day compound exposure quickly) and a journal split by session. After a few dozen plays you'll see which window actually pays and which merely adds commissions — for most traders the edge concentrates in one or two windows, and the rest get traded out of habit.

Most common mistakes

• The wrong time zone. All the windows are in New York time. A chart on your local time or the broker's server time = wrong windows = a different (non-existent) model.
• Trading without a confirmed bullish bias. Without bullish order flow from D1/H4, every pullback looks like an opportunity. The buy setup is traded exclusively on bullish days — day selection is half the model.
• Skipping the MSS on the M5. The discount zone is a location, not a signal. Touching discount without a shift in structure is catching a falling knife.
• Ignoring the session boundaries. A "London" trade taken at 12:00 PM ET is outside the window — different liquidity, different statistics. The clock is part of the edge.
• Greedy targets. 15–30 pips is the definition of the model. Holding a scalp "for 100 pips" turns it into a swing trade without a plan — and usually gives the profit back in the correction after 10:00 AM ET.
• Trading the London Close every day. That setup requires two conditions at once: London and NY moving in agreement, and a higher-timeframe premium array already reached. Without both — it doesn't exist.
• Forcing the Asian session from Europe. Entering at the local middle of the night, in a limited range with thin liquidity, costs most people more (in sleep and mistakes) than it earns. The model doesn't require you to trade all four windows.

Bread & Butter buy teaches something no single setup can: reading the whole day as a sequence of sessions, each with its own role. Before you try it live, make sure you've mastered the killzones, the premium and discount zones and the Judas Swing — and then test the model on your own instrument and your own journal. A 1:1 statistic at 15–30 pips forgives very few execution errors; this is a model for a craftsman, not an adrenaline hunter. The variant for down days is covered in the article on the Bread & Butter sell setup.

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