ICT Power of Three: accumulation, manipulation, distribution

The power of three is an ICT model that describes each trading day, or any session, as three phases: accumulation, manipulation and distribution (AMD). Price builds a range, makes a false move against the real direction, then moves in the real direction. Traders use the model to look for a trade after the false move.

In short

• The power of three has three phases: accumulation, where price forms a range, manipulation, a false break of that range that runs stops, and distribution, the sustained move in the other direction.
• On a daily candle it shows up as a candle that opens, moves against its eventual direction, then closes the other way. In forex, the Asian session is usually treated as accumulation, the London open as manipulation and the New York session as distribution.
• Traders use the daily open as a reference. In a bullish day they look to buy below it after the manipulation, and in a bearish day they look to sell above it.
• The pattern is far easier to see afterwards than to trade live. Many days fail to follow it, and a false break can turn into the real move.
• On a $25,000 Direct account, a 16-pip stop on 1.56 lots of EUR/USD risks about $250, a quarter of the $1,000 daily loss limit.

What is the ICT power of three?

The ICT power of three is a model that says the market moves in three stages in a day or a session. It is also called AMD, for accumulation, manipulation and distribution. Michael Huddleston, the Inner Circle Trader (ICT), uses it to explain why price often makes a false move early in a session and then runs the other way.

The name is slightly misleading, because the three things are phases and not three trades. Accumulation is the period in which price moves sideways and the range builds. Manipulation is a sharp move out of that range in the wrong direction, which takes out stops and tempts breakout traders. Distribution is the move that follows, which carries price well past the range in the intended direction.

The explanation given is that large participants need liquidity to fill big orders. They build a position quietly in the range, push price to where stops and breakout orders sit so they can take the other side, then let price run. Nobody outside those firms can see their orders, so the story is a way to read price, not a fact about who is trading. The practical form is simpler: the first break of a range is often a trap, and a trade after it can have a nearby stop.

What are the three phases?

Accumulation is a range with no clear direction, manipulation is a false break that quickly reverses, and distribution is the lasting move in the opposite direction. In an up day the manipulation is a drop below the range, and in a down day it is a rise above it.

Accumulation looks quiet. Candles overlap, the highs and lows get tested but not held, and both sides of the range collect stop orders. On a forex chart this is usually the Asian session, roughly 8:00 pm to midnight in New York or 1:00 to 5:00 am in the UK in ICT's teaching, which is the same range described in the ICT trading guide.

Manipulation is the part that gives the model its trade. Price runs through the high or the low of the range, often at the London open or the start of the New York session, and then comes back inside. It is a liquidity sweep of the range. In a model bullish day, price drops below the Asian low, takes the sell stops, and recovers.

Distribution is the run to the target. In the model it carries price through the other side of the range and on to the next pool of liquidity, such as the previous day's high or low. It often comes with large candles and fair value gaps, the gaps described in the fair value gap guide.

How does the power of three look on a daily candle?

On a daily candle, the open is the start of accumulation, the wick against the final direction is manipulation, and the body that closes the other way is distribution. A bullish day opens, drops below the open, then closes above it. A bearish day opens, rises above the open, then closes below it.

This is why traders mark the day's opening price. In a bullish day the manipulation happens below the open, so a buy below the open is a buy at a discount to the day's start. In a bearish day the manipulation is above the open, and a sell above it is at a premium. The reference works as a bias filter, in the same spirit as the premium and discount ideas in the smart money concepts guide.

Which open counts is a point of difference. Many ICT traders use midnight New York time as the true day open, not the 5:00 pm New York close that many forex platforms use for the daily candle. The two can differ by many pips, and a daily candle drawn on the platform's default will not match the model's. Check your chart's time zone and candle open before you draw anything.

The pattern is a description of what the day looked like at the close. At 8:00 am it is a hypothesis. A drop below the Asian low at 8:00 am might be the manipulation of a bullish day, or the start of a bearish one, and nothing on the chart tells you which. The model gives you a place to wait for evidence, such as a recovery above the Asian low or a break of structure, not a forecast.

Which sessions match accumulation, manipulation and distribution?

In the usual forex reading, the Asian session is accumulation, the London open is manipulation and the New York session is distribution. The same three-part idea is also applied to single sessions, to a week and to an hour, which means any chart has several overlapping examples.

These times are a convention, not a measured pattern, and teachers shift them. Sometimes London is the accumulation and New York the manipulation, and sometimes the move is spread across both. The ICT Silver Bullet guide uses one-hour windows inside this structure, and the model also works on the 1-hour chart with each hour as a mini-cycle, where it is much less reliable because small ranges are full of noise.

One useful check is whether the previous day's high or low was taken. Traders see the manipulation as more credible when it sweeps a level that was visible beforehand, like an Asian high with equal highs beside it, than when it simply drifts a few pips past a quiet range.

How do you trade the power of three?

To trade the power of three, mark the accumulation range, wait for a run through one side of it, look for price to close back inside the range and break structure, then enter with the stop beyond the manipulation extreme and a target at the next liquidity.

A workable sequence:

1. Mark the day's open and the Asian range. Note the previous day's high and low.
2. Form a bias, from the higher timeframe trend or a premium and discount reading, so you know which side you would prefer to trade.
3. Wait for the sweep: a run through one side of the range, ideally through the side that fits your bias.
4. Wait for the return. A candle closes back inside the range, and a lower timeframe breaks structure, as in the break of structure guide.
5. Enter on a pullback to a gap, a block or a retracement, such as the zone in the optimal trade entry guide.
6. Put the stop beyond the manipulation extreme, and the target at the other side of the range or the previous day's high or low.

The model also gives a plain reason to stand aside. If price breaks the range and keeps closing beyond it, the day is trending from the open and the "manipulation" is the real move. Without the return inside the range, there is no setup.

What does a power of three trade look like with real numbers?

This example is a short trade on a 5-minute EUR/USD chart, for a $25,000 account risking 1% per trade, which is $250. One standard lot of EUR/USD moves $10 per pip. The prices are made up for the illustration.

1. Midnight New York time, the day's open, is 1.0835. The Asian range runs from 1.0820 to 1.0842, 22 pips. The previous day's low is 1.0775, and the previous day's high is 1.0880.
2. The 4-hour chart has turned down. The bias is bearish, so the manipulation you want is a run above the range, in the premium zone above the open.
3. At the London open, at 3:10 am New York time, price runs to 1.0856, 14 pips above the Asian high, then closes a 5-minute candle at 1.0838, back inside the range.
4. A 5-minute candle then closes at 1.0817, below the Asian low of 1.0820 and below the last minor swing low. This is the break of structure.
5. The move leaves a bearish fair value gap from 1.0841 to 1.0845. The entry is a sell limit at 1.0843, the gap's midpoint.
6. The stop goes at 1.0859, 3 pips above the manipulation high. The risk is 16 pips.
7. Size: $250 divided by (16 pips × $10) is 1.56 lots after rounding down. Each pip is $15.60, so the risk is 16 × $15.60 = $249.60.
8. Set the target at 1.0780, 5 pips above the previous day's low. That is 63 pips, so the reward at the target is 63 × $15.60 = $982.80, about 3.9 times the risk.

The reward to risk is high because the stop is close and the target is far. That is also what makes the setup rare. If the move to 1.0856 had been followed by a close at 1.0870, above the stop, the sweep would have been the real breakout, and the loss would be $249.60. If price had never come back to 1.0843, no trade is placed. Simulated trades carry trading costs, as they would on a live account: a commission when a trade opens and when it closes, a price markup on some symbols, and holding costs on positions kept overnight.

How this works on CMC Funded

On a $25,000 Direct account the daily loss limit is $1,000 on the first day, 4% of that day's starting equity, measured on equity. The maximum loss is 6%, fixed, so the floor is $23,500. The $249.60 trade above uses a quarter of the daily limit, and four stop-outs cost $998.40 before costs. On a $25,000 Classic account the limit is $1,250 and the floor $22,500, so the same trade uses a fifth of it.

The model gives two reasons for caution. The first is that the manipulation is a spike, and a spike is when floating loss is largest. If a second spike takes price back to 1.0856 after you are filled at 1.0843, you are 13 pips against, and at 1.56 lots that is $202.80 of floating loss, which counts against the limit at that price even if the candle closes back inside. The second is that a day with a failed setup is often a trending day, where it is easy to take a second and third loss in the same direction.

The limit is worked out afresh each day from that day's starting equity, and an unused allowance does not carry over. The pattern begins around midnight in New York, which is early morning in the UK, so many traders set an order and leave it. Holding overnight and over the weekend is allowed, with holding costs, and there is no time limit. News trading is allowed, which matters because US data often arrives during the distribution phase.

For scale, the Direct profit target is 10%, which is $2,500 on $25,000. The $982.80 winner covers about 39% of it. Direct needs at least 3 trading days. The position size calculator turns a stop distance into lots, and the rules page has the full definitions.

Is the power of three the same as Wyckoff accumulation and distribution?

No. The two share an idea, that price builds a range before a larger move, but they are different frameworks. Wyckoff describes multi-week phases with named events, and the ICT model describes a day or a session in three parts, with a sweep as its middle stage.

Wyckoff's accumulation and distribution schematics were written much earlier, and they cover a base that can last weeks or months. The ICT power of three compresses the idea to a daily rhythm and emphasises the stop run at the range's edge. Some traders see the similarity and use both. It is fair to say the two are alike in spirit and that neither has been shown to predict price reliably.

Does the power of three work?

There is no reliable evidence that days follow the accumulation, manipulation and distribution pattern often enough to trade on that alone, and no public win rate for it. The model is a way to read a chart, and on many days it fits only if you are generous about where each phase starts.

Looking back, almost any day can be sliced into three phases, and a range, a spike and a move are present on most charts. That is the central weakness: a model that fits every day tells you little on the day itself. Live, you cannot know that the spike is manipulation and not distribution until after it is over. The tests that matter are about the rules you add on top, such as requiring a visible sweep, a close back inside the range, a structure break and a minimum of room to the target.

With a 2:1 target you need to win more than one trade in three, before costs, to come out ahead. At 3.9:1, a little over one trade in five is enough, but a far target is reached less often. Recording 50 qualifying days on one symbol, with how many reached the target and how many hit the stop, will tell you more than any teacher's chart.

Common mistakes

• Reading the model onto every day. If there is no clear range, no sweep and no return, there is no setup, and forcing one is a common way to take trades the model never offered.
• Treating the first break of the range as manipulation. A break that keeps closing beyond the range is the move, and waiting for the return inside saves you from it.
• Using the wrong open. A platform's daily open at 5:00 pm New York time is not the midnight open the model uses.
• Taking manipulation against a strong higher timeframe trend. A run above the Asian high during a strong 4-hour uptrend is usually just the trend.
• Putting the stop inside the range. With the stop at 1.0848 on the example trade, 5 pips above the entry, the trade is stopped by ordinary noise before the move develops.
• Skipping the sizing because the target is far. A 16-pip stop at 1.56 lots is $249.60. If the stop is moved to 1.0889 to give it room, the loss is 46 pips × $15.60 = $717.60, 72% of a $1,000 Direct daily limit.

Questions traders ask

What does AMD mean in trading?

AMD stands for accumulation, manipulation and distribution, the three phases of the ICT power of three. It is the same model under a different name. Some traders also use it for the cycle on a single session or a single hour, as well as for a day.

Which timeframe suits the power of three?

No timeframe has been shown to be better. Traders read the daily candle for the shape of the day and then trade it on the 5-minute or 15-minute chart, which is where the Asian range, the sweep and the gap are visible. The 1-hour chart suits slower entries with wider stops.

Is the power of three bullish or bearish?

It is both. A bullish version opens, drops below the open and closes above it. A bearish version opens, rises above the open and closes below it. The model does not tell you which you are in, so the bias comes from the higher timeframe or from a premium and discount read.

Does the power of three work on indices and gold?

Traders apply it to indices and gold as well as forex pairs, using the session of the market they trade. The ranges, spreads and sweeps differ, so the stop sizes and the minimum room to the target need to be rebuilt and tested again for each symbol.

Next steps

The Academy lesson on sessions and kill zones shows how to mark the Asian range and the London and New York windows on a chart. The inducement trading guide covers the smaller traps that often appear inside each phase, and the breaker block guide shows a zone entry that suits the return after the manipulation. Size each trade from its stop with the position size calculator, and read the daily and maximum loss limits on the rules page.

You can compare the Classic and Direct routes on the challenges page.