Advanced Forex · ICT · Lesson 11
ICT Power of Three (AMD)
Learn how the ICT Power of Three framework describes accumulation, manipulation and distribution. Use it as a chart-reading hypothesis, not a guaranteed trading signal.
Learning objectives
- Explain the three AMD phases.
- Identify a candidate range, liquidity sweep and expansion on a chart.
- Compare bullish and bearish scenarios.
- Define invalidation and position risk before considering a trade.
1. What is Power of Three?
Power of Three, often shortened to AMD, is an ICT framework that organizes a potential price move into three phases: Accumulation, Manipulation and Distribution. Real markets do not always follow this sequence; labels can only be assigned with uncertainty.
Price consolidates inside a recognizable range. Mark the range high and low, and note session context.
Price briefly trades beyond one side of the range, potentially taking resting liquidity. A breakout is not automatically a trap.
Price may expand directionally away from the sweep, sometimes leaving displacement or an imbalance.
2. Bullish AMD example
- Mark an Asian-session or other clearly defined consolidation range.
- Observe price sweep below the range low. Do not assume every sweep will reverse.
- Look for a shift in short-term structure and decisive upward displacement.
- Only then evaluate a potential retracement entry, stop location below invalidation and a logical upside target.
3. Bearish AMD example
- Identify a range and its high.
- Price trades above the range high, then falls back inside.
- Check for downward displacement and a bearish structure shift.
- Plan invalidation above the sweep high and evaluate nearby sell-side liquidity as a possible objective.
4. Context and confirmation
Compare the suspected AMD setup with higher-timeframe bias, the relevant session, nearby external liquidity, displacement, market structure and economic-news timing. Avoid interpreting every range breakout as manipulation. A pattern recognized only after the move may not have been tradable in real time.
5. Risk management
- Define the entry condition, stop and invalidation before a trade.
- Size positions so a stop-out stays within a pre-set small fraction of account equity.
- Account for spread, slippage and major announcements; a stop order cannot guarantee an exact fill.
- Skip trades where the setup is ambiguous or the reward-to-risk is unattractive.
6. Chart practice
Exercise A: On a historical M15 chart, mark a consolidation high/low. Record whether price swept a boundary and whether a genuine structure shift followed.
Exercise B: Find one breakout that continued rather than reversing. Explain why calling it manipulation would have been incorrect.
Exercise C: Write an entry hypothesis, invalidation, stop distance and target. Paper-trade only; log the result and alternative explanations.
7. Knowledge check
Power of Three: distinguishing a model from hindsight
The Accumulation-Manipulation-Distribution model describes a possible sequence of range formation, a temporary move beyond that range, and an extended directional move. It is a framework for organizing observations, not a claim that every trading day follows three clean phases. Many sessions trend without a convincing manipulation phase; others remain choppy all day. A trader should explicitly label such sessions as no setup rather than forcing the pattern onto them.
Begin with an observable accumulation range: record the session window, range high, range low, and whether volatility was contracting. A later push outside the range may be a liquidity raid, a genuine breakout, or an ordinary reaction to news. To distinguish them, look for a return into the range, displacement in the opposite direction, and a meaningful market structure shift. Do not assume that a wick beyond a boundary proves manipulation or reveals anyone's intent.
Imagine a bullish example where an overnight range spans 1.2500 to 1.2530. Price briefly drops to 1.2485, then reclaims 1.2500 and closes above a lower-timeframe swing high. A trader studying the model may call the drop a possible manipulation phase and the later rally a possible distribution phase. A simulated entry should be planned only after confirmation, with invalidation below the sweep extreme and a target based on a real opposing liquidity level. Check the available reward relative to the stop distance before considering any trade.
In a bearish example, price runs above a range high and then falls back inside with strong downward candles. A retracement toward a bearish imbalance may offer a research entry, but the setup is invalidated if price reclaims the sweep high under the chosen rules. If the market continues upward instead, the apparent manipulation was not a successful reversal setup. Record the failure without changing the labels afterward.
For backtesting, select thirty sessions and mark range boundaries before viewing later candles. Categorize each as bullish AMD, bearish AMD, breakout continuation, or no clear pattern. Record confirmation timing, stop size, target distance, session, and whether high-impact news occurred. Calculate average reward-to-risk and the proportion of ambiguous cases. The purpose is to test whether your precise version of the model has merit, not to claim certainty from a memorable chart.