Smart money concepts explained: order blocks, fair value gaps and liquidity sweeps

A fair value gap (FVG) is a three-candle pattern in which price moves so quickly that the high of the first candle and the low of the third candle do not overlap (or, in a falling market, the low of the first and the high of the third). The price range between them traded only briefly, in one direction, and smart money concepts traders treat it as an "imbalance" that price may come back to.

Smart money concepts (SMC) is the umbrella name for a family of chart-reading ideas that includes fair value gaps, order blocks, liquidity sweeps, break of structure and change of character. The terms are popular, often useful as description, and frequently oversold. This guide explains each one plainly, shows how they relate to older technical analysis, and sets out what can be tested and what cannot.

Where smart money concepts come from

Much of the current vocabulary was popularised by an online educator who trades under the name ICT (Inner Circle Trader), and "ICT concepts" is often used interchangeably with SMC. The ideas have since spread through social media, video channels and paid courses, and different teachers use the same words in slightly different ways. That inconsistency matters, and we come back to it below.

The central story behind SMC is that large institutions ("smart money") need liquidity to fill big orders, that this liquidity sits in clusters of retail stop-loss and pending orders, and that price is therefore engineered to sweep those clusters before making its "real" move. The chart patterns are presented as the footprints of that process.

Treat that story as a framework, not a fact. A retail candlestick chart shows prices and, at best, a broker's tick volume. It does not show who traded, why, or whether any particular move was "engineered". We will come back to how the currency market actually works after defining the terms.

The core terms, defined plainly

Market structure: swing highs and lows

Everything in SMC starts from swing points. In an uptrend, price makes higher highs and higher lows; in a downtrend, lower highs and lower lows. This is the same definition of trend used in classical technical analysis and price action trading.

Break of structure (BOS)

A break of structure is a move beyond the most recent swing point in the direction of the existing trend. In an uptrend, a candle closing above the last swing high is a bullish BOS; it confirms the trend is continuing. Some traders require a candle close beyond the level, others accept a wick. That choice changes the results and must be fixed before testing.

Change of character (CHoCH)

A change of character is the first break against the existing trend. In an uptrend, the first time price breaks below the most recent higher low, SMC traders call it a CHoCH and treat it as an early sign the trend may be changing. Classical analysts would describe the same event as the first lower low, or a break of the trend's last support.

Liquidity and liquidity sweeps

In SMC, "liquidity" usually means the stop-loss and pending orders assumed to sit just beyond obvious levels: above equal highs, below equal lows, beyond round numbers and beyond recent swing points. A liquidity sweep (also called a stop hunt, grab or raid) is a brief move through such a level that quickly reverses.

Older traders call this a stop run or a false breakout. The phenomenon is real enough: stop orders do cluster around obvious levels, as the stop-loss placement guide discusses, and when they trigger they become market orders that can push price a little further before it settles.

Order blocks

An order block is usually defined as the last opposite-coloured candle before a strong move that breaks structure. A bullish order block is the last down candle before a sharp rally that makes a BOS; a bearish order block is the last up candle before a sharp fall. The idea is that institutions placed orders in that candle's range and that price may return to it to "mitigate" unfilled orders.

Functionally, an order block is a small support or resistance zone drawn from a single candle. It closely resembles the supply and demand zones that traders drew long before the SMC label existed.

Fair value gaps

As defined above, a fair value gap is the price range between the first candle's wick and the third candle's wick in a fast three-candle move. Bullish FVGs form in rallies, bearish FVGs in declines. When price later trades back into the range, the gap is said to be "filled" or "rebalanced". Some traders also mark an "inverse" FVG when price closes through a gap and later respects it from the other side.

Older terms for the same idea include imbalances, single prints and inefficiencies. It is related to, but not the same as, a true price gap, where no trading at all happens between two bars, as at a weekend open.

Premium and discount

Premium and discount split a price range in half. Take a swing low and a swing high; the midpoint (50%) is "equilibrium", above it is the premium zone and below it the discount zone. The SMC rule of thumb is to look for buys in discount and sells in premium within a bullish or bearish context respectively. This is very close to the old idea of buying pullbacks in an uptrend, and the 50% level is the same midpoint that appears on many Fibonacci retracement tools.

An illustrative candle example, described in words

The following is an invented example on a 15-minute EUR/USD chart, with round numbers, to show how an SMC trader would label a sequence.

1. The setup. Over the morning, price makes two highs at almost the same level, 1.1050 and 1.1051. An SMC trader marks the area just above as "buy-side liquidity", on the assumption that sellers have stops there and breakout buyers have buy stops there.
2. The sweep. A candle spikes to 1.1058, then closes back at 1.1046. The equal highs have been swept: stops above 1.1051 were likely triggered, and price did not hold above the level.
3. The change of character. Over the next hour price falls below the most recent higher low at 1.1030, closing at 1.1024. In SMC terms that is a bearish CHoCH.
4. The order block. The last up candle before the sharp fall, with a range of 1.1040 to 1.1052, is marked as a bearish order block.
5. The fair value gap. During the drop, three consecutive candles form with the first candle's low at 1.1043 and the third candle's high at 1.1036. The 7-pip range from 1.1036 to 1.1043 traded only once, inside the large middle candle. That range is the bearish FVG.
6. The premium zone. Taking the swing from the 1.1058 high to a new low at 1.1010, equilibrium is 1.1034. The FVG and the order block both sit in the premium half of that range.
7. The trade idea. A typical SMC plan would be a sell limit inside the FVG or order block, a stop above the sweep high at 1.1058, and a target at the next "sell-side liquidity" below, for example under a pair of equal lows.

A classical analyst looking at the same chart would describe it as a false breakout above double-top resistance, a break of the uptrend's last higher low, and a plan to sell a pullback into broken support turned resistance. The trade is almost identical. The vocabulary is different.

How SMC maps onto established concepts

Seeing the overlap is useful, because the older concepts have a longer track record of being discussed, tested and criticised.

• Order blocks are close cousins of supply and demand zones and of support and resistance more generally, drawn from one candle instead of a cluster of touches.
• Liquidity sweeps are stop runs and failed breakouts. Traders have discussed them for decades.
• BOS and CHoCH are the trend definitions of Dow-style market structure: higher highs and higher lows, or the first failure of that sequence.
• Fair value gaps are imbalances: areas of fast, one-sided trade. Volume-profile traders describe similar zones as low-volume nodes or single prints.
• Premium and discount is the idea of buying pullbacks below the midpoint of a range rather than chasing.
• "Kill zones", the session windows many SMC courses emphasise, are the London and New York opens, when liquidity and volatility rise. The market hours guide covers those sessions.

None of this means SMC is useless. A consistent vocabulary can make a trader's process clearer. It does mean the ideas are not secret or new, and that any edge they offer has to come from execution, risk management and selection, not from the labels.

The "smart money" narrative vs how the market actually works

SMC's story is that institutions hunt retail stops and leave visible footprints. Some of the underlying observations are fair; the certainty with which they are often presented is not.

What is reasonably well understood about the currency market, as described in our guide to how the forex market works:

• It is decentralised. There is no single exchange and no single order book. Prices are formed across many dealing banks, electronic platforms and non-bank market makers. A retail broker's chart is one feed among many.
• Large participants are varied. Banks, asset managers, hedge funds, corporates hedging trade flows, central banks and high-frequency market makers all trade for different reasons, many of which have nothing to do with short-term price direction. A corporate converting export revenue is not "hunting" anyone.
• Big orders are often worked to minimise impact. Institutions frequently split orders, use algorithms or trade at benchmark fixes. Finding liquidity matters to them, and clusters of resting orders are one source of it, which is the kernel of truth in SMC.
• Stop clusters do exist. Orders bunch around obvious levels and round numbers. When price reaches them, triggered stops can accelerate a move briefly.

What a retail chart cannot show is intent. A spike through equal highs that reverses might be a large participant using stops as liquidity, a news headline, a thin moment in the order book, or simply random variation. Calling every such move a deliberate "raid" fits the chart after the fact, but it is not evidence. Be wary of anyone who claims to know what banks are doing from a 5-minute chart.

There is a practical consequence too. If you believe a pattern is the footprint of institutions, it is tempting to treat it as more reliable than it is and to size up. The pattern deserves the same scepticism and the same position sizing as any other.

What is testable and what is vague

The good news is that several SMC concepts can be defined mechanically. The bad news is that many teaching versions are not, which makes them impossible to evaluate honestly.

Reasonably testable, once you fix the definitions:

• Fair value gaps. The three-candle rule is precise: candle 1 high below candle 3 low (bullish) or candle 1 low above candle 3 high (bearish). You can add a minimum size, for example a fraction of ATR, and test what happens when price returns.
• Swing points, BOS and CHoCH. If you define a swing high as a bar whose high exceeds the highs of a set number of bars either side, breaks can be identified by code without hindsight.
• Equal highs and lows and their sweeps. "Two highs within X pips of each other, then a wick beyond and a close back inside within N bars" is testable.
• Premium and discount. Once the swing range is defined, the midpoint is arithmetic.

Vague in many presentations:

• Which candle is the order block. If the rule is "the last down candle before the move that matters", someone has to decide which move mattered, usually after seeing it.
• Which liquidity is the target. If any prior high or low can be cited as the liquidity price was "going for", every move can be explained.
• Higher-timeframe bias. "Trade with the higher-timeframe order flow" needs an objective definition of that bias, or it becomes a judgement call made with hindsight. The multiple timeframe analysis guide discusses how to define alignment objectively.
• Discretionary confluence. Requiring "an order block plus an FVG plus a sweep in the kill zone" can be tested only if each element is precisely defined, and the more conditions you add, the fewer trades you get and the greater the risk of curve-fitting.

A useful test of any rule: could two people, given the same chart and your written rules, mark the same zones without talking to each other? If not, the rule is not yet testable.

How to backtest SMC rules objectively

The general method is in the backtesting guide. Applied to SMC, it looks like this:

1. Write the rules down before looking at results. Define each element: swing length, what counts as a break (close or wick), FVG minimum size, how long a zone stays valid, entry type (limit at the zone edge, midpoint, or confirmation candle), stop placement and exit.
2. Use a large sample. A handful of hand-picked screenshots proves nothing. Aim for hundreds of trades across several pairs and different market conditions, including trending and ranging periods.
3. Include costs. Many SMC entries use tight stops on lower timeframes, where the spread and commission can be a large fraction of the risk. A 5-pip stop with a 1-pip spread is giving up 20% of the risk before the trade starts. Model slippage on stop exits as well.
4. Avoid hindsight in zone marking. If testing manually, step through the chart one bar at a time and mark zones only with data available at that moment. Software replay modes help.
5. Measure the right things. Track win rate, average win and loss in R, expectancy, maximum drawdown and the longest losing streak. A high win rate with tiny winners can still lose money.
6. Hold back data. Develop the rules on one period and check them on another you have not looked at. If performance collapses out of sample, the rules were probably fitted to noise.
7. Forward test on demo, then small. Live markets add execution realities that a backtest does not show.
8. Record every trade. A trading journal that logs the exact zone, entry reason and outcome lets you check whether you are following the rules or improvising.

A worked illustration of why costs matter, with invented numbers: suppose an FVG entry on a 5-minute chart uses a 6-pip stop and a 12-pip target, so it looks like a 2R trade. With a 0.8-pip spread and the equivalent of 0.6 pips in commission, the real risk is about 7.4 pips and the real reward about 10.6 pips, a ratio of roughly 1.4R. A strategy that needs about 33% winners to break even at 2R needs about 41% after costs. Small changes like this decide whether a rule set is viable.

The course, signal and "funded account" ecosystem

SMC is heavily marketed. Courses, mentorships, private chat groups, indicator subscriptions and signal channels often use the vocabulary, frequently alongside screenshots of large winning trades or prop-firm payout certificates. Some of this is genuine education; a great deal is sales.

Red flags, many of which appear in our guide to forex scam red flags:

• Results shown only as hand-picked screenshots rather than a complete, independently verifiable track record.
• Lifestyle marketing: cars, travel and claims of quitting a job, with little about drawdowns or losing months.
• Claims of insider knowledge about what banks or "the algorithm" are doing.
• Win rates in the 80s or 90s with no information on average losses or sample size.
• Pressure tactics: limited places, countdown timers, upsells to an "inner circle".
• Referral or affiliate links to brokers or prop firms embedded in the teaching, which create an incentive regardless of whether you succeed.
• Signal groups that ask for account access or offer to manage funds. Handing over account credentials or money to an unregulated individual is a serious risk.
• Hindsight charts: every example is marked up after the move, with no live, pre-announced trades.

Remember the base rates. Brokers' own regulatory disclosures show that most retail CFD accounts lose money, and no chart vocabulary changes the costs, leverage and behavioural traps behind those numbers. The guide to why most retail CFD traders lose goes through them.

Common mistakes with smart money concepts

• Marking zones after the fact. The most common error. Every chart looks full of perfect order blocks once you know where price went.
• Treating every wick as a sweep. If any spike can be called a liquidity grab, the concept explains everything and predicts nothing.
• Over-stacking confluence. Requiring five conditions produces beautiful examples and very few trades, which is a classic route to curve-fitting.
• Ignoring costs on low timeframes. Tight stops make spread and commission a large share of the risk.
• Sizing up because "institutions are behind it". A narrative is not a reason to increase risk per trade.
• Mixing definitions from different teachers. If your FVG, order block and BOS rules come from three sources, write down your own version and stick to it.
• Skipping the reading of plain candles. Understanding wicks, bodies and closes, covered in the candlestick charts guide, comes before any of the SMC labels.

Frequently asked questions

What is a fair value gap in trading?

A fair value gap is a three-candle pattern in which the first candle's high and the third candle's low (in a rally) do not overlap, leaving a range that traded only during the fast middle candle. Smart money concepts traders treat it as an imbalance price may return to. Older terms for similar zones include imbalances and single prints.

Is smart money concepts the same as ICT?

They overlap heavily. Much of the SMC vocabulary was popularised by ICT (Inner Circle Trader), and the two labels are often used interchangeably. Different educators define the terms in slightly different ways, so it is worth writing down your own precise definitions.

Do order blocks really show institutional orders?

A retail chart cannot show who placed orders or why. An order block is a price zone defined from one candle before a strong move. It behaves like a small supply or demand zone. Whether it reflects institutional activity cannot be confirmed from the chart, so judge it by how a clearly defined rule performs in testing.

What is the difference between BOS and CHoCH?

A break of structure (BOS) is a break of the latest swing point in the direction of the existing trend, confirming continuation. A change of character (CHoCH) is the first break against the trend, such as the first lower low in an uptrend, taken as an early warning of a possible reversal.

Can smart money concepts be backtested?

Many parts can, provided they are defined mechanically: fair value gaps, swing-based BOS and CHoCH, equal highs and lows, and premium and discount levels. The parts that rely on judgement, such as which order block or which liquidity pool matters, have to be turned into fixed rules first, or the test will be contaminated by hindsight.

Is SMC better than traditional technical analysis?

There is no reliable public evidence that it is. Much of SMC renames existing ideas such as supply and demand, stop runs and trend structure. Any edge depends on the specific rules, costs, execution and risk management, which is why objective testing matters more than the label.