Smart Money Concepts: Sweeps, Order Blocks & FVGs

Smart Money Concepts (SMC) is a trader-created price-action vocabulary for describing market structure, perceived liquidity around highs and lows, liquidity sweeps, displacement, order blocks, fair value gaps, and related chart states.

The useful part of SMC is that it encourages traders to define location, sequence, invalidation, and response instead of treating one candle as a complete signal. The dangerous part is the story often attached to those labels: that an unnamed bank, hedge fund, dealer, or market maker deliberately engineered every wick, stop-out, reversal, or gap.

A candle chart cannot prove that story. It does not identify the participant, show every resting order, reveal hidden liquidity, reconstruct order intent, or tell you whether a trade opened, closed, hedged, or liquidated a position.

Use SMC labels as testable chart hypotheses, not as proof that you can see institutional activity.

Key takeaways:

SMC terminology is not standardized; write your definitions before testing it.
Buy-side and sell-side liquidity are SMC shorthand, not a complete map of actual market liquidity.
A liquidity sweep needs a frozen reference, tolerance, trigger, return rule, time limit, and failure state.
An order block or FVG is an OHLC construction, not evidence of a participant's identity or inventory.
BOS and CHoCH depend on the swing, hierarchy, and wick-versus-close rules chosen in advance.
Real manipulation is a conduct-and-intent question; it cannot be diagnosed from a wick alone.
ChartMini can support blind candle-replay tests, but it cannot replay institutional orders, a historical order book, or true order flow.

This page owns the broad SMC terminology and testing crosswalk. The market-structure verification guide owns detailed BOS, CHoCH, swing, wick-versus-close, and hierarchy rules. The Order Block guide owns detailed zone selection and testing. The Order Flow guide owns actual DOM, Time and Sales, footprint, Delta, CVD, absorption, and data-feed boundaries.

What SMC Is—and What It Is Not

SMC packages several familiar price-action ideas into a shared vocabulary. Depending on the educator or platform, the package may include:

swing highs and swing lows;
higher-high, higher-low, lower-high, and lower-low structure;
Break of Structure (BOS);
Change of Character (CHoCH) or Market Structure Shift;
buy-side and sell-side liquidity;
equal highs and equal lows;
liquidity sweeps, grabs, or stop runs;
displacement;
Fair Value Gaps (FVGs) or imbalances;
order blocks, breaker blocks, and mitigation blocks;
premium, discount, and equilibrium inside a selected range.

These labels can help organize a chart-review process. They are not an official exchange classification, a regulatory standard, or a single academic theory.

Two traders can look at the same chart and disagree about:

which swing is major;
how equal two highs must be;
whether a wick or close confirms a break;
which candle forms an order block;
whether a gap has been partially or fully filled;
when a sweep expires;
whether structure is bullish, bearish, ranging, or transitional.

If the label changes after the outcome is visible, the analysis cannot be evaluated fairly. The first requirement is therefore not a secret setup. It is a versioned definition.

Translate SMC Labels Into Observable Rules

SMC label	Observable chart question	What the label cannot prove
Buy-side liquidity	Is there a predefined high or high-zone where stop-buy or breakout activity may cluster?	Exact order inventory, participant identity, or whether orders still exist
Sell-side liquidity	Is there a predefined low or low-zone where stop-sell or breakdown activity may cluster?	Exact order inventory, participant identity, or motive
Liquidity sweep	Did price breach a frozen reference and then return according to a written rule?	That an institution deliberately hunted stops
Displacement	Did price move farther or faster than a predefined baseline with reduced overlap?	Who caused the move or whether it must continue
Fair Value Gap	Is there a three-candle non-overlap interval under a fixed rule?	Unfilled institutional inventory or a required future fill
Order block	Does a predefined candle or candle group precede a qualified move?	Institutional accumulation, distribution, or hidden orders
BOS	Did price break a predefined with-structure swing under the selected trigger?	Continuation certainty
CHoCH	Did price break a predefined counter-structure swing under the selected trigger?	Reversal certainty
Premium / discount	Is price above or below the midpoint of a frozen range?	Fair value in an economic or valuation sense

The distinction between an observation and a causal story is central. “Price traded above the prior confirmed high and closed back below it” is observable. “A bank pushed price up to steal retail liquidity” is not observable from OHLC data alone.

Buy-Side and Sell-Side Liquidity: Use the Terms Carefully

In common SMC usage:

Buy-side liquidity (BSL) usually refers to potential buy-stop and breakout activity above a prior high or high-zone.
Sell-side liquidity (SSL) usually refers to potential sell-stop and breakdown activity below a prior low or low-zone.

This is specialized shorthand. It is not the same as the broader market-microstructure definition of liquidity.

Market liquidity can involve:

bid-ask spread;
available depth at multiple prices;
order-book resiliency and quote refresh;
execution size;
cost to trade;
price impact;
fill probability and fill quality;
venue fragmentation;
hidden, reserve, midpoint, or off-venue interest.

CME research notes that book depth and spread alone can give an incomplete picture, and that fill quality and price impact can add important information. A candle chart compresses this further into open, high, low, close, and sometimes volume. It does not preserve the full order lifecycle.

Common reference zones

SMC traders often mark potential liquidity references around:

confirmed swing highs and lows;
equal or near-equal highs and lows;
session highs and lows;
prior-day or prior-week extremes;
range boundaries;
obvious breakout or breakdown levels;
round numbers.

These are plausible locations for clustered decisions because many participants can observe them. That does not mean the exact resting-order quantity is known, or that the level must be swept.

How to Define a Liquidity Sweep Without Hindsight

“Price took liquidity and reversed” is easy to say after the reversal. A useful test needs explicit states.

1. Freeze the reference

Select the eligible reference before the sweep candidate appears. Examples include:

the most recent confirmed swing high;
the highest confirmed high within the previous 20 closed bars;
a session high established before a specified cutoff;
an equal-high zone defined by two or more confirmed pivots.

Do not replace the reference with a more convenient level after seeing the outcome.

2. Define equality and tolerance

Two highs are rarely identical. Choose a tolerance such as:

an absolute price amount;
a number of minimum ticks;
a percentage of price;
a fraction of recent volatility;
a zone based on candle bodies or wicks.

The tolerance is part of the setup version. Wider tolerance creates more equal-level candidates; narrower tolerance creates fewer.

3. Choose the breach trigger

Possible versions include:

any intrabar trade beyond the reference;
a minimum tick or percentage breach;
a close beyond the reference;
a breach followed by a close back through the reference;
a breach followed by a close back through a tolerance zone.

A wick-only and a close-confirmed version are different rules and should not be mixed in one result set.

4. Define the response

A sweep candidate is not automatically a reversal. Record separate outcomes:

rejection: price returns inside the reference zone within the allowed window;
continuation: price remains beyond the reference and extends in the breach direction;
retest: price returns to the reference after acceptance beyond it;
failure: the planned reaction occurs briefly but the adverse boundary is then broken;
expiry: none of the defined outcomes occurs before the time limit;
ambiguous: available bar data cannot establish event order.

5. Set the observation horizon

A two-bar reversal and a two-day reversal are not the same phenomenon. Freeze the horizon in bars, minutes, sessions, or another measurable unit.

Fair Value Gaps: A Chart Interval, Not a Promise

In common SMC terminology, a Fair Value Gap is a three-candle non-overlap interval.

A frequently used bullish definition is:

Low of candle 3 > High of candle 1

The interval is:

[High of candle 1, Low of candle 3]

A frequently used bearish definition is:

High of candle 3 < Low of candle 1

The interval is:

[High of candle 3, Low of candle 1]

The middle candle is commonly expected to show directional expansion or displacement, but the exact body, range, volume, or overlap requirement varies by methodology.

Define FVG states before testing

Use states such as:

created: the three required candles have closed;
untouched: later price has not entered the interval;
entered: later price has traded inside the interval;
partially filled: price has reached a predefined fraction or midpoint;
fully filled: price has reached the opposite boundary;
closed through: price has closed beyond the opposite boundary under a selected rule;
expired: the maximum age has been reached;
merged: overlapping intervals are combined under a written rule;
ambiguous: chart resolution or missing data prevents classification.

Do not state that an FVG is “unfilled institutional inventory” unless you have data that demonstrates that claim. The chart only demonstrates the non-overlap geometry.

A future revisit may occur for many reasons, and some intervals may never be revisited within the selected horizon. “Fair value” in this label should not be confused with fundamental valuation, exchange-calculated fair value, or a guaranteed equilibrium price.

For ordinary session gaps and close-to-open gap measurement, use the price-gap reading guide. An SMC FVG and an overnight opening gap are different constructions.

Order Blocks: Freeze the Selection Rule

An order block is commonly described as a candle or compact candle group preceding a qualified directional move. Definitions vary substantially.

Possible bullish versions include:

the final bearish candle before a bullish displacement;
the full high-low range of that candle;
only its body;
the final down-close candle before a BOS;
the origin of a move that also creates an FVG.

Bearish versions reverse the direction.

Before testing an order block, define:

eligible candle direction;
single candle versus candle group;
body versus wick boundaries;
required displacement;
required structure event;
maximum zone width;
age and expiry;
whether the first touch consumes the zone;
partial-entry, full-entry, rejection, acceptance, break, and retest states;
overlapping-zone precedence.

The detailed Order Block trading guide owns this classification and testing layer. The key SMC boundary is simpler: an order block is a price-zone hypothesis, not participant identification.

BOS and CHoCH Belong to a Defined Structure Model

A Break of Structure cannot exist without a prior definition of structure.

You need to know:

which pivots count;
when a pivot becomes confirmed;
whether equal highs and lows are distinct;
which swing is local and which is major;
whether the break uses a wick, close, buffer, follow-through, or retest;
what the pre-break state was.

A common framework treats:

BOS as a break in the direction of the prevailing predefined structure;
CHoCH as a break against the prevailing predefined structure that signals transition rather than a guaranteed reversal.

The complete rules, hierarchy, failure states, and look-ahead controls belong to the BOS and CHoCH verification guide. The broad market-structure guide retains the higher-high, higher-low, lower-high, lower-low, trend, and range overview.

Displacement and Premium/Discount Need Versions Too

Displacement

“Strong move” is not a testable definition. Possible displacement versions include:

true range above a rolling percentile;
candle body above a volatility multiple;
directional movement across several bars with low overlap;
close location near the bar extreme;
a move that crosses a frozen structural reference;
a move that creates a qualifying FVG.

These versions may select different samples. Do not label only the moves that later worked.

Premium, discount, and equilibrium

SMC often divides a selected range at its midpoint:

Midpoint = (Range high + Range low) / 2

Above the midpoint is commonly called premium.
Below the midpoint is commonly called discount.
The midpoint is commonly called equilibrium.

The result depends entirely on the chosen range anchors. It does not establish intrinsic value or prove that price is economically expensive or cheap.

Record:

range start and end;
wick or close anchors;
whether the range is completed or still developing;
the timestamp at which both anchors became available;
re-anchoring and invalidation rules.

Liquidity Sweep Does Not Mean Illegal Manipulation

Large trades can move prices because they consume available liquidity. Stops, liquidations, hedging, news reactions, index rebalancing, options-related flows, dealer inventory changes, and ordinary changes in supply and demand can also contribute to fast movement around visible levels.

That is different from proving illegal manipulation.

The CFTC describes spoofing around entering bids or offers with intent to cancel before execution. SEC and CFTC enforcement cases rely on order-level evidence such as non-bona-fide orders, genuine orders on the other side, cancellation patterns, timing, and intent.

A candle wick does not contain that evidence.

Use precise language:

Observable: price traded beyond a confirmed high and closed back below it.
Hypothesis: clustered conditional orders may have contributed to activity around the level.
Unsupported from candles alone: a specific institution deliberately manipulated price to take retail stops.

Actual market manipulation exists and regulators enforce against it. The error is treating every adverse fill or failed breakout as proof of manipulation.

Market and Data Boundaries

SMC labels can be drawn on many charts, but the underlying data are not equivalent.

Centralized futures

A listed futures contract has an exchange order book for that venue, but a standard OHLC chart still omits most order-level information. Contract roll, session template, settlement, tick size, and data-feed depth matter.

U.S. equities

Trading can occur across exchanges and off-exchange venues. One chart or one Level 2 feed may not represent all displayed and non-displayed activity. Corporate actions, opening and closing auctions, halts, and extended-hours sessions affect structure.

Crypto

Liquidity is fragmented across exchanges, perpetual contracts, spot markets, and on-chain venues. A sweep on one venue may not appear identically elsewhere. Liquidations and leverage can produce rapid moves without proving a coordinated institutional stop hunt.

Spot forex and CFDs

There is no single consolidated global spot-FX order book. Broker feeds, liquidity providers, session conventions, spread models, and volume fields differ. A retail CFD chart cannot reveal all interbank orders.

Before comparing samples, freeze:

instrument and venue;
symbol and contract;
data provider;
timezone and session;
adjusted or unadjusted data;
chart interval;
completed versus developing bars;
minimum tick and price precision;
missing-data and roll policies.

A Reproducible SMC Testing Workflow

Step 1: Write one question

Examples:

After a wick-only breach of a confirmed prior-day high, how often does price close back below that high within three bars?
After a bullish FVG is created, how often is its midpoint reached within 20 bars?
Does adding a predefined displacement filter change the result versus all three-candle non-overlap intervals?

Do not test “Does SMC work?” as one undifferentiated question.

Step 2: Freeze every input

Record:

chart identity;
swing rule;
equal-level tolerance;
reference hierarchy;
sweep trigger;
return condition;
FVG boundaries;
order-block rule if used;
BOS/CHoCH version if used;
observation horizon;
ambiguity treatment.

Step 3: Separate candidate, confirmation, and outcome

A useful state sequence might be:

reference available;
candidate breach;
rejection or acceptance confirmation;
follow-through;
failure or invalidation;
expiry;
evaluated outcome.

This prevents later bars from silently changing the original label.

Step 4: Compare against simpler baselines

Possible baselines include:

all swing-high or swing-low breaches;
random eligible references;
ordinary support/resistance zones;
all three-candle non-overlap intervals;
matched-volatility events without SMC filters;
the same setup without the order-block or FVG condition.

A complex label is useful only if it adds stable information beyond a simpler rule.

Step 5: Keep chart outcomes separate from trade results

A chart event can be evaluated without assuming an executable trade.

Chart outcomes may include:

return inside the range;
maximum favorable or adverse excursion;
time to midpoint or opposite boundary;
continuation distance;
failure rate;
expiry rate.

A trade simulation additionally requires:

entry timing;
order type;
fill model;
spread and commission;
slippage;
stop and exit rules;
quantity and risk budget;
gap and same-bar event handling.

Step 6: Use chronological evaluation

Develop definitions on an earlier sample, freeze them, and evaluate them on later unseen data. Review results by market, timeframe, volatility, session, and trend/range regime.

Reject or revise a version when:

results depend on one instrument or period;
a small tolerance change reverses the conclusion;
most examples require hindsight relabeling;
costs remove the apparent advantage;
the rule does not outperform a simpler baseline;
ambiguous bars dominate the sample.

SMC Replay Worksheet

Field	Example record
Chart identity	Symbol, venue, provider, interval, session, timezone
Structure version	Pivot rule, hierarchy, equality tolerance, wick/close trigger
Liquidity reference	Prior-day high, confirmed swing low, equal-high zone, range edge
Reference availability	Timestamp when the level became knowable
Sweep version	Breach threshold, return rule, maximum bars
FVG version	Three-candle rule, minimum size, fill definition, expiry
Order-block version	Candle selection, boundaries, displacement/BOS requirement
State	Candidate, confirmed, failed, invalidated, expired, ambiguous
Chart outcome	MFE, MAE, return time, continuation, full/partial fill
Trade assumptions	Optional entry, fill, costs, stop, exit, quantity
Notes	Data issue, gap, halt, roll, news, session transition

This worksheet makes disagreements visible. Two traders may use different SMC definitions, but each version can still be documented and evaluated.

What ChartMini Can and Cannot Do

ChartMini can help with a narrow transfer test:

Open a historical candle replay.
Hide future candles.
Freeze the chart identity and SMC rule version.
Mark the eligible structure and liquidity references using your own notes.
Advance one candle at a time.
Record candidate, confirmation, failure, expiry, and ambiguity states.
Compare the outcome with a simpler baseline.

ChartMini does not provide:

a synchronized historical Level 2 or DOM book;
Time and Sales;
footprint charts, Delta, or CVD;
hidden, reserve, or off-venue orders;
participant identities;
proof of institutional accumulation or distribution;
spoofing or manipulation detection;
broker-specific fills, queue position, or market impact;
automatic SMC labeling or strategy certification.

A candle replay can test whether your chart definition is consistent. It cannot validate the institutional story attached to the label.

Common SMC Mistakes

Explaining every loss as a stop hunt

A stopped trade can result from ordinary volatility, an invalid setup, a poor stop location, changing information, slippage, or random variation. “Manipulation” is not a substitute for review.

Changing the swing after the break

If the eligible swing is selected after the break, BOS and CHoCH classifications contain hindsight.

Treating every wick as a sweep

A sweep needs a predefined reference, threshold, return condition, and horizon. Otherwise almost any wick can be relabeled later.

Treating every FVG as support or resistance

An FVG is first a geometric interval. Reaction behavior must be evaluated rather than assumed.

Treating an order block as an order record

The candle does not show whose orders were placed, canceled, executed, or left unfilled.

Mixing markets and feeds

A futures chart, a CFD feed, a crypto perpetual, and a stock chart do not expose the same liquidity or trading structure.

Stacking labels without testing the contribution

BOS + sweep + order block + FVG + premium/discount may sound selective, but each added condition reduces the sample and increases overfitting risk. Test the incremental value of every filter.

Frequently Asked Questions

What are Smart Money Concepts in trading?

Smart Money Concepts, or SMC, is a trader-created price-action vocabulary for describing market structure, perceived liquidity around highs and lows, liquidity sweeps, displacement, order blocks, fair value gaps, and related chart states. It is not a standardized regulatory or academic model, and an SMC label does not by itself identify an institution or prove manipulation.

What is a liquidity sweep in SMC?

A liquidity sweep is an SMC label for price trading beyond a predefined high, low, or equal-level zone and then reacting around that reference. A reproducible definition must specify the reference-level rule, tolerance, wick-or-close trigger, return condition, time window, and failure state. The chart pattern does not reveal who traded or why.

What is a fair value gap?

In common SMC usage, a fair value gap is a three-candle price interval where candle one and candle three do not overlap. A bullish version is often defined when the low of candle three is above the high of candle one; the bearish version is the reverse. It is a chart construction, not proof of unfilled institutional orders or a required future fill.

Does an order block prove institutional buying or selling?

No. An order block is a chart-zone label whose candle-selection rules vary among traders and platforms. OHLC candles cannot identify the participant, order size, inventory, hidden liquidity, or motive. Treat an order block as a versioned price-zone hypothesis and test its reactions against simpler baselines.

Is a liquidity sweep the same as market manipulation?

No. Market manipulation and spoofing require evidence about conduct and intent, such as entering non-bona-fide orders with intent to cancel before execution. A wick through a prior high or low may result from ordinary order flow, news, thin liquidity, forced orders, hedging, or many other causes. Candle data alone cannot establish manipulation.

Can ChartMini show institutional orders or real liquidity?

No. ChartMini replays historical candles for chart-reading and decision-practice exercises. It does not provide a synchronized historical order book, Time and Sales, footprint data, hidden orders, participant identities, broker fills, queue position, or evidence that an institution created a particular move.

Official Sources and Research

Trading involves risk, including loss of capital. SMC labels are educational chart-analysis tools, not proof of institutional intent, a recommendation, or a guarantee of future results.

Smart Money Concepts: Liquidity Sweeps, Order Blocks, and FVGs