SMC Trading Dissected — Selection Rules for Order Blocks, Liquidity Sweeps, FVG & BOS/CHoCH

The Skeleton of SMC — Big Capital Needs the Other Side of the Trade

SMC (Smart Money Concepts) grew out of a framework that forex trader Michael J. Huddleston taught for years under the name ICT (Inner Circle Trader), and it spread to retail traders worldwide after he released a free mentorship series on YouTube in 2022. The roots aren't new. The lens of imagining the market as a single big player and reading along with its intent was laid out a century ago by Wyckoff; SMC is closer to a derivative system that translates that lens into the language of modern derivatives markets — stops, liquidations, and resting orders.

The entire system stands on one assumption. A buyer trying to acquire $100 million worth has to get matched against the same amount of sell-side size, so it can't build a position at just any price. The place where opposing orders are stacked thickest — where stops and forced liquidations cluster — is nearly the only place big capital can process its size, and therefore price moves toward that place. From this assumption, every tool — liquidity, sweeps, order blocks, FVG, BOS/CHoCH — is deduced. This piece dissects those tools in the order of 'how to identify them on a chart, how to assemble them, and where to admit you're wrong.'

Price moves from liquidity to liquidity — the sentence SMC repeats like a first principle, and the claim this entire piece puts to the test.

Drawing the Liquidity Map — Where Stops and Liquidations Pile Up

In SMC, liquidity means clusters of unfilled orders resting at specific price levels. They tend to pile up in three places. First, just outside the prior swing high/low — stop-loss sells from longs cluster below the prior low, stop-loss buys from shorts above the prior high. Second, around round numbers like $60,000. Third, the pillar unique to crypto futures: forced liquidation. The liquidation price of N× leverage sits roughly 1/N away from the entry (an approximation that ignores maintenance margin), so liquidation size stacks like a ladder above and below the price zones where entries clustered. Stop orders can be canceled, but a liquidation can't — which is why liquidation clusters are considered the most reliable form of liquidity. The mechanics are covered in What Is Liquidation?.

1. Mark higher-timeframe structure — mark swing highs and lows on the 4H and daily. Ignore the lower-timeframe ripples at this stage.
2. Flag stop-cluster candidates — draw horizontal boxes just outside prior extremes, double tops/bottoms, and support/resistance areas that have been defended multiple times. Draw them as zones, not lines.
3. Add round numbers — overlay psychologically sticky prices (e.g., BTC $60,000 and $65,000).
4. Back-calculate the liquidation ladder — from the entry zone where volume piled up, compute and mark the 25× (±4%), 10× (±10%), and 5× (±20%) liquidation prices.
5. Assign priority — treat zones where stops, round numbers, and the liquidation ladder overlap as the thickest liquidity pools, and only move to the next step (watching for a sweep) when price approaches them.

Spotting the Liquidity Hunt — Anatomy of a Sweep

A liquidity hunt (sweep, stop hunt) is a move where price briefly pokes into a dense zone marked on the map, fills the resting orders, and comes back. The identification conditions compress to three. First, the breach of the extreme is left as a wick, not a closing body. Second, within 1–3 candles of the breach, price returns and closes back inside the level. Third, volume and liquidation fills spike at the moment of the poke — hard evidence that resting orders were actually swept. If the close-back-inside condition is missing, your first suspicion should be a breakout in progress, not a sweep. The structure of how stop hunts get engineered is dissected separately in the stop hunt guide.

Order Blocks and FVG — Candidate Rules and Confirmation Steps

An order block is the zone of the last opposite-direction candle (or cluster) right before a sharp change in direction. The last down candle before a strong rally is called a bullish order block; the last up candle before a sharp drop, a bearish order block. On that definition alone, half the chart qualifies — so practical SMC layers three filters on top. First, the move after the block must be strong enough to break the prior structure (displacement — accompanied by the BOS covered in the next section). Second, only untouched blocks that price hasn't yet returned to remain valid candidates. A block that's been pierced deep once is considered spent. Third, blocks formed right after a liquidity sweep get higher priority.

Let's put invalidation and the risk-reward structure into numbers (a hypothetical calculation for learning purposes). Say a bullish order block prints at $2,410–2,440 on the ETH 4H chart, and the invalidation rule is set as 'a close below the bottom of the block.' Taking a retracement touch of the block's top at $2,440 as the hypothetical entry and placing the stop slightly below the bottom at $2,398, the risk is $42 (about 1.7%). Targeting the prior swing high at $2,566 makes the reward $126 — a 1:3 risk-reward structure. The point isn't prediction: it's that the block defines the invalidation level and stop distance before the entry is even considered.

BOS and CHoCH — Reading Structure Shifts

The raw material for reading structure is swing points. Mark the highs that stand above their neighboring candles and the lows that sit below them, and the market simplifies into a sequence of highs and lows. A BOS (Break of Structure) is a new extreme in the direction of the trend — in an uptrend, a close above the prior high — read as confirmation of trend continuation. A CHoCH (Change of Character) is the opposite: the first moment structure breaks against the trend. In an uptrend, a prior swing low giving way on a closing basis is the first reversal warning. Judging both by closes rather than wicks is the minimum safeguard against hindsight reinterpretation.

1. Judge the trend — on the 4H/daily, a string of BOS reads as trend continuation; a CHoCH gets classified as a reversal alert.
2. Map the liquidity — use the procedure from section 2 to pre-mark liquidity candidates beyond the extremes.
3. Watch for the sweep — when price touches a candidate zone, check the three conditions (wick-only breach, close back inside, burst of fills).
4. Confirm on the lower timeframe — right after the sweep, watch whether a CHoCH prints on roughly the 15-minute chart. A CHoCH without a sweep, or a sweep without a CHoCH, gets demoted to a half signal.
5. Record the zone and invalidation — mark the order block/FVG at the reversal origin as the retracement watch zone, and write down the invalidation (a close beyond the far side of the block) first.
6. Review — whether the scenario worked or failed, record which link in the chain broke, building your own verification data.

The order is the whole point. Grabbing a later link while an earlier link in the sweep → CHoCH → retracement chain is missing — calling some down candle an order block when there wasn't even a sweep — is exactly where this method turns to mush. And whatever the scenario, once invalidation is confirmed on a close, you scrap it on the spot. On a leveraged account, 'let's just watch a little longer' is the shortest path to forced liquidation.

The Reproducibility Debate — Everything Looks Right in Hindsight

Now for the honest part. On a hindsight chart, SMC is always right. Every down candle before a bounce looks like an order block, every empty zone looks like an FVG, every long wick looks like a sweep. But apply the same rules forward from the right edge of the chart and the candidates balloon into dozens — and which one works can only be known after the fact. The looser the rules, the higher the hindsight hit rate and the blurrier the forward predictive power — that's the core of the confirmation-bias criticism that trails SMC.

More fundamentally, critics persistently note that SMC's rules aren't mathematically closed, which makes independent backtesting difficult, and that no publicly verified long-term performance record has been confirmed. The live-trading results of ICT himself, the founder, have faced ongoing verification controversy as well. 'Smart money's intent' can't be observed, so it can be neither verified nor falsified. Even so, it's not all worth discarding. That stops and liquidations pile up beyond the extremes, and that volatility gets amplified through cascading fills once that zone is touched, is observable structure, not narrative. One question remains — instead of guessing 'where' that liquidity sits from chart shapes, is there a way to measure it?

Measured on Whale Story — a Map of Real Liquidation Prices, Not Guesses

The Whale Levels on Whale Story's live tracker overlay the actual average entries and liquidation prices of top Hyperliquid whales onto the chart. That means the liquidity pool SMC guesses 'should be piled up around here' can be cross-checked against a measured value: the price bands where real liquidation prices stack layer on layer. Hyperliquid makes this comparison possible because positions are public on-chain (What Is Hyperliquid?). Put the liquidation heatmap — an estimated density map — next to the measured levels, and you can grade your evidence by reliability: chart guess → estimated map → measured value.

The observation routine is simple. After drawing your liquidity map, cross-check whether real whale liquidation prices actually cluster in your candidate zones, and record whether liquidation fills actually print in succession on the feed as price approaches. If nothing happened in the zones you marked, the map was wrong — and that record itself becomes the data with which you verify SMC with your own hands. In every case, these tools are raw material for detection and observation; they don't pick a direction. The choice of direction and the bearing of losses always belong to the account's owner.