Smart money reversal (SMR)

A turn can print anywhere on a chart, and the smart money reversal is ICT’s name for one that prints at a place marked in advance: a sweep and a structure shift inside an array drawn on a larger timeframe.

The construction

Three parts, in order, each defined on its own page. The location is a higher-timeframe array: a fair value gap, an order block, a rejection block or an old high or low, drawn on a larger timeframe than the one the rest of the sequence is read on. The sweep happens there: price trades beneath a prior swing low inside the array and closes back above it within a stated window, a liquidity sweep by that entry’s test. The shift follows: within a stated number of candles, a market structure shift the other way, a close through the last qualifying lower high, delivered with displacement where the definition requires it.

That is the bullish form, which ends a decline, and teaching looks for it at an array in the discount half of the dealing range, in the terms of the PD array entry. The bearish form ends a rally at an array in the premium half: a swing high inside it is swept, and the shift breaks the last higher low. The detector campaign registers the event as a sweep plus a shift at an array from a larger timeframe, in its own row of the claims ledger.

Where the name comes from

The name carries the account of large participants that ICT teaching shares with the smart money concepts vocabulary. In its account of the turn, the sell orders resting beneath the swept low, the stops of earlier buyers and the entries of breakout sellers, give large buyers the other side of their trades, and the shift is read as that buying showing on the chart. Candle data records none of it, only a trade beneath a price, a close back above it and a break of a swing; who traded, and why, is not in the data. The name is kept because readers search for it, and nothing in the definition depends on the account being true.

Pages that rank for the term use it more loosely: one for any sweep followed by a shift, another for any turn it reads as large participants changing sides, confirmed with volume or momentum readings. This entry keeps the narrower meaning the market maker model entry and the campaign’s row share, because the location fixed in advance is what separates the reversal from sweeps and shifts everywhere else.

A worked case

Constructed numbers, with the elections stated first and made for the example, not as defaults: arrays are four-hour fair value gaps; the sequence is read on fifteen-minute candles; the sweep’s extreme must lie inside the gap and its candle must close back across the swept low; and the shift must close through an intermediate swing, a short-term swing flanked by lesser short-term swings as the market structure shift entry grades the tiers, with displacement read as a body of at least 0.60 of the range and a range of at least 1.5 times ATR(14), within eight candles of the sweep.

On GBP/USD a four-hour bullish gap spans 1.2640 to 1.2655, from the first candle’s high to the third candle’s low. Days later price declines into it. On the fifteen-minute chart a swing low prints at 1.2651, inside the gap, and the bounce from it prints short-term highs at 1.2663, 1.2672 and 1.2667, so 1.2672, flanked by lower short-term highs, is an intermediate high under the stated rule. Price then trades down to 1.2644, seven pips beneath the low and still inside the gap, and that candle closes at 1.2657: the sweep. Four candles later a run of candles that each pass the displacement election closes at 1.2680, above 1.2672: the shift, inside the window. Under these elections the sequence is a bullish smart money reversal, and its extreme is 1.2644. A later trade beneath 1.2644 voids the reading. What price does after the close at 1.2680 is part of no definition.

The bearish mirror

Every direction reverses, and the timeframe pair moves up to daily arrays and hourly candles, the other elections unchanged. On EUR/USD a daily bearish gap spans 1.0890 to 1.0905, from the third candle’s high to the first candle’s low. On the hourly chart price rallies into it, prints a swing high at 1.0897 inside the gap, and pulls back through short-term lows at 1.0881, 1.0874 and 1.0879, so 1.0874, flanked by higher short-term lows, is an intermediate low.

Price then trades up to 1.0902, five pips above the high and still inside the gap, and that candle closes at 1.0893. Three candles later a displaced run closes at 1.0868, beneath 1.0874: a bearish smart money reversal, with its extreme at 1.0902. The numbers are constructed again, and a reversal on one timeframe pair is a different event from one on another.

The elections

Each part carries its own page’s parameters, and the reversal adds six. The timeframe pair: which larger timeframe supplies arrays for which smaller one. The array list: which types count; the wider it is, the likelier some array sits near any turn.

The location rule: the sweep’s extreme inside the array, or within a stated distance where the array is a single price such as an old high. The array’s life: whether a gap already traded through, or an array reached once before, still qualifies. The window between sweep and shift. And the shift’s grade: which swing tier qualifies, a close or a wick through it, and whether displacement is required. Change one election and a different set of events exists, which is why a count of reversals travels with its elections.

Inside the market maker model

The market maker model is the campaign drawn around this turn: a first half that leaves an original consolidation in stages and ends at a higher-timeframe array, the smart money reversal there, and a second half back through the earlier pauses. Read alone, the reversal sets no objectives. Inside the model it is the state at which objectives are named, nearest pause first, and its extreme marks the model failed if price trades back beyond it before the consolidation is reached. On a lower timeframe the turn often has the shape of the ICT 2022 model.

Identifying one, step by step

1. Declare the pair and the list: the larger timeframe for arrays, the smaller for the sequence, and the array types that count.
2. Draw arrays as their candles close on the larger timeframe, and keep only those still live under the stated rule. A four-hour gap does not exist until its third candle closes, and a record that uses it earlier has look-ahead bias.
3. Mark the swings inside an array once price reaches it: the swing lows of a decline, or the swing highs of a rally, on the smaller timeframe.
4. Flag the sweep: a trade beyond one of those swings, with its extreme inside the array and a close back across the swing within the reclaim window.
5. Flag the shift: a close through the last qualifying swing the other way, within the window and at the stated grade.
6. Record the sweep’s extreme at once, because a reading that names nothing that would void it cannot be scored.

Trading it, as taught

Teaching treats the reversal as a trigger. The entry is usually taken on a return into the gap or block the shift’s move left, with the stop beyond the sweep’s extreme and the objective at resting liquidity on the far side: a draw on liquidity or, inside the market maker model, the nearest earlier pause. One guide also grades reversals by the swept level’s significance, the force of the move away and whether the sequence finished inside a kill zone. Each is an election, and a grade chosen after the outcomes are known sorts a record by its results instead of forecasting them.

What it is not

Not a sweep alone: a sweep inside an array that no shift follows stays a sweep. Not a shift alone: a break with no sweep before it, or far from any array, is only a shift. And not a reversal in the ordinary sense, which is a verdict about later travel: the sequence is complete when the shift prints, before any later travel exists, and the name borrows the verdict’s word for it.

Common mistakes

Drawing the array after the turn, by searching the larger timeframe for whichever zone sits nearest the low. Reading the array and the sequence on the same timeframe, which removes the location condition the definition depends on. Changing the array list from one analysis to the next. Counting the sweeps at arrays that turned while forgetting the ones price passed straight through. And writing that smart money reversed the market, when what printed was a sweep and a break on stated candles.

What the reversal does not settle

The sequence is checkable: an array, a sweep inside it and a shift after it, which two readers with the same elections mark identically. The forecast attached to it is a separate claim: that the sequence at a higher-timeframe array is followed by travel in the new direction more often, or further, than the same sweep and shift away from any array, and than an array reached without them. Testing it takes those counts, marked as each reversal prints, with every reversal voided beyond its extreme counted beside every one that travelled. The smart money reversal driver has a row in the claims ledger, where that verdict renders, and the two lenses keep the sequence apart from the forecast.