· Mis à jour
Pourquoi les niveaux évidents sont balayés : clusters de stops, l'anatomie d'un piège en 3 bougies, et comment distinguer les vraies cassures des fausses.
Auteur: Charts Quest
Créez un compte ChartsQuest gratuit pour lire l'article complet — et obtenez les deux premiers niveaux de la quête.
Inscription gratuiteEvery beginner who learns about support and resistance has the same experience within their first few weeks. They spot a beautiful, obvious level — say, resistance at $100 that price has bounced off three times. Price finally pushes through it. They buy the breakout, exactly as the textbook says. Price climbs a little further, stalls, reverses, and within a few candles is back below $100, taking out their stop and continuing down. The "breakout" was false. The obvious level failed in the most painful way possible: not by holding, but by breaking just enough to pull people in before reversing.
If this has happened to you, here is the uncomfortable and oddly liberating truth: it was not bad luck, and it was not a one-off. False breakouts are a structural feature of how markets work, and they happen because the level was obvious, not despite it. This guide explains the mechanics from the ground up — where stop orders pile up and why, what "liquidity" actually means, how a sweep unfolds candle by candle, and how to tell a genuine breakout from a trap. By the end, the move that used to feel like the market personally targeting you will look like a predictable pattern you can recognize, plan around, and sometimes simply step aside from.
One promise up front: nothing here will let you avoid every trap. Nobody can. The realistic goal is smaller — to stop being the most predictable person at the level, to define your invalidation before you enter, and to lose small and planned amounts when you are wrong instead of large and surprised ones.
Start with a simple observation: everyone is looking at the same chart. The support level you found at $50 is not your private discovery. Every trader, every piece of charting software, and every automated system watching that asset can see that price bounced at $50 three times. Horizontal levels, round numbers, prior highs and lows — these are obvious precisely because they are visible to all participants simultaneously.
Now think about what an obvious level produces: synchronized behavior. Thousands of independent traders, having read the same textbooks, do the same three things around the same price:
Notice what has accumulated just below the obvious level: a dense layer of sell orders — stop-losses from longs (which execute as market sells) and entry orders from breakdown shorts. The level everyone trusts has, by the very act of being trusted, manufactured a pocket of one-directional orders right beneath it.
This is the seed of every false breakout. An obvious level does not fail randomly. It fails because the crowd's textbook-correct behavior around it creates something valuable to other participants: a concentration of guaranteed orders at a known price. To understand who finds that valuable and why, we need one more concept — liquidity.
"Liquidity" sounds abstract, but in this context it means something concrete: orders waiting to be filled. A market is liquid where there are many willing buyers and sellers, and illiquid where there are few.
Here is the constraint that drives everything in this article: every buy needs a seller, and every sell needs a buyer. If you want to buy 100 shares, someone must sell you 100 shares at that price. For a small trader this is never a problem — your 100 shares get filled instantly without moving the price. But imagine a participant who needs to buy a very large position — a fund, a market maker, any entity trading size. If they simply slam a huge market buy into a quiet market, their own order eats through all the available sell orders and pushes the price up against them as they buy. They get progressively worse prices. Their size is a handicap.
What a large participant needs is a moment when lots of people are selling at once, so there is plenty of supply to buy from without moving the price. And where, on the entire chart, is there a known, pre-built cluster of sell orders waiting to fire all at once?
Just below obvious support. The stop-losses of every textbook long, plus the entry sells of every breakdown short, sitting in a tidy pile at $49.80 and below.
This is why traders use the phrase liquidity pool: a price zone where resting orders have accumulated. Stops below support are a pool of sell-side liquidity. Stops above resistance — from short sellers who placed their stops just above the obvious ceiling, plus breakout buyers' entry orders — are a pool of buy-side liquidity. These pools are not hidden. Anyone who understands crowd behavior can predict where they sit, because the crowd builds them in the same places every time:
Once you see levels this way, the chart changes character. A clean triple-bottom is not only "strong support." It is also a large, well-advertised pile of sell orders sitting just below — fuel, waiting for a spark.
A sweep (also called a stop hunt, stop run, or liquidity grab) is what happens when price is pushed into one of those pools and the resting orders fire. It is worth walking through slowly, because the mechanics explain everything about how false breakouts look on a chart.
Set the scene with numbers. A stock has bounced off $50.00 support three times over two weeks. The order landscape below the level looks roughly like this:
Now the sequence:
Step 1 — the push. Price drifts down to $50.10. Selling pressure — perhaps ordinary, perhaps deliberate from a participant who would like those stops to fire — nudges price through $50.00. It only takes modest selling to break the surface, because everyone defending the level is already long; few fresh buyers step in at $49.95.
Step 2 — the cascade. The first stops trigger at $49.90. A stop-loss for a long position is a market sell order — when it fires, it sells at whatever price is available, pushing price lower. That lower price triggers the next layer of stops at $49.80, which push price lower still, triggering $49.70... Each stop that fires is fuel for the next. Meanwhile, the breakdown shorts' entries are firing too, adding more selling. Price doesn't drift through the pool; it accelerates through it, often in one or two fast, ugly candles. The market did not suddenly discover bad news at $49.90. The move is mechanical — a chain reaction of pre-placed orders detonating in sequence.
Step 3 — the absorption. Here is the pivotal moment. All that cascading selling is supply, and supply is exactly what a large buyer was waiting for. Into the flood of stop-loss sells around $49.40–$49.70, large resting buy orders absorb the selling. The big participant fills a position at better prices than were available all week, courtesy of the crowd's stops. Notice the elegant brutality: the sellers at the bottom of the flush are mostly people who were buyers at $50.20 yesterday, now selling their shares at $49.50 — at a loss — directly to the participant who wanted them.
Step 4 — the snap-back. Once the pool is drained, something important is true: everyone who was going to sell has sold. The stops are gone. The breakdown shorts are in — and they are now trapped, because the selling pressure is exhausted and price has stopped falling. Even modest buying now moves price up quickly through the vacuum. As price climbs back above $50.00, the trapped shorts' own stop-losses (which are buy orders, sitting above their entries) begin to fire, adding buy fuel. Price snaps back inside the old territory, often as fast as it fell — and frequently keeps going, because the level is now genuinely cleared of sellers.
On the chart, the whole event compresses into a long wick stabbing below $50 and closing back above it. Beginners look at that wick and see chaos or bad luck. You should now see four mechanical steps: push, cascade, absorption, snap-back.
To be fair and accurate: not every sweep is a deliberate "hunt" by some specific actor. Often it is simply emergent — the orders are there, ordinary selling tips the first domino, and the cascade-then-vacuum dynamic plays out with no mastermind required. The mechanics, and the chart pattern they leave, are the same either way. You do not need to know who swept the level to recognize that it was swept.
Sweeps leave a recognizable footprint. While real charts are messier than diagrams, the classic false breakout resolves into three acts — often literally three candles on whatever timeframe you are watching, sometimes spread over a few more.
A strong-looking candle drives through the level. Through resistance at $100, say: a tall green candle closing at $100.80. It looks like everything a breakout should look like, which is precisely the trap — this candle's job (functionally speaking) is to trigger the resting orders beyond the level and to convince breakout traders that the move is real. Volume often spikes here, which beginners read as confirmation. But volume on the push is ambiguous: it can be genuine breakout demand, or it can be the stop cascade itself plus the trapped traders piling in.
This is the tell, and it is subtle. A real breakout should find acceptance beyond the level: continued buying, follow-through, price comfortably building above $100. Instead, the next candle goes nowhere. A small body, perhaps a doji, perhaps a feeble push to $101.10 that immediately fades — price hovering just beyond the level without conviction. Mechanically, the stall means the breakout has run out of fuel: the stops have all fired (that fuel is spent), and new buyers — people willing to initiate fresh positions above $100 — are not showing up. The cascade bought the move its first candle; only genuine demand can buy it a second, and the second candle is where genuine demand fails to appear. Experienced range traders watch candle 2 like hawks. The longer price stalls just beyond a level without extending, the worse the odds for the breakout.
The confirmation. A decisive candle closes back through the level — back below $100 — putting price inside its old territory. This candle is the trap snapping shut: every breakout buyer above $100 is now underwater, and their stops (sell orders, sitting just below their entries... which is to say, right around the level) begin to fire, accelerating the move away. The candle-3 close is the moment the false breakout becomes readable rather than merely suspected. Before it, you have a push and a pause, which might still resolve upward. After it, you have a failed breakout with trapped participants — one of the more reliable situations in chart reading, because the trapped traders' exits provide mechanical fuel for the reversal.
The three-act structure — push, stall, close back inside — is worth burning into memory, because it converts a chaotic experience into a checklist. When you feel the urge to chase a breakout, the anatomy tells you exactly what you are risking: being the candle-1 buyer who provides candle-3's fuel.
Real breakouts exist. Levels do genuinely fail, trends do extend, and a trader who fades every breakout gets steamrolled by the real ones. The skill is not predicting which is which in advance — nobody can do that consistently — but knowing which evidence to wait for and what each clue is worth.
The single most useful distinction: where did the candle close? A wick beyond a level means price visited the territory; a close beyond the level means price was accepted there. A long upper wick poking above $100 that closes at $99.60 is evidence of rejection — sellers met the excursion and pushed it back. A full-bodied candle closing at $101.50 is evidence of acceptance. Many disciplined traders refuse to treat any breakout as real until they see a full candle close beyond the level on their timeframe — and the more meaningful the timeframe (a daily close versus a 5-minute close), the stronger the evidence. This one filter, "closes, not wicks," removes a large fraction of trap entries at the cost of slightly later entries on real moves. That trade-off — worse price for better information — is usually worth taking.
The second tell comes after the break: what happens when price returns to the broken level? In a genuine breakout, the old ceiling becomes the new floor. Price breaks above $100, pulls back to $100.20, finds buyers, and turns up — the retest holds, demonstrating that participants now treat $100 as support. In a trap, the retest fails: price comes back to $100, slices through without hesitation, and the "support" evaporates because it was never real acceptance, only a stop cascade. Waiting for the retest costs you the trades that never pull back, but it gives you two gifts on the trades you do take: confirmation that the level flipped, and a natural, close-by invalidation point (if price closes back below the retested level, you are simply wrong, and your stop is small).
A breakout's location in the larger structure matters enormously. A breakout in the direction of a well-established trend, after a tight consolidation, has structural logic behind it: the trend's participants are adding, and the consolidation built orderly fuel. A breakout at the extreme of a long sideways range, against no trend, in quiet conditions, is statistically much more likely to be a sweep — because ranges are where the liquidity-pool dynamic is strongest (next section). Similarly, the speed and follow-through after the break carries information: real breakouts tend to extend and stay extended; traps tend to stall within a candle or two. None of these clues is certain alone. Stacked together — close beyond the level, successful retest, sensible context, real follow-through — they shift the odds meaningfully.
A trading range is a stretch where price oscillates between a defined floor and ceiling — say, bouncing between $48 and $52 for weeks. Ranges deserve their own section because they are where false breakouts are most frequent, and the reason follows directly from everything above.
Think about what a long range builds. Every touch of $52 creates shorts with stops just above $52. Every touch of $48 creates longs with stops just below $48. Week after week, the pools at both edges grow deeper — a range is essentially a machine for accumulating liquidity just beyond its own boundaries. Meanwhile, the range teaches everyone watching it that the edges hold, so each new touch attracts more participants playing the bounce, stacking more stops onto the pile.
The result is a market with dense, well-advertised order pools at both extremes and — this is the key — no committed trend pushing in either direction. When a sweep happens inside a trend, genuine directional demand can turn the sweep into a real breakout. In a trendless range, there is usually no such demand: price pokes above $52, fires the shorts' stops, finds no fresh buyers willing to own the asset above the range, and falls back inside. The pattern is so common that experienced range traders often expect the boundaries to be swept before any real move, and treat a clean poke-and-reclaim at a range extreme as a setup in itself rather than a surprise.
Two practical consequences for a beginner trading ranges. First, breakout entries at range boundaries are the lowest-quality version of breakout trading — demand more confirmation there (a strong close beyond, then a held retest), or skip them. Second, if you trade the bounces within a range, place your stop with the sweep in mind: not at the absolute minimum distance beyond the edge where it sits in the densest part of the pool, but beyond the zone where a typical sweep of that chart would reach — and size the position smaller to keep the same dollar risk, exactly as the position sizing formula dictates.
Abstract mechanics stick better with numbers attached. Both examples are hypothetical and simplified, but the structure is what you will see on real charts.
A stock has ranged between $48.00 and $52.00 for six weeks, with three clean bounces off $48. You can infer the order landscape: heavy long stop-losses from $47.90 down to about $47.40, plus breakdown shorts set to trigger below $47.80.
Monday: price drifts to $48.10. Tuesday, a red candle pushes through — price drops to $47.45 in under an hour, fast and vertical (the cascade: each layer of stops selling into the next). At $47.45 the fall just... stops. The next 5-minute candles are small, two-sided, going nowhere (absorption — the selling is being bought). Then price lifts, reclaims $48.00, and closes the day at $48.60 — a daily candle with a long lower wick, closed back inside the range.
Walk the participants. The longs who bought $48.20 with stops at $47.80 were stopped out at the lows — they sold the bottom. The breakdown shorts who entered at $47.75 are trapped the moment price reclaims $48.00; their buy-stops above fuel the recovery. And a trader who understood the anatomy had a different option entirely: after the reclaim of $48.00 (candle 3 — the close back inside), enter long around $48.20 with a stop below the sweep low at $47.30. Stop distance: $0.90. With a $100 risk budget: $100 ÷ $0.90 = 111 shares. The invalidation is structural and unambiguous — if price goes back below the sweep low, the reclaim failed and the idea is simply wrong. Whether this particular trade wins or loses, the risk was defined before entry and the loss, if it comes, is exactly one budgeted unit.
Same range, other edge, told from the loser's seat — because that seat teaches more. Price approaches $52.00 for the fourth time. A trader has been waiting weeks for "the breakout." A strong green 15-minute candle closes at $52.40 on a burst of volume. He buys $52.45, stop at $51.90 — risking $0.55 per share, 180 shares for his $100 budget. So far, textbook.
Candle 2: price tries $52.60, fades, closes $52.35 — a stall. Candle 3: a heavy red candle drives to $51.70 and closes there, back inside the range. His stop fills around $51.88 (a few cents of slippage in the rush — stops are market orders). Loss: about $103. Annoying, but here is the point: survivable and planned. He risked one budgeted unit on a defined idea with a defined invalidation, the idea was wrong, and he paid the pre-agreed price.
Now replay it without the discipline. Same entry, but no stop — "I'll watch it." At $51.70 he holds, because the range floor at $48 is "support." Two days later the range floor gets swept too (ranges often break properly after enough traps have cleared one side's fuel — in this case the failed upside move trapped buyers whose later selling pressured the range), and he finally exits at $47.60 in despair — a $4.85 per-share loss on 180 shares: roughly $873, nearly nine budgeted units, on a trade planned to risk one. The false breakout did not destroy him. The absence of pre-defined invalidation did.
Everything in this article points at one practice: decide, before you enter, the exact price at which your idea is provably wrong — and exit there, mechanically, every time.
False breakouts are precisely the environment where this discipline earns its keep, for two reasons. First, traps are designed (emergently or otherwise) to generate hope and rationalization at the worst moment. The candle-1 push looks so convincing that when candle 3 closes back inside, the trapped trader's brain floods with reasons to hold: "it'll come back," "the level is still basically intact," "I'll just wait for a bounce to exit." Pre-defined invalidation removes the decision from that compromised moment. The exit was decided by the calm version of you, at planning time; the stressed version of you merely executes it.
Second, structural invalidation is what makes the math work. Recall the core position sizing formula: size equals risk budget divided by stop distance. The formula is only honest if the stop is genuinely where the idea dies and genuinely gets honored. In example 1, the invalidation was "below the sweep low" — a price with objective meaning (the reclaim failed). In example 2, it was "back inside the range" — equally objective (the breakout failed). Neither was "where my loss reaches an amount I dislike." When invalidation is structural and respected, every trap costs exactly one planned unit, and a strategy can absorb being trapped repeatedly while staying intact. When invalidation is improvised, a single trap can do a month of damage — and the trap, remember, is the common case at obvious levels, not the rare one.
A short pre-entry ritual covers it: Where is the level? Where are the stops clustered beyond it (i.e., where would a sweep reach)? Where, beyond that, is my idea objectively dead? What size does my budget allow at that stop distance? If you cannot answer all four, you do not have a trade — you have a hunch with money attached.
Pattern recognition here is trainable, and you can train it without risking anything.
Mark levels in advance, then watch. On any chart, mark the obvious levels — range edges, equal lows, round numbers — and write down where you believe the stop clusters sit. Then watch (or scroll forward through history) and grade yourself: did the level hold cleanly, break cleanly, or get swept? After a few dozen reps you will be startled how often the answer is "swept."
Collect the anatomy. Build a small screenshot folder of push–stall–reclaim sequences. Note the timeframe, the context (range edge versus trend), where the wick stopped, and how far beyond the level the sweep reached. The "how far" question calibrates your future stop placement better than any rule of thumb.
Replay your own losses. If you have already been caught in false breakouts, journal them retrospectively: which candle did you enter on (it is almost always candle 1), what would the candle-close filter have changed, where did the candle-3 close occur, and what did exiting (or failing to exit) cost in budgeted units? Your own trapped trades are the highest-value study material you own.
Practice the patient versions. In simulation, take only retest entries for a few weeks — break, close beyond, return to the level, hold, then enter. You will miss some runners. Track honestly what the filter saves versus what it costs. Most beginners discover the trade-off heavily favors patience.
No — your individual stop is far too small for anyone to care about. What gets targeted (or simply gets swept by emergent order flow) is the cluster: thousands of stops stacked in the same obvious zone. The practical lesson is not paranoia, it is predictability: if your stop sits exactly where every textbook reader's stop sits — a few cents below the obvious level — it sits in the densest part of the pool. Place stops beyond where a typical sweep of that chart reaches, accept the wider distance, and reduce size to keep your dollar risk constant.
No. Real breakouts exist and drive some of the best directional moves on any chart; fading every break is its own losing strategy. The adjustment is evidentiary, not prohibitive: stop buying candle-1 pushes at obvious levels, and start demanding confirmation — a full candle close beyond the level on a meaningful timeframe, ideally a retest that holds, and context that supports continuation (trend, tight prior consolidation). You will enter later and at worse prices than the lucky candle-1 buyers on real breakouts. In exchange, you will skip a large fraction of the traps. Over many trades, that exchange is the entire point.
All of them — the logic is fractal because the cause (orders clustering beyond visible levels) exists on every timeframe. A 5-minute chart sweeps intraday stops; a daily chart sweeps swing traders' stops; a weekly chart sweeps investors' stops. Two practical notes: higher-timeframe closes carry more evidentiary weight (a daily close above resistance means more than a 5-minute close), and lower timeframes contain more noise, so sweeps there are more frequent but less individually meaningful. Whatever timeframe you trade, judge breaks by the closes of that timeframe or higher.
There is no universal number — it depends on the asset's volatility and how deep the stop cluster runs — and anyone who gives you a fixed percentage is guessing. The honest method is empirical: study the recent history of the specific chart you trade and measure past sweeps. Did pokes beyond range edges typically run 0.3%, 1%, 3% before reclaiming? Volatile assets sweep deeper. Use that measured distance for two things: placing your own stops beyond typical sweep depth, and judging when a "breakout" has extended far past sweep territory and is more likely real.
This is exactly the moment pre-defined invalidation exists for. If your plan said "I am wrong on a close back inside the level," then you already have your answer, and every minute of hesitation is the trap's psychology working on you — trapped traders holding and hoping are the fuel for the move against them. Exit, log it, and review later with a calm head. If you find you didn't define invalidation before entering, treat that as the real lesson: the problem was not this trade, it was entering without an exit plan, and the fix happens before the next entry, not during this one.
They mean levels are zones of interaction, not magic lines of protection. Support and resistance still matter enormously — they are where the orders are, which is exactly why sweeps happen there. The refined model is: levels attract activity; that activity includes stop clusters; stop clusters invite sweeps; and what happens after the sweep (reclaim or acceptance) is often more informative than the original touch of the level. Traders who understand this use levels with more nuance — expecting wicks beyond them, judging closes rather than touches, and treating a swept-and-reclaimed level as evidence rather than betrayal.
False breakouts make a lot more sense once stop placement and position sizing are second nature — the trap only ruins traders who let it. If you have not already, read our guide to risk management and position sizing, which covers the budget-and-formula system that makes every trap survivable. And if you want to practice spotting sweeps, reclaims, and range traps on real historical charts with guided lessons, that is exactly what ChartsQuest was built for — short interactive modules, one pattern at a time.
This article is educational content only, not financial advice; trading involves substantial risk of loss, and you should never trade with money you cannot afford to lose.
ChartsQuest est fourni à des fins éducatives uniquement. Rien ici ne constitue un conseil financier, juridique ou de trading.
Certains contenus de ChartsQuest peuvent être créés, édités ou accélérés à l'aide d'outils d'IA.
Les quatre figures les plus courantes : anatomie, règles de confirmation, objectifs mesurés et — surtout — comment chacune échoue.
Lire →
MembresUne approche structurelle des tendances : comment les définir objectivement, tradez les pullbacks plutôt que chasser, et reconnaître la cassure qui les termine.
Lire →
Que signifient les unités de temps, pourquoi le même marché raconte des histoires différentes selon l'échelle, et un workflow multi-échelles calme pour débutants.
Lire →