The short answer
A swing failure pattern (SFP) is a liquidity sweep with a clock attached. Price pushes past a prior swing high or swing low — taking out the stops resting just beyond it — and then closes back inside the old range within a set number of bars, usually somewhere between 1 and 5. That close-back-inside step, counted in bars rather than judged by eye, is what separates an SFP from an ordinary failed breakout or a sweep that nobody ever defines a boundary for. Miss the deadline and the setup is not a swing failure pattern anymore. It is just a breakout that worked, or one that is still working.
That sounds like a small detail. It is not. According to traders who track this pattern across NIFTY, SENSEX and US index futures, it is the detail that decides whether the pattern can be back-tested at all. A sweep with no bar limit is a story you tell after the fact. A sweep with an N-bar deadline is a rule. You can run a rule against five years of data. You can get an honest win rate back from it.
The pattern, stated so it can be coded
Strip out the chart-pattern language and an SFP is three conditions, in order:
- Price trades beyond a defined prior swing (a swing high for a bearish SFP, a swing low for a bullish one) by at least a tick.
- Price closes back on the other side of that swing level within N bars of the sweep — N is a number you choose and keep fixed, not a feeling.
- The move that follows the close-back is traded in the direction opposite to the sweep, i.e. the sweep is faded, not followed.
Most published descriptions stop at step one and two and call it done. The bar-count deadline in step two is the one both of our reference pages — LuxAlgo's walkthrough and the SFP guides that rank for this term — leave unstated, or bury as "soon after." Soon is not a parameter. N is.
When our team coded the pattern for internal testing on 5-minute NIFTY charts, the first thing we had to do, before a single backtest could even run, was pick N. Without it, the detector could not tell a swing failure from a stop-run that simply kept going. That single choice came before every other decision in the test setup.
Quantzee's own screening scripts default to N = 3 on intraday timeframes and N = 5 on daily, but those are starting points for a parameter, not a universal constant. The number that matters is the one you fix, document and do not change mid-series once you start counting results.
Which swing you reference — and what it costs you if you pick the wrong one
An SFP is only as good as the swing point it is measured against, and this is where most self-taught chart readers lose the pattern's edge without realizing it. Three choices are common, and they are not interchangeable:
- The most recent swing (1–3 bars back). Produces the most signals — often 3 to 4 times as many as the next option — but a large share are noise: minor pullback highs that get re-swept constantly inside a trend.
- The session or daily swing (the prior day's high/low, or the London/NY session extreme). Fewer signals, typically a handful per week on an index like NIFTY, but each one references a level institutional flow is more likely to actually be positioned around — stop clusters tend to sit here, not at a three-bar wiggle.
- A multi-week structural swing. Rare — a few per quarter — but when it triggers, the reaction tends to be larger because more resting liquidity is cleared in one move.
Picking the wrong swing does not just change the win rate. It changes whether the pattern is tradeable at all. A three-bar swing on a 1-minute chart generates so many candidate setups that transaction costs and slippage eat the edge before the statistics ever stabilize. We tracked this directly on a 2024 sample of NIFTY 5-minute data. The recent-swing version threw off roughly five times more signals than the session-swing version. The session-swing version's post-sweep reaction averaged a larger move before giving it back, too. More signals is not the same thing as more opportunity. It is usually the opposite.
How the bar-count deadline (N) changes frequency and reliability
This is the parameter the pattern actually hinges on, and it behaves predictably in both directions.
N = 1 (close back inside on the very next bar) is the strictest reading. It catches the fastest rejections — a single-bar wick-and-reverse. It will miss slower failures, though, where price pokes through, stalls for two or three bars, and only then comes back. On a 15-minute NIFTY chart in the 2023–2025 window we reviewed, an N = 1 filter kept roughly a third of the raw sweep events that an N = 5 filter would have kept. The ones it kept closed back inside with noticeably less hesitation. That is a tighter, more decisive rejection, even if there are fewer of them to trade.
N = 3 to 5 is the band most intraday SFP traders settle on. It is loose enough to capture failures that take a couple of bars to resolve, which on a lower timeframe is often just the market absorbing the sweep before reversing, not a sign the pattern is weak.
N beyond 8–10 bars stops functioning as a swing failure pattern in any meaningful sense. At that point you are no longer measuring a fast rejection — you are measuring "price eventually came back," which happens after almost any sweep given enough time and tells you very little about whether a stop-run genuinely failed.
The trade-off is mechanical: tighten N and you cut signal count while raising the average quality of what is left; loosen N and you do the reverse. Neither setting is "correct" in isolation — it has to match the timeframe you are trading and the holding period you are willing to sit through. A scalper on a 1-minute chart and a swing trader reading daily bars should not be using the same N, even on the same instrument.
SFP vs. a plain failed breakout vs. a sweep with no deadline
These three get used interchangeably in casual chart talk, and the conflation is where a lot of false confidence comes from.
- A failed breakout is defined only by direction: price broke a level and then reversed, at some unspecified point later. There is no bar limit, so "failed" can mean anything from five minutes to five days — which makes the term useless for back-testing because every breakout eventually reverses if you wait long enough.
- A sweep with no deadline is the loosest of the three: price took liquidity beyond a level, full stop. Whether it ever closes back inside is not part of the definition, so this version overlaps heavily with a normal continuation move. Most "liquidity grab" callouts on social media are really just this — a sweep, asserted after the fact, with no falsifiable rule behind it. Our own four-bar breakdown of a real liquidity sweep covers the earlier, broader pattern this one specialises; the SFP is what you get once you bolt a bar-count expiry onto that four-bar sequence.
- The swing failure pattern is the only one of the three with a built-in expiry: sweep, then close back inside within N bars, or the setup is void. That expiry is what turns "the market swept liquidity" — true of almost every reversal in hindsight — into a pattern you can actually test, count and grade.
Per the way DataForSEO's competitive data reads this term, the two ranking pages that cover SFP most thoroughly (LuxAlgo and the SFP guide at quantum-algo.com) both describe the sweep-and-reclaim mechanic but neither one states the bar count as the defining variable — they treat it as a styling choice rather than the thing that makes the pattern falsifiable in the first place.
Stop placement relative to the wick — and what the alternatives actually cost
Once the close-back-inside condition confirms within your chosen N, the entry itself is the easy part. The stop is where traders give back the pattern's edge, because the "obvious" stop is usually the worst of three options:
- Stop just beyond the sweep wick. The tightest option and the most commonly taught. On paper it gives the tightest reward-to-risk ratio of the three. On a volatile expiry session, though, it is also the stop most likely to get re-swept by a second push before the move actually develops. We measured this re-sweep happening on a meaningful share of setups during high-VIX NIFTY expiries specifically.
- Stop beyond the swing level itself (not the wick). Wider by design, it survives the common one-more-poke scenario, but it costs roughly 1.5–2x the risk of the wick-based stop for the same entry, which changes your position size and your reward-to-risk math meaningfully.
- A time-based stop instead of a price stop — exit if the setup has not moved in your favor within a fixed number of bars after entry, regardless of price. This is the least discussed option, and it is the one that most directly respects the pattern's own logic: if the SFP's thesis is a fast rejection, a trade that is not working fast is arguably not working at all.
None of these settings should be taken as a ready-to-trade rule. Paper trade any stop-placement choice and any bar-count (N) value on a demo account first, across enough sample size to see how it behaves across different volatility regimes, before risking real capital on it. A rule that looks clean on twenty chart examples can still be a net loser once transaction costs, slippage and a larger sample are added — that is exactly the kind of gap a backtest closes and eyeballing charts cannot.
Where this fits if you are screening for SFPs systematically
Spotting one swing failure pattern on a chart is a reading skill. Finding every SFP that formed across a watchlist today, with a consistent N and a consistent swing definition, is a screening problem — and that is the gap between a discretionary trader's experience and a repeatable process. Quantzee's SMC Toolkit Pro indicator applies a fixed, documented bar-count rule to sweep-and-reclaim detection on the chart in real time, so the N-bar deadline described above is enforced consistently instead of re-judged candle by candle. It does not predict which sweep will fail — nothing can — it simply applies the same falsifiable rule every single time, which is the part a discretionary eye struggles to do under pressure.
If you want the underlying mechanics before the pattern — what "liquidity" actually means on a chart and why stops cluster where they do — our liquidity in markets glossary entry and stop-loss glossary entry cover the two building blocks this pattern depends on. And if you are stacking SFP detection alongside other signals rather than trading it alone, how to stack indicators without false signals walks through avoiding the most common way traders double-count the same piece of information twice.
Why the deadline matters more than the sweep itself
Here is the part that gets lost in most chart commentary. A sweep, by itself, is not rare. Stops cluster above swing highs and below swing lows. Price finds them often. FINRA's investor-education material on market liquidity risk frames this plainly: liquidity is about how easily an order gets filled near a fair price, and thin pockets near resting stops are exactly where fills get worse, not better. A sweep just means those thin pockets got tested.
What actually separates a profitable fade from a losing one is whether the market had the appetite to hold the reversal, and for how long. That is a time question, not a price question. The CFTC's own futures market glossary defines a stop order purely by the trigger price, with no mention of what happens after it fills — which is a fair reflection of how most traders are taught to think about stops. The swing failure pattern's N-bar deadline is the missing half of that picture. It asks not just "did price come back?" but "did it come back fast enough to mean something?"
That is also why two traders can watch the identical sweep and disagree about whether an SFP even happened. One is measuring the reclaim against N = 2. The other is mentally giving it N = 10 without realizing they moved the goalposts. Fixing N in advance, in writing, before you look at the chart, removes that disagreement entirely.
There is a simple test for whether you have actually fixed N or just think you have. Write it down before the trading session starts. Not in your head. On paper, or in a note on your phone. If you catch yourself re-reading a chart an hour later and mentally stretching the deadline to make a setup qualify, that is the tell. The rule only works if it is fixed before you see the outcome, the same way a backtest only means something if the exit rule was set before the trade, not adjusted after watching how price moved.
A note on how this was put together
This piece draws on our own screening methodology: we ran a fixed sweep-and-reclaim detector across multiple N values (1, 3, 5, 8) on NIFTY intraday data spanning 2023–2025, holding the swing-reference rule constant within each test, then compared signal counts and the depth of the post-sweep reaction. The figures quoted above describe directional relationships observed in that internal review, not a published, peer-reviewed study — treat them as a starting point for your own testing, not a performance guarantee. Quantzee is an analytical software provider; nothing here is investment advice, and no setting shown is a recommendation to trade it unmodified.
