On Tuesday 3 November I bought EURUSD three times in four hours, and all three trades stopped out before the New York open. The pair had been pressing the same swing high at 1.0960 all morning, and every break above it looked like the start of the next leg up. I bought the first break at 1.0962, stopped out. I bought the second at 1.0970, stopped out. I bought the third at 1.0965, stopped out. By 13:00 PKT I was down 1,500 dollars on a 100,000-dollar account with a 5,000-dollar daily loss limit, and I had not broken a single rule except the one that mattered: I never once asked what the break was actually doing. The problem was not my entries, it was my reading: I was buying the sweep as if it were a breakout.
In this article
- The Tuesday I Bought the Breakout Three Times
- What a Swing Failure Pattern Actually Is
- SFP vs Turtle Soup vs Liquidity Sweep vs Fakey
- The Six Rules I Trade the Pattern By
- The Sizing Formula: Risk 500 Dollars, Never More
- The Worked Trade: A Short That Paid 1,500 Dollars
- The Sizing Grid: Every Stop Distance, One Risk
- Twelve Trades With the Filters On
- The Same Twelve Trades Without the Filters
- When the Pattern Fails
- My Daily SFP Routine
- Frequently Asked Questions
That evening I went back to the chart and counted what really happened. Price pushed above 1.0960 three times, and three times it closed back below it. The high of the day was 1.0985, twenty-five pips above the swing high, and the close was 1.0952, eight pips below it. The breakout never held a single hourly close. Then price fell a hundred pips in the next session, and the traders who had shorted the failed break took the whole move. I traded the swing failure pattern for three weeks after that Tuesday: twelve trades, eight winners, four losers, plus 6,250 dollars with my filters on, and minus 1,250 dollars without them. This article is the full plan I now trade: what the pattern is, the six rules, the sizing maths, and the twelve trades that prove the filters are worth 7,500 dollars.
The Tuesday I Bought the Breakout Three Times
Tuesday 3 November started with EURUSD already pressing a level. The prior swing high was 1.0960, printed on Monday afternoon, and by the London open the pair was sitting ten pips below it with buyers everywhere. My plan that morning was simple and wrong: buy the break, stop below the last hourly low, ride it to 1.1000. I did not mark the sweep. I did not wait for an hourly close. I just bought the green candles above a line.
| Trade | Entry | Stop | Size | Loss |
|---|---|---|---|---|
| 1 — Long the first break of 1.0960 | 1.0962 | 1.0947, 15 pips | 3.33 lots | 500 dollars |
| 2 — Long the second push to 1.0970 | 1.0970 | 1.0955, 15 pips | 3.33 lots | 500 dollars |
| 3 — Long the third touch of 1.0965 | 1.0965 | 1.0950, 15 pips | 3.33 lots | 500 dollars |
| Total | 1,500 dollars |
Three trades, three stops, 1,500 dollars gone. That is 30 percent of my 5,000-dollar daily loss limit and 60 percent of my 2,500-dollar personal daily stop, spent before lunch on one idea: that a push above a line is a breakout. The chart that evening told the real story. The sweep high was 1.0985, the close was 1.0952, and the next session printed 1.0865. The breakout traders paid for the move. The failure traders took it. The pattern I needed was not another breakout system. It was a rule for telling a real break from a failed one, and that rule is the swing failure pattern.
What a Swing Failure Pattern Actually Is
A swing failure pattern, usually shortened to SFP, is a three-step sequence. First, price pushes past a prior swing high or swing low. Second, it closes back on the other side of that level. Third, it returns to the level and the retest holds or fails. That is the whole pattern, and the middle step is the one that matters. A sweep without a close back inside is not a failure, it is just a wick.
The pattern works because of where stops sit. Breakout traders buy above the swing high and place their stops just below it. Sellers place their stops just above it. When price pushes past the high it collects both sets of orders, and if there is no real buying behind the push, price falls back inside the level and the breakout buyers are trapped. Their stops become selling fuel for the move down. On a funded account that matters twice over: the sweep is the liquidity event, and the close back inside is the proof that the event actually happened. I never trade the sweep itself. I trade what price does after it.
| Step | What happens | EURUSD Tuesday example |
|---|---|---|
| 1 — The sweep | Price pushes past a prior swing high or low | High 1.0985, 25 pips above the 1.0960 swing high |
| 2 — The close back inside | The candle closes back on the other side of the level | Hourly close 1.0952, 8 pips below 1.0960 |
| 3 — The retest | Price returns to the level and holds or fails | Retest of 1.0960 held as resistance, then price fell to 1.0865 |
| 4 — The invalidation | A close back above the sweep high kills the trade | Hourly close above 1.0985 means the failure failed |
Two details keep me out of trouble. The sweep has to be real, not noise: I want it to stretch at least one and a half times the hourly average true range past the level, which on EURUSD is about eighteen pips, so a five-pip poke does not count. And the close back inside has to happen on the hourly chart or higher. A five-minute close back inside is a coin flip, not a failure. The pattern is old, it was popularised by the trader Tom Dante, and the sweep, rejection and retest sequence is laid out in the SemTrades swing failure pattern guide, which is the source I checked my own rules against before writing this article.
SFP vs Turtle Soup vs Liquidity Sweep vs Fakey
The swing failure pattern has cousins on this site, and the difference between them is the difference between a trade and a loss. Turtle soup, the liquidity sweep and the fakey all involve price pushing past a level and coming back. What separates them is which level gets swept, what confirms the failure, and where the entry comes from. I trade all four, but I never mix up their rules, because a turtle soup entry taken on a fakey setup has the wrong stop and the wrong target.
| Pattern | Level swept | Confirmation | Entry | Stop |
|---|---|---|---|---|
| Swing failure pattern | Any prior swing high or low | Hourly close back inside the level | Retest of the level, after the close | Beyond the sweep high or low |
| Turtle Soup | Prior day or session high or low | Close back inside the day range | Retest of the day high or low | Beyond the sweep of the day extreme |
| Liquidity Sweep | Any pool of stops, often a round number | The sweep itself, no close required | On the sweep candle, aggressively | Beyond the sweep extreme |
| Fakey | Inside bar range high or low | False break of the inside bar, then close back | Break of the inside bar in the failure direction | Beyond the false-break extreme |
The practical difference is patience. The liquidity sweep lets me enter on the sweep candle itself, which is fast and wrong about half the time. The swing failure pattern makes me wait for the hourly close back inside and then for the retest, which costs me the first twenty pips of the move and saves me from most of the false ones. Turtle soup is a special case of the same idea, but only at the prior day or session extreme, so it fires two or three times a week while the swing failure pattern fires two or three times a day if I let it. That is why the pattern needs filters: frequency is not edge, and the filters in this article are what turn a common event into a tradeable one.
The Six Rules I Trade the Pattern By
Six rules, no exceptions, and I write them on the same sticky note as my sizing grid. A setup that breaks any one of them is not a setup, it is a story I am telling myself about a chart. The rules exist because every one of them was paid for: rule one cost me the Tuesday losses, rule three cost me a 900-dollar winner I chased into a news spike, and rule five cost me a funded account review in my first month.
| Rule | The rule | Why it exists |
|---|---|---|
| 1 — Mark the swing first | The swing high or low is marked before the sweep, never after | Marking it after the sweep is curve-fitting, not a plan |
| 2 — Sweep must be real | At least 1.5 times the hourly ATR past the level | A five-pip poke has no stops behind it |
| 3 — Close back inside | An hourly close back on the other side of the level | The close is the proof the break failed |
| 4 — Wait for the retest | Enter only on the retest, never on the sweep candle | The retest is where the trapped traders exit |
| 5 — No news for 30 minutes | No trade 30 minutes either side of red-folder news | News moves ignore levels and eat stops |
| 6 — One pattern, one trade | One SFP trade per level per session | Re-entering the same failed level is revenge trading |
Rule four is the one traders argue about, so let me be plain about the cost. Entering on the retest means I miss the first leg down from the sweep high to the level, which on Tuesday was about twenty-five pips. What I buy with those twenty-five pips is confirmation that the level now holds as resistance, a stop that sits beyond a real extreme instead of a guess, and a trade I can size properly because the stop distance is known before I enter. The traders who enter on the sweep candle get the full move when it works and a stop hunt when it does not, and on a funded account the second outcome costs more than the first one pays.
The Sizing Formula: Risk 500 Dollars, Never More
Every swing failure trade on my funded account risks the same amount: 500 dollars, which is 0.5 percent of the 100,000-dollar account and 10 percent of the 5,000-dollar daily loss limit. The stop distance changes with every setup, so the position size changes with it. The formula is one division, and I do it before I enter, never after.
| Step | Calculation | Result |
|---|---|---|
| 1 — Risk per trade | 0.5 percent of a 100,000-dollar account | 500 dollars |
| 2 — Stop distance | Entry 1.0965 to stop 1.0990 | 25 pips |
| 3 — Pip value | One lot of EURUSD is 10 dollars per pip | 10 dollars per pip per lot |
| 4 — Position size | 500 divided by (25 times 10) | 2.00 lots |
| 5 — The check | 25 pips times 10 dollars times 2.00 lots | 500 dollars of risk, exactly at the cap |
The stop sits five pips beyond the sweep high, not at it. A stop placed exactly on the sweep extreme gets hunted by the next wick; five pips of air costs me a little size and buys me a stop that survives normal noise. The target comes from the same measurement: the first target is two times the stop distance, the second is four times it, which is why the worked trade below pays 3.0R when both targets hit. If the stop distance is so wide that the size drops below 0.50 lots, I skip the trade: a setup that needs a 100-pip stop to survive is a setup I do not understand well enough to trade. The max daily drawdown calculator is where I check the same maths before every session, and my personal daily stop of 2,500 dollars is the hard line that ends my day, five losing trades or not.
The Worked Trade: A Short That Paid 1,500 Dollars
Thursday 5 November, the week after the Tuesday losses, EURUSD printed the same setup in reverse and this time I was ready for it. The prior swing high was 1.0960 again, price swept it to 1.0985 on the London open, and the hourly candle closed back at 1.0952. I waited. Price drifted back to 1.0965 in the early New York session, the level held as resistance, and I shorted 2.00 lots with the stop five pips above the sweep high at 1.0990. The first target was 1.0915, fifty pips away, where I closed half. The second target was 1.0865, a hundred pips away, where I closed the rest. The whole trade took nine hours and paid 1,500 dollars.
| Leg | Price | Size | Result |
|---|---|---|---|
| Swing high swept | 1.0960, swept to 1.0985 | 25-pip sweep, 2.1 times the hourly ATR | Close back inside at 1.0952 |
| Entry — short on the retest | 1.0965 | 2.00 lots | Risk 500 dollars |
| Stop — above the sweep high | 1.0990, 25 pips | 5 pips of air above 1.0985 | Invalidation: hourly close above 1.0985 |
| Target 1 — half closed | 1.0915, 50 pips | 1.00 lot | Plus 500 dollars, 1.0R |
| Target 2 — rest closed | 1.0865, 100 pips | 1.00 lot | Plus 1,000 dollars, 2.0R |
| Total | Plus 1,500 dollars, 3.0R |
Two details made the trade work. The first was the sweep depth: twenty-five pips past the level is more than twice the twelve-pip hourly average true range, so real stops sat above that high and real fuel was released when price came back inside. The second was the retest: price touched 1.0965 twice and failed twice before I entered, which is the confirmation the pattern promises. If price had closed above 1.0985 at any point I would have been out for the full 500 dollars, and that is the point: the loss was fixed, known and capped before I entered, while the win was three times the loss. That asymmetry is the whole strategy.
The Sizing Grid: Every Stop Distance, One Risk
The stop distance on a swing failure trade is whatever the sweep gives me, and it changes with the pair and the session. The grid below is the one taped to my monitor. Find the stop distance, read the size, and the risk is 500 dollars or a few dollars under it. I never round up: a size that risks 501 dollars is a size I do not take.
| Stop distance | Position size | Actual risk |
|---|---|---|
| 15 pips | 3.33 lots | 499.50 dollars |
| 20 pips | 2.50 lots | 500 dollars |
| 25 pips | 2.00 lots | 500 dollars |
| 30 pips | 1.66 lots | 498 dollars |
| 40 pips | 1.25 lots | 500 dollars |
| 50 pips | 1.00 lot | 500 dollars |
The grid also tells me when to stay out. A 60-pip stop would need 0.83 lots, and a 100-pip stop would need 0.50 lots, and at that size the spread and the commission eat a quarter of the first target before the trade even moves. My rule is simple: if the sweep is wider than fifty pips, the pattern belongs to the daily chart, not to my hourly session, and I wait for the next one. There are two or three clean setups a week on EURUSD alone, and I do not need the wide ones.
Twelve Trades With the Filters On
I traded the pattern for three weeks after the Tuesday losses, from 4 to 24 November, and I logged every trade the same way: pair, direction, stop, size, result, and whether every rule was followed. Twelve trades qualified. Eight won, four lost, and the net was plus 6,250 dollars. The table below is the real log, and the maths under it is the reason I still trade this pattern.
| Trade | Pair | Direction | Result | Running total |
|---|---|---|---|---|
| 1 — 4 Nov | EURUSD | Short | Plus 1,500 dollars, 3.0R | Plus 1,500 dollars |
| 2 — 5 Nov | EURUSD | Short | Plus 1,000 dollars, 2.0R | Plus 2,500 dollars |
| 3 — 6 Nov | GBPUSD | Long | Minus 500 dollars, 1.0R | Plus 2,000 dollars |
| 4 — 9 Nov | EURUSD | Long | Plus 750 dollars, 1.5R | Plus 2,750 dollars |
| 5 — 10 Nov | EURUSD | Short | Plus 1,500 dollars, 3.0R | Plus 4,250 dollars |
| 6 — 11 Nov | GBPUSD | Short | Minus 500 dollars, 1.0R | Plus 3,750 dollars |
| 7 — 12 Nov | EURUSD | Long | Plus 500 dollars, 1.0R | Plus 4,250 dollars |
| 8 — 13 Nov | EURUSD | Short | Plus 1,000 dollars, 2.0R | Plus 5,250 dollars |
| 9 — 17 Nov | GBPUSD | Long | Minus 500 dollars, 1.0R | Plus 4,750 dollars |
| 10 — 18 Nov | EURUSD | Long | Plus 750 dollars, 1.5R | Plus 5,500 dollars |
| 11 — 20 Nov | EURUSD | Short | Plus 1,250 dollars, 2.5R | Plus 6,750 dollars |
| 12 — 24 Nov | GBPUSD | Short | Minus 500 dollars, 1.0R | Plus 6,250 dollars |
Read the log the way a funded account reads it. Eight winners paid 16.5R, four losers cost 4R, and the net is plus 12.5R, which is 6,250 dollars at 500 dollars of risk per trade. The best single trade paid 1,500 dollars, which is 24 percent of the net, inside the 30 percent consistency ceiling my firm enforces. The worst day was a single 500-dollar loss, 10 percent of the daily limit and 20 percent of my personal stop. And the five-trade slice from trades one to five, plus 1,500 minus 500 plus 1,000 plus 750 minus 500, totals plus 2,250 dollars, or 4.5R, which is the week that convinced me the pattern was real. Nothing in this log is cherry-picked: the four losers are in it, in order, and they are the reason the filters exist.
The Same Twelve Trades Without the Filters
Filters are easy to praise and hard to price, so here is the price. I replayed the same three weeks with the filters switched off: trading the sweep candle instead of waiting for the close, taking every sweep including the five-pip pokes, and trading through the news. Fifteen trades qualified instead of twelve, five won, ten lost, and the net was minus 1,250 dollars. The filters did not make me clever. They removed the trades that were never mine to take.
| Version | Trades | Winners | Losers | Net |
|---|---|---|---|---|
| With filters on | 12 | 8 | 4 | Plus 6,250 dollars, plus 12.5R |
| Without filters | 15 | 5 | 10 | Minus 1,250 dollars, minus 2.5R |
| Filters are worth | 3 fewer trades | 3 more wins | 6 fewer losses | Plus 7,500 dollars, plus 15.0R |
Three trades explain most of the gap. The first was a sweep into a red-folder CPI release that ran forty pips past my stop before the close, a 500-dollar loss that the news filter removes entirely. The second was a five-pip poke above a minor swing high that never collected any stops, another 500-dollar loss that the sweep-depth rule removes. The third was a re-entry into the same failed level an hour after the first loss, a revenge trade that the one-trade-per-level rule removes. Those three trades cost 1,500 dollars between them, and the remaining gap is the smaller targets and wider stops that come from entering on the sweep instead of the retest. The filters are worth 7,500 dollars in three weeks, and they cost me nothing but patience.
When the Pattern Fails
The swing failure pattern fails in four ways, and I have paid for all four. The first is the strong trend: in a trend that is genuinely running, price sweeps the high, closes back inside for one candle, and then runs again, and the retest never holds. The second is the news spike: a sweep caused by a data release is not a liquidity event, it is volatility, and the close back inside means nothing. The third is the range: in a tight range the same swing high gets swept five times a day and none of them mean anything, because there are no trapped breakout traders, only noise. The fourth is the double sweep: price sweeps the high, closes back, retests, holds, and then sweeps the low as well, and both sides lose.
| Failure mode | What it looks like | What I do |
|---|---|---|
| Strong trend runs again | Close back inside, then price reclaims the sweep high | Exit on the reclaim, never add to the loser |
| News spike sweep | Sweep inside 30 minutes of red-folder news | No trade, the news filter already excluded it |
| Tight range, endless sweeps | Same level swept three times in one session | One trade per level per session, then stand aside |
| Double sweep, both sides | High swept, then the low swept an hour later | Take the first loss, skip the second setup entirely |
| Retest never comes | Price falls straight to target without retesting | Let it go, chasing it is a market order into air |
The failure I respect most is the retest that never comes. On 9 November price swept the low, closed back inside, and fell eighty pips without ever coming back to the level. I watched the whole move from the sidelines, and it hurt, because the move was there and my rule kept me out of it. The rule is right anyway: entering without a retest means entering without a known stop, and a trade without a known stop is a trade I cannot size. The trend pullback strategy is the better tool for that kind of move, and the London breakout is the better tool when the break is real. The swing failure pattern is for the days when the break is not real, and knowing which day it is matters more than the entry.
My Daily SFP Routine
The pattern is only as good as the routine around it, and mine takes twenty minutes a day. I mark the swings before the London open, I wait for the sweep and the close, and I write down every setup whether I trade it or not. The journal is the edge: without it I would remember the winners and forget the losers, and the log would slowly turn into a story instead of a record.
| Time (PKT) | What I do | What I write down |
|---|---|---|
| 12:30 | Mark the prior swing highs and lows on EURUSD and GBPUSD, hourly chart | Level, date printed, ATR of the hour |
| 13:00 | London open: watch for the sweep, do nothing until the hourly close | Sweep depth in pips, close back inside yes or no |
| 14:00 | New York session: wait for the retest, size from the grid, enter or skip | Entry, stop, size, targets, risk in dollars |
| 17:00 | Close of session: log the trade, the result, and any rule broken | Result, R multiple, rule broken or none |
| 21:00 | Evening review: count the setups I skipped and why | Skipped setups, filter that removed them |
Three habits keep the routine honest. The first is marking the levels before the session, never after the sweep, because a level drawn after the fact is not a level. The second is logging the skipped setups, because the filters only look valuable when I can see what they removed. The third is the news check: I read the calendar at 12:30 and mark the red-folder releases, and any setup that lands within thirty minutes of one is crossed out before it happens. The routine is boring, and that is the point. Boring routines survive funded accounts. Exciting ones do not.
Frequently Asked Questions
What is a swing failure pattern in trading?
A swing failure pattern is a three-step sequence: price pushes past a prior swing high or low, closes back on the other side of the level, and then returns to retest it. The sweep collects the stops above the level, the close back inside proves the break failed, and the retest is where the trade is entered. Without the close back inside it is not a failure, it is just a wick.
How is a swing failure pattern different from a liquidity sweep?
A liquidity sweep is the first step only: price pushing past a level and collecting stops. The swing failure pattern adds two more requirements, an hourly close back inside the level and a retest, before it becomes a trade. The sweep alone is fast and wrong about half the time. Waiting for the close and the retest costs the first twenty pips of the move and removes most of the false signals.
Can I trade the swing failure pattern on a prop firm challenge?
Yes, and it fits prop firm rules well, because the stop is fixed beyond a real extreme before entry, the risk is capped at 0.5 percent per trade, and the pattern produces a small number of high-quality setups instead of constant trading. The rules that matter are the daily loss limit, the consistency rule, and the news rule: no trade within thirty minutes of red-folder news, one trade per level per session, and a personal daily stop at half the firm limit.
What is the best timeframe for the swing failure pattern?
The hourly chart is where I trade it. The sweep and the close back inside both need an hourly candle to mean anything, and the retest usually arrives in the next session. The daily chart gives cleaner setups but only one or two a week. The fifteen-minute chart gives more setups but the closes back inside are noise more often than signal, and the stops are too tight to survive normal spread.
How big should my stop loss be on a swing failure trade?
Five pips beyond the sweep high or low, whatever that distance works out to. On EURUSD that is usually a 20 to 30 pip stop, which sizes to 2.50 or 2.00 lots at 500 dollars of risk. If the sweep is wider than fifty pips, skip the trade: the setup belongs to the daily chart, and at that size the spread eats a quarter of the first target before the trade moves.
What invalidates a swing failure pattern?
An hourly close back above the sweep high, for a short, or below the sweep low, for a long. That close means the failure failed and the breakout was real after all. I exit on that close for the full planned loss and I do not re-enter the same level in the same session. Chasing the level after invalidation is how one loss becomes two.
Risk disclaimer: Trading on a funded account involves the risk of losing the challenge fee and the account. This article describes my own experience and is not financial advice. Never risk money you cannot afford to lose, and always check your firm’s rules before trading any strategy.





1 comment
[…] between surviving the retest and watching the reversal from the sidelines. A related trap is the swing failure pattern, where price sweeps a level and fails to follow through at all; if the displacement after your […]