Turtle Soup: Fading the Failed Breakout
Turtle soup is the fade of a fresh breakout that fails fast: price pushes through an obvious high or low, takes the stops resting beyond it, and snaps back within a handful of candles, and the trade is the snap, short above a swept high or long below a swept low. The name belongs to the technique. The behavior it fades is older than any of the tools in this block: most breakouts of obvious levels fail, and the failure is tradeable because the traders caught on the wrong side have to get out.
The setup needs three ingredients, and the first is the pool. The level has to be obvious, a prior day's low, a prior week's high, an overnight range edge, the kind of level any chart shows by default. Obvious levels concentrate stops just beyond them, and concentrated stops are the reward the sweep is chasing. The second ingredient is the speed of failure. A level that breaks and builds value on the far side is a genuine breakout. A level that breaks by a small margin, holds for minutes, and reverses is a soup. The difference is acceptance, and acceptance is measured in closes, not wicks.
The Fade
The mechanics of the fade are the mechanics of the trap, viewed from the exit side. When price takes the stops above an obvious high, two crowds are created at once: the breakout buyers who entered on the break, and the trapped shorts whose stops just fired as market orders into the rally. Both crowds become fuel. The breakout buyers are the last holders of an expensive position with no follow-through behind it, and the former shorts, already out, watch the level fail and re-enter against it. When the reversal comes, it comes from both flows at once, which is why the fade's first move is often the fastest part of the day.
The entry discipline is the same one the whole block runs on. The sweep alone is not the signal. The signal is the failure to accept: price back on the near side of the level, with displacement away from the extreme. The tightest version enters on the retrace after that displacement, with the stop beyond the sweep extreme. The stop is structurally honest: if price gets back above the swept high and holds, the soup thesis is simply wrong, and the trade exits for a small, defined loss.
Timing is not decoration here. The classic soups print at the session's real opens, where the day's first directional act gets decided, because that is when someone with size has an interest in clearing the obvious levels before the real move. The same sweep in the dead stretch after lunch is far less likely to be either a trap or a launch; it is drift crossing a line.
One blunt warning: the fade is a fighter's trade. It stands in front of momentum, and the cost of being early is real. Traders who fade every break of every level are not trading turtle soup; they are donating to it. The setup earns its edge from the pool being obvious, the failure being fast, and the displacement being genuine. All three, every time, or no trade.
Soup or Breakdown
The decision between fading a break and respecting it is the skill, and the evidence sorts into a short checklist. Speed: a soup fails within minutes; a breakout that is going somewhere usually does not hand the level back on the same candles that took it. Depth: soups overshoot the level by a small margin, enough to touch the stops, not enough to build value; a drive through the level that keeps going is acceptance. Closes: a wick through and a close back on the near side is the soup's signature; consecutive closes on the far side are the breakdown's. Context: a sweep against the higher-timeframe direction, in a window where sweeps are normal, at a level the whole market can see, is the highest-grade soup the chart offers.
Volume sharpens the read when it is available. A sweep on a burst of volume that immediately dies, with the reversal candle carrying more commitment than the break itself, fits the trap. A break that holds its volume and keeps building is a move with participants, and fighting it because the level was obvious is a bias, not a read.
There is also the failed second test, which is where the fade most often gets punished. The first sweep of a level is the classic soup. Price often comes back for a second look, and a second sweep that holds can be a legitimate fade again, but a second sweep that keeps going is the market announcing that the first read was the wrong one. The rule is mechanical: each successive test of the same level weakens the fade, and a soup that fails twice in the same window is retired for the session.
A Worked Example: One Sweep, One Snap
The following numbers are invented for illustration, a hypothetical index with round levels. Nothing here describes a real session.
The prior day's low sits at 4,286, and the overnight session has been building value above it. In the New York morning window, price pushes through in one fast move, touches 4,279, seven points past the level, and stops. The next candle closes back above 4,286 and displaces upward, leaving a fair value gap behind the reversal. The fade goes on at 4,288 on the retrace into the gap. The stop sits below the sweep extreme at 4,273: 15 points of risk. The target is the old high near 4,340, the opposite side of the overnight range: 52 points, about 3.5 to 1.
| Element | Level | Role it played |
|---|---|---|
| Prior day low | 4,286 | The obvious level holding the stops |
| Sweep low | 4,279 | Seven points past the level, then failure |
| Entry | 4,288 | The retrace into the reversal gap |
| Stop | 4,273 | Beyond the sweep extreme |
| Target | 4,340 | The opposite pool of the range |
The invalidation is acceptance on the far side. Two consecutive closes below 4,286 after the entry, or any trade that holds beneath the sweep extreme, ends the fade. The distinction matters because a soup that comes back for a second test can still work, but a soup that gets accepted was a breakout wearing the wrong label, and the market has no obligation to hand back the loss.
The Limits of the Fade
The fade's edge decays in predictable places. Trend days are the first: a market that has already displaced through three levels in one direction will eat a fourth level for breakfast, and fading the fourth because it was obvious is ignoring everything the session said. News windows are the second: a repricing that gaps through a level carries no pool logic at all, and the bounce after it is continuation more often than failure. And thin sessions are the third: a sweep in dead hours may be nothing but drift, and drift has no trapped crowd to fuel the reversal.
What the fade shares with every tool in this block is the dependency chain. The pool must be obvious, the failure must be fast, the displacement must be real, and the higher-timeframe direction must not be screaming the other way. Remove any one and the setup is a story. Keep all of them, and the fade is the same trade the model and the divergence produce, viewed from the shortest distance: the market takes liquidity, finds none beyond it, and returns to where the real business is.
The next lesson moves from levels to the clock itself: the specific minutes of the day when this block's setups are statistically most likely to print.
Turtle Soup Questions
How far past the level should the sweep run?
Far enough to touch the stops resting beyond it, and no further. A marginal overshoot fits the trap; a drive that keeps building on the far side is acceptance. There is no fixed number, and the closes matter more than the depth: a wick past the level with closes back on the near side is the soup's shape.
Is the first sweep or a later one the better fade?
The first sweep of a level carries the cleanest edge, because the trapped crowd is fresh and the exits are largest. Later sweeps of the same level work less often, and a second or third failed fade in one window is a signal to stop trading the level, not to size up.
Can turtle soup be traded in quiet hours?
Rarely with an edge. The trap depends on a crowd being present to trap, and dead hours have neither the stops nor the follow-through. The same sweep at a session open or a macro window carries the participation that makes the reversal fast.
What timeframes suit the fade?
Low to medium intraday charts, where the failure shows within a handful of candles and the stop beyond the sweep extreme stays small. On higher timeframes the same behavior exists but the "fast failure" condition stretches into sessions, and the trade becomes a swing read on acceptance rather than a quick fade.
The fade depends on the hour as much as the level. The next lesson gives that hour a name: the fixed minutes when the day's algorithmic business concentrates.