The short answer
ICT Turtle Soup is a reversal setup that trades the failed breakout, the moment price pushes through an obvious level to trigger breakout entries and stop losses, then snaps back and traps everyone who chased the break. It is one of the oldest liquidity plays in trading, and it has a real history most pages skip.
The name is not ICT's invention. Linda Raschke and Larry Connors coined Turtle Soup in their 1995 book Street Smarts as a tongue-in-cheek jab at the famous Turtle Traders' breakout system, and the original rule faded a failed twenty-day breakout.
The ICT version adapts the same idea to liquidity. Price sweeps a buy-side or sell-side liquidity pool, fails to continue, and closes back inside the range, and that sweep-and-fail is the trigger I wait for.
If the method is new, the Smart Money Concepts hub shows where this sits in the framework.
Where the name comes from
The Turtle Traders were a famous experiment run by Richard Dennis and William Eckhardt in the 1980s, which taught a mechanical breakout system to a group of novices. The system bought twenty-day breakouts and sold twenty-day breakdowns, and for a time it printed money.
Raschke and Connors noticed that many of those breakouts failed, and they built a counter-strategy that took the other side. They called it Turtle Soup, because it made its money out of the Turtles' failed entries, and they published the rules in Street Smarts in 1995.
I lean on that history because it is the one part of this setup that is genuinely documented, in a real book by named authors, rather than lore passed around forums. The ICT reading layers on top of that older idea.
The original Turtle Soup rule
The original Raschke rule was precise and mechanical, which is why it survived thirty years. Price makes a new twenty-day low, then closes back above the prior twenty-day low within a few bars, and you fade the breakdown by buying the reversal.
The mirror applied to highs. Price makes a new twenty-day high, fails, closes back below the prior high, and you sell the failed breakout.
The setup worked because breakouts fail often enough that fading the obvious ones, with a tight stop, paid over a large sample.
I keep the original rule in mind because it tells you what the setup really is, a fade of failed breakout trading, not a prediction of where price is going. The edge is in the failure, not in the direction.
The ICT version: sweep, fail, reverse
The ICT adaptation swaps the twenty-day rule for a liquidity level. Instead of a twenty-day high or low, ICT traders mark an obvious swing high or low where buy stops or sell stops cluster, which is the buy-side or sell-side liquidity.
Price sweeps the level, taking the resting stops and triggering the breakout entries, then fails to continue and closes back inside the prior range. The close back inside is the rejection, and the entry follows that rejection in the opposite direction.
The ICT layer adds the liquidity read on top of the old failed-breakout logic. Where Raschke faded a chart breakout, ICT fades a liquidity sweep, and the mechanism that makes both reverse is the same.
Turtle Soup versus a liquidity sweep or grab
These terms get tangled, so I separate them once. A liquidity pool is the resting orders, a sweep is the event of price taking them, and Turtle Soup is the specific trade you take when the sweep fails and reverses.
The general sweep and how to trade its reversal is covered on the liquidity grab page, which owns the mechanic. This page owns the structured Turtle Soup setup, which adds the failed-breakout lineage and a defined entry, stop, and target on top.
The distinction that matters is the failure. A sweep that keeps going is a genuine breakout and you follow it, while a sweep that closes back inside is a Turtle Soup candidate and you fade it, and telling those two apart is the whole skill.
| Term | What it means | What you do |
|---|---|---|
| Liquidity pool | The resting stops at a level | Mark it, do not trade it yet |
| Sweep or grab | Price takes the pool | Wait to see if it fails |
| Turtle Soup | The sweep fails and closes back inside | Fade the failed break |
| Run | The sweep keeps going | Do not fade, it is a real breakout |
The Turtle Soup entry, stop, and target
The entry comes on the close back inside the range, ideally confirmed by a lower-timeframe shift in your favour. I do not enter on the sweep itself, because the sweep is the trap being set, not the signal to act.
The stop sits just beyond the sweep extreme, because a clean close beyond that extreme means the break was real and the setup is invalid. The risk is the distance from entry to that stop, which is one unit.
The target is the next opposing pool of liquidity, which is where the reversal is said to be heading. A long off a failed breakdown aims at the buy-side liquidity above, and a short off a failed breakout aims at the sell-side liquidity below.
A worked example: fading a failed breakdown
The setup is clearest with a real shape, so here is a bullish Turtle Soup long, step by step. The numbers are illustrative, but the ratios are how I actually manage the trade.
Suppose a currency pair prints an obvious equal low at 1.0850 beside a prior swing low, stacking dense sell-side liquidity just below the market. Price dips into the New York open, sweeps to 1.0835 to clear the sell stops, and snaps back to close the four-hour candle at 1.0855, back above the level.
The sweep and the close back inside are the setup. On the fifteen-minute chart price shifts up and breaks a minor lower high, and I enter long at 1.0858 with a stop at 1.0830, just below the sweep low, risking twenty-eight points.
The target is the buy-side liquidity at the prior swing high near 1.0940, which is over eighty points from entry and better than two and a half to one. If price closes back below the sweep low instead, the stop removes me for a one-unit loss and I wait for the next setup.
Where Turtle Soup fails
The single biggest risk in this setup is the genuine breakout, the sweep that does not fail and keeps right on going. Fading a real breakout is how Turtle Soup traders lose, because the move that looked like a trap was the start of a trend.
I protect against that with the close back inside the range, which is the evidence the break failed. A sweep that holds beyond the level and does not close back inside is not a Turtle Soup, it is a run, and forcing a fade on it is the error that pays the most tuition.
The second failure is the sweep that partial-fills, where price closes back inside just enough to look like a rejection and then resumes the break. A tighter confirmation on the lower timeframe, waiting for a real shift rather than a single wick, filters most of those out.
What is actually proven about Turtle Soup
The honesty panel matters here because the setup straddles heritage and doctrine. The original Raschke failed-breakout fade is documented in a real 1995 book by named authors, and the broader fact that breakouts fail often is observable on any chart.
The mechanism behind the reversal has primary research behind it. Carol Osler's work at the Federal Reserve Bank of New York found that stop-loss orders cluster at obvious levels and that triggering them produces the sharp moves and snapbacks the setup trades, which is the documented engine of the failed break.
What is not proven is the ICT-specific layer. The claim that institutions deliberately engineer these sweeps to trap breakout traders is a plausible story that fits the observation, but it is doctrine rather than a documented mechanism, and no peer-reviewed study validates the specific ICT Turtle Soup rules or their win rate.
Common mistakes
The losses on this setup come from the same short list of errors, and most of them are timing mistakes. These four are the ones I see most.
Fading every sweep is the first. Not every sweep fails, and entering on the sweep with no close-back-inside confirmation is how you fade a genuine breakout and get run over.
Ignoring the higher timeframe is the second. A failed breakdown inside a higher-timeframe downtrend is counter-trend, and those fail more often than they pay, so I trade Turtle Soup only in the direction of the prevailing trend.
Placing the stop at the level instead of beyond the sweep is the third. The sweep wick often extends past the obvious level, and a stop there gets picked off before the reversal runs.
Treating the setup as guaranteed is the fourth. Even the documented failed-breakout edge loses on the real breakouts, so a Turtle Soup without a stop and a risk plan is a hope, not a strategy.
