Market Structure (ICT)
Turtle Soup
A close beyond the prior 20-bar high or low is the exact entry signal the original Turtle system trades. Turtle Soup takes the other side of that same signal, betting the breakout is a trap.
Two ledgers, one signal, opposite sides
The tape prints live and a rolling 20-bar Donchian channel (gold lines) recomputes on every bar. Every time a close prints outside that channel, two trades open at the same price: a Turtle trade that follows the breakout direction, and a Soup trade that fades it. Both are marked to a fixed 8-bar forward exit, and the outcome of each pair is always mirror-image opposite, since one side's gain is the other's loss on the same price move. Switch the regime toggle to rebuild the tape with different underlying drift and volatility, and watch which ledger tends to come out ahead.
Loading tape…
Cumulative % return per ledger, one point per resolved 8-bar trade
How it works
- The signal is mechanical: a close beyond the last 20 bars. The channel top is the highest high of the 20 bars before the current one, the channel bottom is the lowest low. When a close prints above the top or below the bottom, that is the breakout, the same trigger the original Turtle Trading System uses to enter.
- Twenty bars is a convention, not a magic number. The original Turtle rules used a 20-day channel for one system and a 55-day channel for a slower one. Nothing about market structure requires exactly 20, it's a lookback that was picked and back-tested, and shorter or longer windows shift how often signals fire and how reliable they are.
- Turtle Soup fades the exact same trigger the Turtle system would take. Larry Williams' framing is that a fresh extreme often gets bought or sold by traders reacting to the breakout itself, and once that reactive volume is spent, price often snaps back inside the range it just left.
- Both ledgers are graded the same way: price change over the next 8 bars. Turtle P&L is the move in the breakout direction, Soup P&L is the same move measured against the fade. They are exact mirror images of each other by construction, so this tool never rigs one side to look better, it measures the same eight bars twice.
- Which ledger wins is regime-dependent, not fixed. A trending tape tends to reward following the breakout, because the move that triggered the channel break often keeps going. A choppy, range-bound tape tends to reward fading it, because breakouts in chop are more likely to be false starts that snap back. Neither ledger wins every single trade in either regime, this is a tendency across many signals, not a guarantee on any one of them.
Where this breaks
Regime is only obvious after the fact
This tool lets you pick the regime ahead of time and watch the expected ledger pull ahead, which makes the edge look cleaner than it is in live trading. In practice nobody rings a bell to announce that the next 50 bars will be trending or choppy, and a market can flip from one to the other mid-stream with no warning, turning a string of winning Soup fades into a string of Turtle-favoring breakouts overnight. This chart is also price-only: it can measure the size of the breakout and how the channel and candles are structured, but it has no way to see real volume or which participants were actually behind the move, which on a live chart matters a great deal for judging whether a breakout has real conviction behind it or not. A mechanical 20-bar rule applied blindly, without any read on regime or participation, is closer to a coin flip with extra steps than an edge.