Actualizado
Turtle Breakout Donchian Channel Graded & Measured
Turtle Breakout — Donchian Channel, Graded & Edge-Measured
The classic Donchian channel read as the Turtle breakout system, with an honesty layer bolted on. Instead of just plotting a channel and asserting the breakout works, this engine forward-tests its own breaks and shows you — on the chart — where they actually pay and where they don't.
WHAT IT IS
A confirmed close beyond the prior N-bar high is a bullish break; beyond the prior N-bar low, a bearish one (classic turtle entry, default 20 bars). The trend line that is drawn and ridden is the turtle EXIT channel (default 10 bars) — the trailing stop. Every break is graded 0-5 on how decisive the breaking bar was, can be gated to trending conditions, and carries a dynamic entry / stop / target plan.
THE MEASUREMENT (why this is not just another Donchian script)
Every armed break is resolved past-only under a chosen exit model and scored in R, per grade, as a Wilson-floored win-rate and net R after costs — shown beside two baselines:
- A RANDOM-DIRECTION NULL: the same bars and the same exit, but a coin-flip direction. If a grade does not beat the null, its "edge" is just drift and payoff geometry.
- A FADE: the opposite trade. If fading earns more than taking the break, the signal is better inverted.
A break only earns trust when it beats the null AND clears zero after costs.
WHAT THE MEASUREMENT REVEALS — AND THE SCRIPT SHOWS IT
Across timeframes, LOW-timeframe breakouts (roughly 1-5m) tend to LOSE, and lose to the null — they get whipsawed, and fading pays more. The breakout edge appears on HIGHER timeframes (about 1h / 4h / 1D), where it turns net-positive and beats the null — consistent with the original Turtle being a higher-timeframe method. The GRADE rates how decisive the breaking bar looked but does NOT reliably predict the outcome, so treat it as context, not a quality score. Load the measured-edge table and read it on YOUR market and timeframe before trusting a setup, and prefer higher timeframes for the with-trend breakout.
ALSO INCLUDED
- A 9-timeframe alignment table: trend, fresh break, grade, and the channel level on each of 1m / 3m / 5m / 15m / 30m / 1h / 2h / 4h / 1D.
- A regime gate: Kaufman efficiency ratio and/or higher-TF alignment, so breaks arm only in a trend.
- A dynamic trade plan: entry, a trailing stop (the turtle exit channel or a chandelier-ATR), and R- or ATR-based targets.
- A trend ribbon, and EXP_* data-window outputs for multi-indicator setups.
HOW TO USE
Add it to a chart — ideally a higher timeframe. Keep the breaks that pass the strength filter and the regime gate. Open the measured-edge table (research toggle) and read the ALL row against the NULL row; switch the exit model and watch which one turns the entry edge into a positive net R. Let the table accumulate enough resolved breaks before leaning on any number.
LIMITATIONS
All measured statistics are IN-SAMPLE on the current chart's own history — they are context and a transparency tool, NOT a backtest, a forecast, or a promise of future performance. Breaks confirm on bar close; higher-timeframe values use closed bars only (no look-ahead). Research and education only; not investment advice.
CONCEPTS USED (operationalised here, not third-party code): Donchian channel (Richard Donchian), the Turtle breakout system (Dennis & Eckhardt), ATR and the Chandelier exit (Welles Wilder / Chuck LeBeau), the efficiency ratio (Perry Kaufman), triple-barrier labelling (in the spirit of Marcos Lopez de Prado), and the Wilson score interval. The grading, regime gate, 9-timeframe tally, ribbon and the per-grade edge measurement are this script's own synthesis.
The classic Donchian channel read as the Turtle breakout system, with an honesty layer bolted on. Instead of just plotting a channel and asserting the breakout works, this engine forward-tests its own breaks and shows you — on the chart — where they actually pay and where they don't.
WHAT IT IS
A confirmed close beyond the prior N-bar high is a bullish break; beyond the prior N-bar low, a bearish one (classic turtle entry, default 20 bars). The trend line that is drawn and ridden is the turtle EXIT channel (default 10 bars) — the trailing stop. Every break is graded 0-5 on how decisive the breaking bar was, can be gated to trending conditions, and carries a dynamic entry / stop / target plan.
THE MEASUREMENT (why this is not just another Donchian script)
Every armed break is resolved past-only under a chosen exit model and scored in R, per grade, as a Wilson-floored win-rate and net R after costs — shown beside two baselines:
- A RANDOM-DIRECTION NULL: the same bars and the same exit, but a coin-flip direction. If a grade does not beat the null, its "edge" is just drift and payoff geometry.
- A FADE: the opposite trade. If fading earns more than taking the break, the signal is better inverted.
A break only earns trust when it beats the null AND clears zero after costs.
WHAT THE MEASUREMENT REVEALS — AND THE SCRIPT SHOWS IT
Across timeframes, LOW-timeframe breakouts (roughly 1-5m) tend to LOSE, and lose to the null — they get whipsawed, and fading pays more. The breakout edge appears on HIGHER timeframes (about 1h / 4h / 1D), where it turns net-positive and beats the null — consistent with the original Turtle being a higher-timeframe method. The GRADE rates how decisive the breaking bar looked but does NOT reliably predict the outcome, so treat it as context, not a quality score. Load the measured-edge table and read it on YOUR market and timeframe before trusting a setup, and prefer higher timeframes for the with-trend breakout.
ALSO INCLUDED
- A 9-timeframe alignment table: trend, fresh break, grade, and the channel level on each of 1m / 3m / 5m / 15m / 30m / 1h / 2h / 4h / 1D.
- A regime gate: Kaufman efficiency ratio and/or higher-TF alignment, so breaks arm only in a trend.
- A dynamic trade plan: entry, a trailing stop (the turtle exit channel or a chandelier-ATR), and R- or ATR-based targets.
- A trend ribbon, and EXP_* data-window outputs for multi-indicator setups.
HOW TO USE
Add it to a chart — ideally a higher timeframe. Keep the breaks that pass the strength filter and the regime gate. Open the measured-edge table (research toggle) and read the ALL row against the NULL row; switch the exit model and watch which one turns the entry edge into a positive net R. Let the table accumulate enough resolved breaks before leaning on any number.
LIMITATIONS
All measured statistics are IN-SAMPLE on the current chart's own history — they are context and a transparency tool, NOT a backtest, a forecast, or a promise of future performance. Breaks confirm on bar close; higher-timeframe values use closed bars only (no look-ahead). Research and education only; not investment advice.
CONCEPTS USED (operationalised here, not third-party code): Donchian channel (Richard Donchian), the Turtle breakout system (Dennis & Eckhardt), ATR and the Chandelier exit (Welles Wilder / Chuck LeBeau), the efficiency ratio (Perry Kaufman), triple-barrier labelling (in the spirit of Marcos Lopez de Prado), and the Wilson score interval. The grading, regime gate, 9-timeframe tally, ribbon and the per-grade edge measurement are this script's own synthesis.
Notas de prensa
MTF feature added for ease of visual refer
Exención de responsabilidad
La información y las publicaciones no constituyen, ni deben considerarse como, asesoramiento o recomendaciones financieras, de inversión, de trading u otro tipo, proporcionadas o respaldadas por TradingView. Obtenga más información en Condiciones de uso.
Exención de responsabilidad
La información y las publicaciones no constituyen, ni deben considerarse como, asesoramiento o recomendaciones financieras, de inversión, de trading u otro tipo, proporcionadas o respaldadas por TradingView. Obtenga más información en Condiciones de uso.