Обновлено
Ehlers Roofing Filter [RC Tools]
RC Tools — Ehlers Roofing Filter
────────────────────────────────────────────────────────────────────
█ OVERVIEW
Price is three things layered on top of each other: a slow trend drift, a band of cyclical swings, and high-frequency noise. Most oscillators blur all three together. John Ehlers' Roofing Filter separates them. It removes the slow drift, then removes the bar-to-bar noise, and leaves only the middle band: the market's cyclical component, as a clean zero-centred oscillator.
█ WHAT IT DOES
Runs price through a two-pole high-pass filter (strips the trend) followed by a two-pole Super Smoother low-pass filter (strips the noise), then divides the result by its own recent root-mean-square so the reading is in comparable units on any asset. Classifies each confirmed bar into one of four phases — Positive & Rising, Positive & Fading, Negative & Falling, Negative & Recovering — colours the background accordingly, and shows a table with the current phase, how long it has lasted, the reading, and the active passband.
█ THE THEORY BEHIND IT
A standard oscillator such as RSI or a detrended price series still carries slow drift from the underlying trend, so its zero line or midline means different things at different times. A high-pass filter removes that drift by discarding cycles longer than a chosen cutoff. That alone leaves jagged, noisy output, so a low-pass Super Smoother, a two-pole filter that smooths with less lag than a simple moving average of similar length, removes cycles shorter than a second cutoff. What survives is a "passband" of cycle lengths: the swings long enough to matter and short enough not to be trend. In Ehlers' terminology, the filter's output "roofs" over both ends of the spectrum.
█ HOW IT IS CALCULATED
1. Two-pole high-pass filter on price, with a cutoff period of 48 bars by default (cycles longer than this are removed as drift).
2. Two-pole Super Smoother low-pass filter on the high-pass output, with a cutoff of 10 bars by default (cycles shorter than this are removed as noise).
3. Divide the filtered value by its own root-mean-square over a rolling window (100 bars by default) so the output is in comparable units.
4. Classify: above zero and higher than the previous confirmed bar = Positive & Rising; above zero and lower = Positive & Fading; below zero and lower = Negative & Falling; below zero and higher = Negative & Recovering.
State classification occurs ONLY on confirmed bar close. The plot, background and table update together on closed bars, so nothing flips mid-bar.
█ SETTINGS & CONFIGURATION
• Source (default close)
• High-Pass Cutoff (default 48 bars) — shorter removes trend more aggressively
• Low-Pass Cutoff (default 10 bars) — longer is smoother but slower. Capped just below the high-pass cutoff if set higher
• RMS Normalisation Window (default 100 bars)
• Table position, state colours, pane background and optional main-chart background tint are all configurable
█ HOW TO USE IT
Use it to see the market's cyclical rhythm without the trend in the way. Swings that stretch beyond roughly ±2 on the normalised scale are large relative to recent history. Phase changes, such as Positive & Rising turning into Positive & Fading, show momentum within the cycle shifting before the zero line is crossed. It works best as context alongside your own trend and entry process, not as a standalone signal. Because it removes the trend by design, it is a tool for judging where you are within a swing, not which way the larger trend is going.
Works on any asset and timeframe; tune the two cutoffs to the cycle lengths you trade.
█ LIMITATIONS
• It isolates a band of cycle lengths. It is not a trend indicator, and a long steady trend reads as close to zero by design.
• It is a filter, not a forecast. It is a smoothed view of what price has already done, with a small, unavoidable lag from the low-pass stage.
• Passband choice matters. The cutoffs decide which cycles survive, and there is no single correct pair for every asset or timeframe.
• Recursive filters need time to settle, so the plot starts only after a warm-up period; the earliest bars on a chart are not reliable.
• RMS normalisation makes readings comparable but relative to recent history, so the same absolute move reads smaller after a volatile stretch than after a quiet one.
• Four phases are a deliberate simplification of a continuous reality.
• All classification updates on confirmed bar close only. This script does NOT repaint.
█ DISCLAIMER
For educational and informational purposes only. Nothing here is financial advice. Past behaviour of any cyclical phase does not indicate future results. Trade at your own risk.
────────────────────────────────────────────────────────────────────
█ OVERVIEW
Price is three things layered on top of each other: a slow trend drift, a band of cyclical swings, and high-frequency noise. Most oscillators blur all three together. John Ehlers' Roofing Filter separates them. It removes the slow drift, then removes the bar-to-bar noise, and leaves only the middle band: the market's cyclical component, as a clean zero-centred oscillator.
█ WHAT IT DOES
Runs price through a two-pole high-pass filter (strips the trend) followed by a two-pole Super Smoother low-pass filter (strips the noise), then divides the result by its own recent root-mean-square so the reading is in comparable units on any asset. Classifies each confirmed bar into one of four phases — Positive & Rising, Positive & Fading, Negative & Falling, Negative & Recovering — colours the background accordingly, and shows a table with the current phase, how long it has lasted, the reading, and the active passband.
█ THE THEORY BEHIND IT
A standard oscillator such as RSI or a detrended price series still carries slow drift from the underlying trend, so its zero line or midline means different things at different times. A high-pass filter removes that drift by discarding cycles longer than a chosen cutoff. That alone leaves jagged, noisy output, so a low-pass Super Smoother, a two-pole filter that smooths with less lag than a simple moving average of similar length, removes cycles shorter than a second cutoff. What survives is a "passband" of cycle lengths: the swings long enough to matter and short enough not to be trend. In Ehlers' terminology, the filter's output "roofs" over both ends of the spectrum.
█ HOW IT IS CALCULATED
1. Two-pole high-pass filter on price, with a cutoff period of 48 bars by default (cycles longer than this are removed as drift).
2. Two-pole Super Smoother low-pass filter on the high-pass output, with a cutoff of 10 bars by default (cycles shorter than this are removed as noise).
3. Divide the filtered value by its own root-mean-square over a rolling window (100 bars by default) so the output is in comparable units.
4. Classify: above zero and higher than the previous confirmed bar = Positive & Rising; above zero and lower = Positive & Fading; below zero and lower = Negative & Falling; below zero and higher = Negative & Recovering.
State classification occurs ONLY on confirmed bar close. The plot, background and table update together on closed bars, so nothing flips mid-bar.
█ SETTINGS & CONFIGURATION
• Source (default close)
• High-Pass Cutoff (default 48 bars) — shorter removes trend more aggressively
• Low-Pass Cutoff (default 10 bars) — longer is smoother but slower. Capped just below the high-pass cutoff if set higher
• RMS Normalisation Window (default 100 bars)
• Table position, state colours, pane background and optional main-chart background tint are all configurable
█ HOW TO USE IT
Use it to see the market's cyclical rhythm without the trend in the way. Swings that stretch beyond roughly ±2 on the normalised scale are large relative to recent history. Phase changes, such as Positive & Rising turning into Positive & Fading, show momentum within the cycle shifting before the zero line is crossed. It works best as context alongside your own trend and entry process, not as a standalone signal. Because it removes the trend by design, it is a tool for judging where you are within a swing, not which way the larger trend is going.
Works on any asset and timeframe; tune the two cutoffs to the cycle lengths you trade.
█ LIMITATIONS
• It isolates a band of cycle lengths. It is not a trend indicator, and a long steady trend reads as close to zero by design.
• It is a filter, not a forecast. It is a smoothed view of what price has already done, with a small, unavoidable lag from the low-pass stage.
• Passband choice matters. The cutoffs decide which cycles survive, and there is no single correct pair for every asset or timeframe.
• Recursive filters need time to settle, so the plot starts only after a warm-up period; the earliest bars on a chart are not reliable.
• RMS normalisation makes readings comparable but relative to recent history, so the same absolute move reads smaller after a volatile stretch than after a quiet one.
• Four phases are a deliberate simplification of a continuous reality.
• All classification updates on confirmed bar close only. This script does NOT repaint.
█ DISCLAIMER
For educational and informational purposes only. Nothing here is financial advice. Past behaviour of any cyclical phase does not indicate future results. Trade at your own risk.
Информация о релизе
Updated screenshot
Отказ от ответственности
Информация и публикации не предназначены для предоставления и не являются финансовыми, инвестиционными, торговыми или другими видами советов или рекомендаций, предоставленных или одобренных TradingView. Подробнее читайте в Условиях использования.
Отказ от ответственности
Информация и публикации не предназначены для предоставления и не являются финансовыми, инвестиционными, торговыми или другими видами советов или рекомендаций, предоставленных или одобренных TradingView. Подробнее читайте в Условиях использования.