OPEN-SOURCE SCRIPT
Zaktualizowano Mann-Kendall Trend Significance [RC Tools]
RC Tools — Mann-Kendall Trend Significance
────────────────────────────────────────────────────────────────────
█ OVERVIEW
Most trend tools answer "which way is price going." This one answers a different question: "how statistically unlikely is it that this trend is just noise." It applies the Mann-Kendall test — a nonparametric hypothesis test from statistics, most commonly used in hydrology and climate-science time-series analysis — to price, turning "trend" into a proper standardised test statistic rather than a slope or a moving-average read.
█ WHAT IT DOES
Computes a standardised Z-statistic for monotonic trend over a rolling window and classifies each confirmed bar as a Significant Uptrend or Significant Downtrend once that statistic crosses a configurable significance threshold. Colours the chart background accordingly, plots both the smoothed and raw Z line in a dedicated pane against static threshold lines, and shows a table with the current state, how long price has been in it, and historical base rates (average forward return and win rate) for each state.
█ THE THEORY BEHIND IT
The Mann-Kendall test was developed to detect a monotonic trend in a time series without assuming any particular distribution or that the trend is linear — it only asks whether values tend to rise (or fall) more often than chance would predict. It does this by comparing every pair of points in a window and tallying how often the later point is higher versus lower than the earlier one. Under the null hypothesis of no trend, that tally has a known variance, which lets the raw count be converted into a Z-score — the same logic behind any standard statistical significance test. A Z-score of 1.645, for example, corresponds to the classic 90% one-tailed critical value: at that level, the observed pattern would be expected by pure chance less than 10% of the time.
This is a meaningfully different question from "has price been going up." A choppy market can have more up-days than down-days without ever producing a statistically significant Z-score; a genuinely persistent trend will.
█ HOW IT IS CALCULATED
1. Over a rolling window, compute S — the sum, across every pair of points in the window, of the sign of (later value − earlier value). A persistent uptrend pushes S strongly positive; a persistent downtrend pushes it strongly negative; a directionless window keeps it near zero.
2. Under the null hypothesis of no trend, S has a known variance: Var(S) = n(n−1)(2n+5) / 18, where n is the window length (this assumes no tied values, a reasonable approximation for continuous price data).
3. Standardise S into a Z-score, with the standard continuity correction applied.
4. Optionally smooth the Z-statistic (it is naturally "steppy," since individual pairs enter and exit the window discretely as new bars form).
5. When smoothed Z rises above the long threshold, the state flips to Significant Uptrend. When it falls below the (negative) short threshold, it flips to Significant Downtrend. Otherwise the state holds — this is hysteresis, not noise.
Classification occurs ONLY on confirmed bar close — the plotted Z, the background colour and the table all update together, so nothing here can disagree mid-bar or flip back and forth as the current bar forms.
█ SETTINGS & CONFIGURATION
• Source (default close)
• Window Length (default 20, capped at 50 to keep the pairwise comparison fast)
• Long / Short Significance Thresholds (default 1.645 each, the classic 90% one-tailed critical value) — set independently so long and short conviction can be tuned separately rather than assuming symmetric behaviour
• Smoothing Length and Type (default 3-period EMA) — reduces the raw statistic's step-like behaviour
• Forward Return Window (default 20 bars) — the horizon used for the base-rate table
• Table visibility, position and colours are fully configurable; the main-chart background painting can be toggled off if you only want the statistics pane
█ HOW TO USE IT
Use it as a trend-confirmation filter, not a standalone entry trigger. Because it requires the statistic to clear a significance threshold rather than simply cross zero, it tends to flag fewer, more deliberate trend changes than a typical oscillator — useful for filtering out other tools' false starts in choppy conditions. Check the base-rate table's sample count before treating any single state as meaningfully predictive.
Works on any asset and timeframe with sufficient history for the Window Length.
█ LIMITATIONS
• Mann-Kendall tests for a MONOTONIC trend within the window. It says nothing about the trend's magnitude, and any use of it as a precision entry/exit signal is a misuse.
• The variance formula assumes no tied values, which is reasonable for continuous price data but can be mildly optimistic on assets with heavy price discretisation (e.g. very low-priced or thinly-traded instruments).
• The window length is capped at 50 to keep the pairwise comparison fast — larger structural trends spanning more bars are not captured directly.
• The raw Z statistic is discrete and "steppy" by construction; smoothing trades responsiveness for a cleaner state transition.
• Historical base-rate stats need a meaningful sample count (check N) before being trusted, especially in a low-frequency-flip regime or on a short history.
• This script does NOT repaint. All classification updates on confirmed bar close only.
█ DISCLAIMER
For educational and informational purposes only. Nothing here is financial advice. Past behaviour of any trend-significance state does not indicate future results. Trade at your own risk.
────────────────────────────────────────────────────────────────────
█ OVERVIEW
Most trend tools answer "which way is price going." This one answers a different question: "how statistically unlikely is it that this trend is just noise." It applies the Mann-Kendall test — a nonparametric hypothesis test from statistics, most commonly used in hydrology and climate-science time-series analysis — to price, turning "trend" into a proper standardised test statistic rather than a slope or a moving-average read.
█ WHAT IT DOES
Computes a standardised Z-statistic for monotonic trend over a rolling window and classifies each confirmed bar as a Significant Uptrend or Significant Downtrend once that statistic crosses a configurable significance threshold. Colours the chart background accordingly, plots both the smoothed and raw Z line in a dedicated pane against static threshold lines, and shows a table with the current state, how long price has been in it, and historical base rates (average forward return and win rate) for each state.
█ THE THEORY BEHIND IT
The Mann-Kendall test was developed to detect a monotonic trend in a time series without assuming any particular distribution or that the trend is linear — it only asks whether values tend to rise (or fall) more often than chance would predict. It does this by comparing every pair of points in a window and tallying how often the later point is higher versus lower than the earlier one. Under the null hypothesis of no trend, that tally has a known variance, which lets the raw count be converted into a Z-score — the same logic behind any standard statistical significance test. A Z-score of 1.645, for example, corresponds to the classic 90% one-tailed critical value: at that level, the observed pattern would be expected by pure chance less than 10% of the time.
This is a meaningfully different question from "has price been going up." A choppy market can have more up-days than down-days without ever producing a statistically significant Z-score; a genuinely persistent trend will.
█ HOW IT IS CALCULATED
1. Over a rolling window, compute S — the sum, across every pair of points in the window, of the sign of (later value − earlier value). A persistent uptrend pushes S strongly positive; a persistent downtrend pushes it strongly negative; a directionless window keeps it near zero.
2. Under the null hypothesis of no trend, S has a known variance: Var(S) = n(n−1)(2n+5) / 18, where n is the window length (this assumes no tied values, a reasonable approximation for continuous price data).
3. Standardise S into a Z-score, with the standard continuity correction applied.
4. Optionally smooth the Z-statistic (it is naturally "steppy," since individual pairs enter and exit the window discretely as new bars form).
5. When smoothed Z rises above the long threshold, the state flips to Significant Uptrend. When it falls below the (negative) short threshold, it flips to Significant Downtrend. Otherwise the state holds — this is hysteresis, not noise.
Classification occurs ONLY on confirmed bar close — the plotted Z, the background colour and the table all update together, so nothing here can disagree mid-bar or flip back and forth as the current bar forms.
█ SETTINGS & CONFIGURATION
• Source (default close)
• Window Length (default 20, capped at 50 to keep the pairwise comparison fast)
• Long / Short Significance Thresholds (default 1.645 each, the classic 90% one-tailed critical value) — set independently so long and short conviction can be tuned separately rather than assuming symmetric behaviour
• Smoothing Length and Type (default 3-period EMA) — reduces the raw statistic's step-like behaviour
• Forward Return Window (default 20 bars) — the horizon used for the base-rate table
• Table visibility, position and colours are fully configurable; the main-chart background painting can be toggled off if you only want the statistics pane
█ HOW TO USE IT
Use it as a trend-confirmation filter, not a standalone entry trigger. Because it requires the statistic to clear a significance threshold rather than simply cross zero, it tends to flag fewer, more deliberate trend changes than a typical oscillator — useful for filtering out other tools' false starts in choppy conditions. Check the base-rate table's sample count before treating any single state as meaningfully predictive.
Works on any asset and timeframe with sufficient history for the Window Length.
█ LIMITATIONS
• Mann-Kendall tests for a MONOTONIC trend within the window. It says nothing about the trend's magnitude, and any use of it as a precision entry/exit signal is a misuse.
• The variance formula assumes no tied values, which is reasonable for continuous price data but can be mildly optimistic on assets with heavy price discretisation (e.g. very low-priced or thinly-traded instruments).
• The window length is capped at 50 to keep the pairwise comparison fast — larger structural trends spanning more bars are not captured directly.
• The raw Z statistic is discrete and "steppy" by construction; smoothing trades responsiveness for a cleaner state transition.
• Historical base-rate stats need a meaningful sample count (check N) before being trusted, especially in a low-frequency-flip regime or on a short history.
• This script does NOT repaint. All classification updates on confirmed bar close only.
█ DISCLAIMER
For educational and informational purposes only. Nothing here is financial advice. Past behaviour of any trend-significance state does not indicate future results. Trade at your own risk.
Informacje o Wersji
Removed Performance TableSkrypt open-source
W zgodzie z duchem TradingView twórca tego skryptu udostępnił go jako open-source, aby użytkownicy mogli przejrzeć i zweryfikować jego działanie. Ukłony dla autora. Korzystanie jest bezpłatne, jednak ponowna publikacja kodu podlega naszym Zasadom serwisu.
Wyłączenie odpowiedzialności
Informacje i publikacje nie stanowią i nie powinny być traktowane jako porady finansowe, inwestycyjne, tradingowe ani jakiekolwiek inne rekomendacje dostarczane lub zatwierdzone przez TradingView. Więcej informacji znajduje się w Warunkach użytkowania.
Skrypt open-source
W zgodzie z duchem TradingView twórca tego skryptu udostępnił go jako open-source, aby użytkownicy mogli przejrzeć i zweryfikować jego działanie. Ukłony dla autora. Korzystanie jest bezpłatne, jednak ponowna publikacja kodu podlega naszym Zasadom serwisu.
Wyłączenie odpowiedzialności
Informacje i publikacje nie stanowią i nie powinny być traktowane jako porady finansowe, inwestycyjne, tradingowe ani jakiekolwiek inne rekomendacje dostarczane lub zatwierdzone przez TradingView. Więcej informacji znajduje się w Warunkach użytkowania.