OPEN-SOURCE SCRIPT
Güncellendi Price Projection [FEELS]
What happened the last time price looked like this? Price Projection finds every similar window in the chart's history and shows what came after. Descriptive, not predictive.
The script takes the most recent window of bars as a search pattern, scans the chart's own history for past windows that moved in a similar shape, and aggregates what happened after those past windows into a forward picture: ghost candles for the typical path, a band for the spread of outcomes, and a dashboard of descriptive statistics. Every number on the chart describes past behavior on the current symbol. Nothing is a prediction.
How it works
The engine runs entirely on the chart's open, high, low and close. There are no external data feeds, no security() calls, no off-chart inputs.
Similarity is measured as a Pearson correlation between the current window and each historical window, computed on log prices so that a move is compared by shape rather than by absolute size. A coarse pass over sampled points ranks every candidate window; the best matches are then confirmed at full resolution. A spacing filter prevents a single historical episode from supplying most of the sample, and similarity-weighting gives closer matches more influence on the aggregate path.
Each accepted analog is anchored to the close at the end of its window. The forward bars of every analog are expressed as ratios to that anchor, then re-based onto the current price. The ghost candles are the weighted aggregate of those analog continuations. The band contains the middle portion of analog outcomes at each step (interquartile by default, adjustable). The dashboard reports the median outcome, the share of analogs that closed higher, and the best and worst analog result.
The walk-forward check
A projection tool is only as honest as its track record, so the script builds one into the dashboard. It re-runs the same analog engine at evenly spaced historical checkpoints. At each checkpoint it uses only the data that existed at that point in time, with no lookahead, and compares the analog statistics against what price actually did afterward.
Two results are reported. Direction hits is the share of checkpoints where the realized move matched the direction of the analog median. Close inside band is the share where the realized close landed inside the outcome band; a well-calibrated band sits near its coverage setting, which is shown as the target next to it.
These numbers are not always flattering, and that is the point. On some symbols and timeframes the analogs carry information and the figures are strong; on others they sit near a coin flip, and the script shows that plainly instead of hiding it. Read the walk-forward row before trusting the projection on any given chart.
The screenshot above shows the check returning weak figures on a chart where the recent structure has not been informative. That is the tool working correctly, not failing: it is telling you the analogs are noise here.
How to read it
-Ghost candles and the average path show the typical analog continuation, re-based onto current price.
-The band shows the spread of outcomes, not a guarantee. Wider band means analogs disagreed more.
-The dashboard separates the median outcome (the typical analog) from best and worst (the extremes). When the median and the average path point different ways, the sample is skewed and the band is doing the real work.
-The walk-forward row tells you how much the analogs have been worth on this exact symbol and timeframe.
Inputs
Pattern window sets how many recent bars form the search pattern. Scan depth sets how far back the search runs. Analogs used, Min similarity and Min analogs control how many matches are accepted and how strict the match must be; if too few pass, the sample auto-relaxes toward the next-best matches down to a floor, and the dashboard flags when it has done so. Projection horizon sets how far ahead the forward picture is drawn. Band coverage sets the width of the outcome band and the calibration target for the walk-forward band check. The walk-forward section controls the checkpoint count and per-checkpoint scan depth. Visual toggles cover aggregate candles, band, analog paths, average path, price targets, the live-bar divider, the current-pattern box, and the analog-window highlights and labels.
A larger pattern window means fewer but broader matches: on a weekly chart, widening it toward 100 bars finds a handful of full-cycle analogs instead of many short ones. Tune it to the structure your strategy trades.
Repainting
This is a forward projection and it recalculates. On every new bar the engine re-scans and redraws, so the projected path, band and statistics update as the pattern changes. The projection is always anchored to the most recent confirmed bar, not to the still-forming bar. Treat the forward drawing as a description of historical analogs as of the current bar, which will move as new bars arrive. The walk-forward statistics are computed on confirmed historical bars and do not repaint.
Notes
The script is a descriptive analysis tool. It does not issue buy or sell signals and the statistics are not financial advice. Historical analog behavior does not guarantee future results.
Mashup statement: this is an original implementation. The correlation-based analog search, the log-price similarity space, the anchored ratio aggregation, the outcome band, and the built-in walk-forward check are written from scratch and are not derived from another author's script.
The script takes the most recent window of bars as a search pattern, scans the chart's own history for past windows that moved in a similar shape, and aggregates what happened after those past windows into a forward picture: ghost candles for the typical path, a band for the spread of outcomes, and a dashboard of descriptive statistics. Every number on the chart describes past behavior on the current symbol. Nothing is a prediction.
How it works
The engine runs entirely on the chart's open, high, low and close. There are no external data feeds, no security() calls, no off-chart inputs.
Similarity is measured as a Pearson correlation between the current window and each historical window, computed on log prices so that a move is compared by shape rather than by absolute size. A coarse pass over sampled points ranks every candidate window; the best matches are then confirmed at full resolution. A spacing filter prevents a single historical episode from supplying most of the sample, and similarity-weighting gives closer matches more influence on the aggregate path.
Each accepted analog is anchored to the close at the end of its window. The forward bars of every analog are expressed as ratios to that anchor, then re-based onto the current price. The ghost candles are the weighted aggregate of those analog continuations. The band contains the middle portion of analog outcomes at each step (interquartile by default, adjustable). The dashboard reports the median outcome, the share of analogs that closed higher, and the best and worst analog result.
The walk-forward check
A projection tool is only as honest as its track record, so the script builds one into the dashboard. It re-runs the same analog engine at evenly spaced historical checkpoints. At each checkpoint it uses only the data that existed at that point in time, with no lookahead, and compares the analog statistics against what price actually did afterward.
Two results are reported. Direction hits is the share of checkpoints where the realized move matched the direction of the analog median. Close inside band is the share where the realized close landed inside the outcome band; a well-calibrated band sits near its coverage setting, which is shown as the target next to it.
These numbers are not always flattering, and that is the point. On some symbols and timeframes the analogs carry information and the figures are strong; on others they sit near a coin flip, and the script shows that plainly instead of hiding it. Read the walk-forward row before trusting the projection on any given chart.
The screenshot above shows the check returning weak figures on a chart where the recent structure has not been informative. That is the tool working correctly, not failing: it is telling you the analogs are noise here.
How to read it
-Ghost candles and the average path show the typical analog continuation, re-based onto current price.
-The band shows the spread of outcomes, not a guarantee. Wider band means analogs disagreed more.
-The dashboard separates the median outcome (the typical analog) from best and worst (the extremes). When the median and the average path point different ways, the sample is skewed and the band is doing the real work.
-The walk-forward row tells you how much the analogs have been worth on this exact symbol and timeframe.
Inputs
Pattern window sets how many recent bars form the search pattern. Scan depth sets how far back the search runs. Analogs used, Min similarity and Min analogs control how many matches are accepted and how strict the match must be; if too few pass, the sample auto-relaxes toward the next-best matches down to a floor, and the dashboard flags when it has done so. Projection horizon sets how far ahead the forward picture is drawn. Band coverage sets the width of the outcome band and the calibration target for the walk-forward band check. The walk-forward section controls the checkpoint count and per-checkpoint scan depth. Visual toggles cover aggregate candles, band, analog paths, average path, price targets, the live-bar divider, the current-pattern box, and the analog-window highlights and labels.
A larger pattern window means fewer but broader matches: on a weekly chart, widening it toward 100 bars finds a handful of full-cycle analogs instead of many short ones. Tune it to the structure your strategy trades.
Repainting
This is a forward projection and it recalculates. On every new bar the engine re-scans and redraws, so the projected path, band and statistics update as the pattern changes. The projection is always anchored to the most recent confirmed bar, not to the still-forming bar. Treat the forward drawing as a description of historical analogs as of the current bar, which will move as new bars arrive. The walk-forward statistics are computed on confirmed historical bars and do not repaint.
Notes
The script is a descriptive analysis tool. It does not issue buy or sell signals and the statistics are not financial advice. Historical analog behavior does not guarantee future results.
Mashup statement: this is an original implementation. The correlation-based analog search, the log-price similarity space, the anchored ratio aggregation, the outcome band, and the built-in walk-forward check are written from scratch and are not derived from another author's script.
Sürüm Notları
v1.1- New alerts: bullish and bearish analog consensus (configurable threshold), and median outcome turning positive or negative. Available in the standard alert dialog.
- New "Text size" setting scales all on-chart captions: the info label, price targets and pattern labels.
Existing setups are not affected, the analog engine and all statistics work exactly as they did.
Açık kaynak kodlu komut dosyası
Gerçek TradingView ruhuyla, bu komut dosyasının mimarı, yatırımcıların işlevselliğini inceleyip doğrulayabilmesi için onu açık kaynaklı hale getirdi. Yazarı tebrik ederiz! Ücretsiz olarak kullanabilseniz de, kodu yeniden yayınlamanın Topluluk Kurallarımıza tabi olduğunu unutmayın.
💰Free Channel — t.me/FeelsStrategy
🤵For more info contact — t.me/VolodymyrFeels
🔗 Links — feelsstrategy.com/links
🤵For more info contact — t.me/VolodymyrFeels
🔗 Links — feelsstrategy.com/links
Feragatname
Bilgiler ve yayınlar, TradingView tarafından sağlanan veya onaylanan finansal, yatırım, alım satım veya diğer türden tavsiye veya öneriler anlamına gelmez ve teşkil etmez. Kullanım Koşulları bölümünde daha fazlasını okuyun.
Açık kaynak kodlu komut dosyası
Gerçek TradingView ruhuyla, bu komut dosyasının mimarı, yatırımcıların işlevselliğini inceleyip doğrulayabilmesi için onu açık kaynaklı hale getirdi. Yazarı tebrik ederiz! Ücretsiz olarak kullanabilseniz de, kodu yeniden yayınlamanın Topluluk Kurallarımıza tabi olduğunu unutmayın.
💰Free Channel — t.me/FeelsStrategy
🤵For more info contact — t.me/VolodymyrFeels
🔗 Links — feelsstrategy.com/links
🤵For more info contact — t.me/VolodymyrFeels
🔗 Links — feelsstrategy.com/links
Feragatname
Bilgiler ve yayınlar, TradingView tarafından sağlanan veya onaylanan finansal, yatırım, alım satım veya diğer türden tavsiye veya öneriler anlamına gelmez ve teşkil etmez. Kullanım Koşulları bölümünde daha fazlasını okuyun.