PolyFilter [BackQuant]PolyFilter
A flexible, low-lag trend filter with three smoothing engines—optimized for clean bias, fewer whipsaws, and clear alerting.
What it does
PolyFilter draws a single “intelligent” baseline that adapts to price while suppressing noise. You choose the engine— Fractional MA , Ehlers 2-Pole Super Smoother , or a Multi-Kernel blend . The line can color itself by slope (trend) or by position vs price (above/below), and you get four ready-made alerts for flips and crosses.
What it plots
PolyFilter line — your smoothed trend baseline (width set by “Line Width”).
Optional candle & background coloring — choose: color by trend slope or by whether price is above/below the filter.
Signal markers — Arrows with L/S when the slope flips or when price crosses the line (if you enable shapes/alerts).
How the three engines differ
Fractional MA (experimental) — A power-law weighting of past bars (heavier focus on the most recent samples without throwing away history). The Adaptation Speed acts like the “fraction” exponent (default 0.618). Lower values lean more on recent bars; higher values spread weight further back.
Ehlers 2-Pole Super Smoother — Classic low-lag IIR smoother that aggressively reduces high-frequency noise while preserving turns. Great default when you want a steady, responsive baseline with minimal parameter fuss.
Multi-Kernel — A 70/30 blend of a Gaussian window and an exponential kernel. The Gaussian contributes smooth structure; the exponential adds a hint of responsiveness. Useful for assets that oscillate but still trend.
Reading the colors
Trend mode (default) — Line & candles turn green while the filter is rising (signal > signal ) and red while it’s falling.
Above/Below mode — Line & candles reflect price’s position relative to the filter: green when price > filter, red when price < filter. This is handy if you treat the filter like a dynamic “fair value” or bias line.
Inputs you’ll actually use
Calculation Settings
Price Source — Default HLC/3. Switch to Close for stricter trend, or HLC3/HL2 to soften single-print spikes.
Filter Length — Window/period for all engines. Shorter = snappier turns; longer = smoother line.
Adaptation Speed — Only affects Fractional MA . Lower it for faster, more local weighting; raise it for smoother, more global weighting.
Filter Type — Pick one of: Fractional MA, Ehlers 2-Pole, Multi-Kernel.
UI & Plotting
Color based off… — Choose Trend (slope) or > or < Close (position vs price).
Long/Short Colors — Customize bull/bear hues to your theme.
Show Filter Line / Paint candles / Color background — Visual toggles for the line, bars, and backdrop.
Line Width — Make the filter stand out (2–3 works well on most charts).
Signals & Alerts
PolyFilter Trend Up — Slope flips upward (the filter crosses above its prior value). Good for early continuation entries or stop-tightening on shorts.
PolyFilter Trend Down — Slope flips downward. Often used to scale out longs or rotate bias.
PolyFilter Above Price — The filter line crosses up through price (filter > price). This can confirm that mean has “caught up” after a pullback.
PolyFilter Below Price — The filter line crosses down through price (filter < price). Useful to confirm momentum loss on bounces.
Quick starts (suggested presets)
Intraday (5–15m, crypto or indices) — Ehlers 2-Pole, Length 55–80. Trend coloring ON, candle paint ON. Look for pullbacks to a rising filter; avoid fading a falling one.
Swing (1H–4H) — Multi-Kernel, Length 80–120. Background color OFF (cleaner), candle paint ON. Add a higher-TF confirmation (e.g., 4H filter rising when you trade 1H).
Range-prone FX — Fractional MA, Length 70–100, Adaptation ~0.55–0.70. Consider Above/Below mode to trade mean reversion to the line with a strict risk cap.
How to use it in practice
Bias line — Trade in the direction of the filter slope; stand aside when it flattens and color chops back and forth.
Dynamic support/resistance — Treat the line as a moving value area. In trends, entries often appear on shallow tags of the line with structure confluence.
Regime switch — When the filter flips and holds color for several bars, tighten stops on the opposing side and look for first pullback in the new color.
Stacking filters — Many users run PolyFilter on the active chart and a slower instance (longer length) on a higher timeframe as a “macro bias” guardrail.
Tuning tips
If you see too many flips, lengthen the filter or switch to Multi-Kernel.
If turns feel late, shorten the filter or try Ehlers 2-Pole for lower lag.
On thin or very noisy symbols, prefer HLC3 as the source and longer lengths.
Performance note: very large lengths increase computation time for the Multi-Kernel and Fractional engines. Start moderate and scale up only if needed.
Summary
PolyFilter gives you a single, trustworthy baseline that you can read at a glance—either as a pure trend line (slope coloring) or as a dynamic “above/below fair value” reference. Pick the engine that matches your market’s personality, set a sensible length, and let the color and alerts guide bias, entries on pullbacks, and risk on reversals.
Backquant
Kalman Adjusted Average True Range [BackQuant]Kalman Adjusted Average True Range
A volatility-aware trend baseline that fuses a Kalman price estimate with ATR “rails” to create a smooth, adaptive guide for entries, exits, and trailing risk.
Built on my original Kalman
This indicator is based on my original Kalman Price Filter:
That core smoother is used here to estimate the “true” price path, then blended with ATR to control step size and react proportionally to market noise.
What it plots
Kalman ATR Line the main baseline that turns up/down with the filtered trend.
Optional Moving Average of the Kalman ATR a secondary line for confluence (SMA/Hull/EMA/WMA/DEMA/RMA/LINREG/ALMA).
Candle Coloring (optional) paint bars by the baseline’s current direction.
Why combine Kalman + ATR?
Kalman reduces measurement noise and produces a stable path without the lag of heavy MAs.
ATR rails scale the baseline’s step to current volatility, so it’s calm in chop and more responsive in expansion.
The result is a single, intelligible line you can trade around: slope-up = constructive; slope-down = caution.
How it works (plain English)
Each bar, the Kalman filter updates an internal state (tunable via Process Noise , Measurement Noise , and Filter Order ) to estimate the underlying price.
An ATR band (Period × Factor) defines the allowed per-bar adjustment. The baseline cannot “jump” beyond those rails in one step.
A direction flip is detected when the baseline’s slope changes sign (upturn/downturn), and alerts are provided for both.
Typical uses
Trend confirmation Trade in the baseline’s direction; avoid fading a firmly rising/falling line.
Pullback timing Look for entries when price mean-reverts toward a rising baseline (or exits on tags of a falling one).
Trailing risk Use the baseline as a dynamic guide; many traders set stops a small buffer beyond it (e.g., a fraction of ATR).
Confluence Enable the MA overlay of the Kalman ATR; alignment (baseline above its MA and rising) supports continuation.
Inputs & what they do
Calculation
Kalman Price Source which price the filter tracks (Close by default).
Process Noise how quickly the filter can adapt. Higher = more responsive (but choppier).
Measurement Noise how much you distrust raw price. Higher = smoother (but slower to turn).
Filter Order (N) depth of the internal state array. Higher = slightly steadier behavior.
Kalman ATR
Period ATR lookback. Shorter = snappier; longer = steadier.
Factor scales the allowed step per bar. Larger factors permit faster drift; smaller factors clamp movement.
Confluence (optional)
MA Type & Period compute an MA on the Kalman ATR line , not on price.
Sigma (ALMA) if ALMA is selected, this input controls the curve’s shape. (Ignored for other MA types.)
Visuals
Plot Kalman ATR toggle the main line.
Paint Candles color bars by up/down slope.
Colors choose long/short hues.
Signals & alerts
Trend Up baseline turns upward (slope crosses above 0).
Alert: “Kalman ATR Trend Up”
Trend Down baseline turns downward (slope crosses below 0).
Alert: “Kalman ATR Trend Down”
These are state flips , not “price crossovers,” so you avoid many one-bar head-fakes.
How to start (fast presets)
Swing (daily/4H) ATR Period 7–14, Factor 0.5–0.8, Process Noise 0.02–0.05, Measurement Noise 2–4, N = 3–5.
Intraday (5–15m) ATR Period 5–7, Factor 0.6–1.0, Process Noise 0.05–0.10, Measurement Noise 2–3, N = 3–5.
Slow assets / FX raise Measurement Noise or ATR Period for calmer lines; drop Factor if the baseline feels too jumpy.
Reading the line
Rising & curving upward momentum building; consider long bias until a clear downturn.
Flat & choppy regime uncertainty; many traders stand aside or tighten risk.
Falling & accelerating distribution lower; short bias until a clean upturn.
Practical playbook
Continuation entries After a Trend Up alert, wait for a minor pullback toward the baseline; enter on evidence the line keeps rising.
Exit/reduce If long and the baseline flattens then turns down, trim or exit; reverse logic for shorts.
Filters Add a higher-timeframe check (e.g., only take longs when the daily Kalman ATR is rising).
Stops Place stops just beyond the baseline (e.g., baseline − x% ATR for longs) to avoid “tag & reverse” noise.
Notes
This is a guide to state and momentum, not a guarantee. Combine with your process (structure, volume, time-of-day) for decisions.
Settings are asset/timeframe dependent; start with the presets and nudge Process/Measurement Noise until the baseline “feels right” for your market.
Summary
Kalman ATR takes the noise-reduction of a Kalman price estimate and couples it with volatility-scaled movement to produce a clean, adaptive baseline. If you liked the original Kalman Price Filter (), this is its trend-trading cousin purpose-built for cleaner state flips, intuitive trailing, and confluence with your existing
Key Levels: Daily, Weekly, Monthly [BackQuant]Key Levels: Daily, Weekly, Monthly
Map the market’s “memory” in one glance—yesterday’s range, this week’s chosen day high/low, and D/W/M opens—then auto-clean levels once they break.
What it does
This tool plots three families of high-signal reference lines and keeps them tidy as price evolves:
Chosen Day High/Low (per week) — Pick a weekday (e.g., Monday). For each past week, the script records that day’s session high and low and projects them forward for a configurable number of bars. These act like “memory levels” that price often revisits.
Daily / Weekly / Monthly Opens — Plots the opening price of each new day, week, and month with separate styling. These opens frequently behave like magnets/flip lines intraday and anchors for regime on higher timeframes.
Auto-pruning — When price breaks a stored level, the script can automatically remove it to reduce clutter and refocus you on still-active lines. See: (broken levels removed).
Why these levels matter
Liquidity pockets — Prior day’s high/low and the daily open concentrate stops and pending orders. Mapping them quickly reveals likely sweep or fade zones. Example: previous day highs + daily open highlighting liquidity:
Context & regime — Monthly opens frame macro bias; trading above a rising cluster of monthly opens vs. below gives a clean top-down read. Example: monthly-only “macro outlook” view:
Cleaner charts — Auto-remove broken lines so you focus on what still matters right now.
What it plots (at a glance)
Past Chosen Day High/Low for up to N prior weeks (your choice), extended right.
Current Daily Open , Weekly Open , and Monthly Open , each with its own color, label, and forward extension.
Optional short labels (e.g., “Mon High”) or full labels (with week/month info).
How breaks are detected & cleaned
You control both the evidence and the timing of a “break”:
Break uses — Choose Close (more conservative) or Wick (more sensitive).
Inclusive? — If enabled, equality counts (≥ high or ≤ low). If disabled, you need a strict cross.
Allow intraday breaks? — If on, a level can break during the tracked day; if off, the script only counts breaks after the session completes.
Remove Broken Levels — When a break is confirmed, the line/label is deleted automatically. (See the demo: )
Quick start
Pick a Day of Week to Track (e.g., Monday).
Set how many weeks back to show (e.g., 8–10).
Choose how far to extend each family (bars to the right for chosen-day H/L and D/W/M opens).
Decide if a break uses Close or Wick , and whether equality counts.
Toggle Remove Broken Levels to keep the chart clean automatically.
Tips by use-case
Intraday bias — Watch the Daily Open as a magnet/flip. If price gaps above and holds, pullbacks to the daily open often decide direction. Pair with last day’s high/low for sweep→reversal or true breakout cues. See:
Weekly structure — Track the week’s chosen day (e.g., Monday) high/low across prior weeks. If price stalls near a cluster of old “Monday Highs,” look for sweep/reject patterns or continuation on reclaim.
Macro regime — Hide daily/weekly lines and keep only Monthly Opens to read bigger cycles at a glance (BTC/crypto especially). Example:
Customization
Use wicks or bodies for highs/lows (wicks capture extremes; bodies are stricter).
Line style & thickness — solid/dashed/dotted, width 1–5, plus global transparency.
Labels — Abbreviated (“Mon High”, “D Open”) or full (month/week/day info).
Color scheme — Separate colors for highs, lows, and each of D/W/M opens.
Capacity controls — Set how many daily/weekly/monthly opens and how many weeks of chosen-day H/L to keep visible.
What’s under the hood
On your selected weekday, the script records that session’s true high and true low (using wicks or body-based extremes—your choice), then projects a horizontal line forward for the next bars.
At each new day/week/month , it records the opening price and projects that line forward as well.
Each bar, the script checks your “break” rules; once broken, lines/labels are removed if auto-cleaning is on.
Everything updates in real time; past levels don’t repaint after the session finishes.
Recommended presets
Day trading — Weeks back: 6–10; extend D/W opens: 50–100 bars; Break uses: Close ; Inclusive: off; Auto-remove: on.
Swing — Fewer daily opens, more weekly opens (2–6), and 8–12 weeks of chosen-day H/L.
Macro — Show only Monthly Opens (1–6 months), dashed style, thicker lines for clarity.
Reading the examples
Broken lines disappear — decluttering in action:
Macro outlook — monthly opens as cycle rails:
Liquidity map — previous day highs + daily open:
Final note
These are not “signals”—they’re reference points that many participants watch. By standardising how you draw them and automatically clearing the ones that no longer matter, you turn a noisy chart into a focused map: where liquidity likely sits, where price memory lives, and which lines are still in play.
Quantile Regression Bands [BackQuant]Quantile Regression Bands
Tail-aware trend channeling built from quantiles of real errors, not just standard deviations.
What it does
This indicator fits a simple linear trend over a rolling lookback and then measures how price has actually deviated from that trend during the window. It then places two pairs of bands at user-chosen quantiles of those deviations (inner and outer). Because bands are based on empirical quantiles rather than a symmetric standard deviation, they adapt to skewed and fat-tailed behaviour and often hug price better in trending or asymmetric markets.
Why “quantile” bands instead of Bollinger-style bands?
Bollinger Bands assume a (roughly) symmetric spread around the mean; quantiles don’t—upper and lower bands can sit at different distances if the error distribution is skewed.
Quantiles are robust to outliers; a single shock won’t inflate the bands for many bars.
You can choose tails precisely (e.g., 1%/99% or 5%/95%) to match your risk appetite.
How it works (intuitive)
Center line — a rolling linear regression approximates the local trend.
Residuals — for each bar in the lookback, the indicator looks at the gap between actual price and where the line “expected” price to be.
Quantiles — those gaps are sorted; you select which percentiles become your inner/outer offsets.
Bands — the chosen quantile offsets are added to the current end of the regression line to draw parallel support/resistance rails.
Smoothing — a light EMA can be applied to reduce jitter in the line and bands.
What you see
Center (linear regression) line (optional).
Inner quantile bands (e.g., 25th/75th) with optional translucent fill.
Outer quantile bands (e.g., 1st/99th) with a multi-step gradient to visualise “tail zones.”
Optional bar coloring: bars trend-colored by whether price is rising above or falling below the center line.
Alerts when price crosses the outer bands (upper or lower).
How to read it
Trend & drift — the slope of the center line is your local trend. Persistent closes on the same side of the center line indicate directional drift.
Pullbacks — tags of the inner band often mark routine pullbacks within trend. Reaction back to the center line can be used for continuation entries/partials.
Tails & squeezes — outer-band touches highlight statistically rare excursions for the chosen window. Frequent outer-band activity can signal regime change or volatility expansion.
Asymmetry — if the upper band sits much further from the center than the lower (or vice versa), recent behaviour has been skewed. Trade management can be adjusted accordingly (e.g., wider take-profit upslope than downslope).
A simple trend interpretation can be derived from the bar colouring
Good use-cases
Volatility-aware mean reversion — fade moves into outer bands back toward the center when trend is flat.
Trend participation — buy pullbacks to the inner band above a rising center; flip logic for shorts below a falling center.
Risk framing — set dynamic stops/targets at quantile rails so position sizing respects recent tail behaviour rather than fixed ticks.
Inputs (quick guide)
Source — price input used for the fit (default: close).
Lookback Length — bars in the regression window and residual sample. Longer = smoother, slower bands; shorter = tighter, more reactive.
Inner/Outer Quantiles (τ) — choose your “typical” vs “tail” levels (e.g., 0.25/0.75 inner, 0.01/0.99 outer).
Show toggles — independently toggle center line, inner bands, outer bands, and their fills.
Colors & transparency — customize band and fill appearance; gradient shading highlights the tail zone.
Band Smoothing Length — small EMA on lines to reduce stair-step artefacts without meaningfully changing levels.
Bar Coloring — optional trend tint from the center line’s momentum.
Practical settings
Swing trading — Length 75–150; inner τ = 0.25/0.75, outer τ = 0.05/0.95.
Intraday — Length 50–100 for liquid futures/FX; consider 0.20/0.80 inner and 0.02/0.98 outer in high-vol assets.
Crypto — Because of fat tails, try slightly wider outers (0.01/0.99) and keep smoothing at 2–4 to tame weekend jumps.
Signal ideas
Continuation — in an uptrend, look for pullback into the lower inner band with a close back above the center as a timing cue.
Exhaustion probe — in ranges, first touch of an outer band followed by a rejection candle back inside the inner band often precedes mean-reversion swings.
Regime shift — repeated closes beyond an outer band or a sharp re-tilt in the center line can mark a new trend phase; adjust tactics (stop-following along the opposite inner band).
Alerts included
“Price Crosses Upper Outer Band” — potential overextension or breakout risk.
“Price Crosses Lower Outer Band” — potential capitulation or breakdown risk.
Notes
The fit and quantiles are computed on a fixed rolling window and do not repaint; bands update as the window moves forward.
Quantiles are based on the recent distribution; if conditions change abruptly, expect band widths and skew to adapt over the next few bars.
Parameter choices directly shape behaviour: longer windows favour stability, tighter inner quantiles increase touch frequency, and extreme outer quantiles highlight only the rarest moves.
Final thought
Quantile bands answer a simple question: “How unusual is this move given the current trend and the way price has been missing it lately?” By scoring that question with real, distribution-aware limits rather than one-size-fits-all volatility you get cleaner pullback zones in trends, more honest “extreme” tags in ranges, and a framework for risk that matches the market’s recent personality.
Deadband Hysteresis Supertrend [BackQuant]Deadband Hysteresis Supertrend
A two-stage trend tool that first filters price with a deadband baseline, then runs a Supertrend around that baseline with optional flip hysteresis and ATR-based adverse exits.
What this is
A hybrid of two ideas:
Deadband Hysteresis Baseline that only advances when price pulls far enough from the baseline to matter. This suppresses micro noise and gives you a stable centerline.
Supertrend bands wrapped around that baseline instead of raw price. Flips are further gated by an extra margin so side changes are more deliberate.
The goal is fewer whipsaws in chop and clearer regime identification during trends.
How it works (high level)
Deadband step — compute a per-bar “deadband” size from one of four modes: ATR, Percent of price, Ticks, or Points. If price deviates from the baseline by more than this amount, move the baseline forward by a fraction of the excess. If not, hold the line.
Centered Supertrend — build upper and lower bands around the baseline using ATR and a user factor. Track the usual trailing logic that tightens a band while price moves in its favor.
Flip hysteresis — require price to exceed the active band by an extra flip offset × ATR before switching sides. This adds stickiness at the boundary.
Adverse exit — once a side is taken, trigger an exit if price moves against the entry by K × ATR .
If you would like to check out the filter by itself:
What it plots
DBHF baseline (optional) as a smooth centerline.
DBHF Supertrend as the active trailing band.
Candle coloring by trend side for quick read.
Signal markers 𝕃 and 𝕊 at flips plus ✖ on adverse exits.
Inputs that matter
Price Source — series being filtered. Close is typical. HL2 or HLC3 can be steadier.
Deadband mode — ATR, Percent, Ticks, or Points. This defines the “it’s big enough to matter” zone.
ATR Length / Mult (DBHF) — only used when mode = ATR. Larger values widen the do-nothing zone.
Percent / Ticks / Points — alternatives to ATR; pick what fits your market’s convention.
Enter Mult — scales the deadband you must clear before the baseline moves. Increase to filter more noise.
Response — fraction of the excess applied to baseline movement. Higher responds faster; lower is smoother.
Supertrend ATR Period & Factor — traditional band size controls; higher factor widens and flips less often.
Flip Offset ATR — extra ATR buffer required to flip. Useful in choppy regimes.
Adverse Stop K·ATR — per-trade danger brake that forces an exit if price moves K×ATR against entry.
UI — toggle baseline, supertrend, signals, and bar painting; choose long and short colors.
How to read it
Green regime — candles painted long and the Supertrend running below price. Pullbacks toward the baseline that fail to breach the opposite band often resume higher.
Red regime — candles painted short and the Supertrend running above price. Rallies that cannot reclaim the band may roll over.
Frequent side swaps — reduce sensitivity by increasing Enter Mult, using ATR mode, raising the Supertrend factor, or adding Flip Offset ATR.
Use cases
Bias filter — allow entries only in the direction of the current side. Use your preferred triggers inside that bias.
Trailing logic — treat the active band as a dynamic stop. If the side flips or an adverse K·ATR exit prints, reduce or close exposure.
Regime map — on higher timeframes, the combination baseline + band produces a clean up vs down template for allocation decisions.
Tuning guidance
Fast markets — ATR deadband, modest Enter Mult (0.8–1.2), response 0.2–0.35, Supertrend factor 1.7–2.2, small Flip Offset (0.2–0.5 ATR).
Choppy ranges — widen deadband or raise Enter Mult, lower response, and add more Flip Offset so flips require stronger evidence.
Slow trends — longer ATR periods and higher Supertrend factor to keep you on side longer; use a conservative adverse K.
Included alerts
DBHF ST Long — side flips to long.
DBHF ST Short — side flips to short.
Adverse Exit Long / Short — K·ATR stop triggers against the current side.
Strengths
Deadbanded baseline reduces micro whipsaws before Supertrend logic even begins.
Flip hysteresis adds a second layer of confirmation at the boundary.
Optional adverse ATR stop provides a uniform risk cut across assets and regimes.
Clear visuals and minimal parameters to adjust for symbol behavior.
Putting it together
Think of this tool as two decisions layered into one view. The deadband baseline answers “does this move even count,” then the Supertrend wrapped around that baseline answers “if it counts, which side should I be on and where do I flip.” When both parts agree you tend to stay on the correct side of a trend for longer, and when they disagree you get an early warning that conditions are changing.
When the baseline bends and price cannot reclaim the opposite band , momentum is usually continuing. Pullbacks into the baseline that stall before the far band often resolve in trend.
When the baseline flattens and the bands compress , expect indecision. Use the Flip Offset ATR to avoid reacting to the first feint. Wait for a clean band breach with follow through.
When an adverse K·ATR exit prints while the side has not flipped , treat it as a risk event rather than a full regime change. Many users cut size, re-enter only if the side reasserts, and let the next flip confirm a new trend.
Final thoughts
Deadband Hysteresis Supertrend is best read as a regime lens. The baseline defines your tolerance for noise, the bands define your trailing structure, and the flip offset plus adverse ATR stop define how forgiving or strict you want to be at the boundary. On strong trends it helps you hold through shallow shakeouts. In choppy conditions it encourages patience until price does something meaningful. Start with settings that reflect the cadence of your market, observe how often flips occur, then nudge the deadband and flip offset until the tool spends most of its time describing the move you care about rather than the noise in between.
Theil-Sen Line Filter [BackQuant]Theil-Sen Line Filter
A robust, median-slope baseline that tracks price while resisting outliers. Designed for the chart pane as a clean, adaptive reference line with optional candle coloring and slope-flip alerts.
What this is
A trend filter that estimates the underlying slope of price using a Theil-Sen style median of past slopes, then advances a baseline by a controlled fraction of that slope each bar. The result is a smooth line that reacts to real directional change while staying calm through noise, gaps, and single-bar shocks.
Why Theil-Sen
Classical moving averages are sensitive to outliers and shape changes. Ordinary least squares is sensitive to large residuals. The Theil-Sen idea replaces a single fragile estimate with the median of many simple slopes, which is statistically robust and less influenced by a few extreme bars. That makes the baseline steadier in choppy conditions and cleaner around regime turns.
What it plots
Filtered baseline that advances by a fraction of the robust slope each bar.
Optional candle coloring by baseline slope sign for quick trend read.
Alerts when the baseline slope turns up or down.
How it behaves (high level)
Looks back over a fixed window and forms many “current vs past” bar-to-bar slopes.
Takes the median of those slopes to get a robust estimate for the bar.
Optionally caps the magnitude of that per-bar slope so a single volatile bar cannot yank the line.
Moves the baseline forward by a user-controlled fraction of the estimated slope. Lower fractions are smoother. Higher fractions are more responsive.
Inputs and what they do
Price Source — the series the filter tracks. Typical is close; HL2 or HLC3 can be smoother.
Window Length — how many bars to consider for slopes. Larger windows are steadier and slower. Smaller windows are quicker and noisier.
Response — fraction of the estimated slope applied each bar. 1.00 follows the robust slope closely; values below 1.00 dampen moves.
Slope Cap Mode — optional guardrail on each bar’s slope:
None — no cap.
ATR — cap scales with recent true range.
Percent — cap scales with price level.
Points — fixed absolute cap in price points.
ATR Length / Mult, Cap Percent, Cap Points — tune the chosen cap mode’s size.
UI Settings — show or hide the line, paint candles by slope, choose long and short colors.
How to read it
Up-slope baseline and green candles indicate a rising robust trend. Pullbacks that do not flip the slope often resolve in trend direction.
Down-slope baseline and red candles indicate a falling robust trend. Bounces against the slope are lower-probability until proven otherwise.
Flat or frequent flips suggest a range. Increase window length or decrease response if you want fewer whipsaws in sideways markets.
Use cases
Bias filter — only take longs when slope is up, shorts when slope is down. It is a simple way to gate faster setups.
Stop or trail reference — use the line as a trailing guide. If price closes beyond the line and the slope flips, consider reducing exposure.
Regime detector — widen the window on higher timeframes to define major up vs down regimes for asset rotation or risk toggles.
Noise control — enable a cap mode in very volatile symbols to retain the line’s continuity through event bars.
Tuning guidance
Quick swing trading — shorter window, higher response, optionally add a percent cap to keep it stable on large moves.
Position trading — longer window, moderate response. ATR cap tends to scale well across cycles.
Low-liquidity or gappy charts — prefer longer window and a points or ATR cap. That reduces jumpiness around discontinuities.
Alerts included
Theil-Sen Up Slope — baseline’s one-bar change crosses above zero.
Theil-Sen Down Slope — baseline’s one-bar change crosses below zero.
Strengths
Robust to outliers through median-based slope estimation.
Continuously advances with price rather than re-anchoring, which reduces lag at turns.
User-selectable slope caps to tame shock bars without over-smoothing everything.
Minimal visuals with optional candle painting for fast regime recognition.
Notes
This is a filter, not a trading system. It does not account for execution, spreads, or gaps. Pair it with entry logic, risk management, and higher-timeframe context if you plan to use it for decisions.
Deadband Hysteresis Filter [BackQuant]Deadband Hysteresis Filter
What this is
This tool builds a “debounced” price baseline that ignores small fluctuations and only reacts when price meaningfully departs from its recent path. It uses a deadband to define how much deviation matters and a hysteresis scheme to avoid rapid flip-flops around the decision boundary. The baseline’s slope provides a simple trend cue, used to color candles and to trigger up and down alerts.
Why deadband and hysteresis help
They filter micro noise so the baseline does not react to every tiny tick.
They stabilize state changes. Hysteresis means the rule to start moving is stricter than the rule to keep holding, which reduces whipsaw.
They produce a stepped, readable path that advances during sustained moves and stays flat during chop.
How it works (conceptual)
At each bar the script maintains a running baseline dbhf and compares it to the input price p .
Compute a base threshold baseTau using the selected mode (ATR, Percent, Ticks, or Points).
Build an enter band tauEnter = baseTau × Enter Mult and an exit band tauExit = baseTau × Exit Mult where typically Exit Mult < Enter Mult .
Let diff = p − dbhf .
If diff > +tauEnter , raise the baseline by response × (diff − tauEnter) .
If diff < −tauEnter , lower the baseline by response × (diff + tauEnter) .
Otherwise, hold the prior value.
Trend state is derived from slope: dbhf > dbhf → up trend, dbhf < dbhf → down trend.
Inputs and what they control
Threshold mode
ATR — baseTau = ATR(atrLen) × atrMult . Adapts to volatility. Useful when regimes change.
Percent — baseTau = |price| × pctThresh% . Scale-free across symbols of different prices.
Ticks — baseTau = syminfo.mintick × tickThresh . Good for futures where tick size matters.
Points — baseTau = ptsThresh . Fixed distance in price units.
Band multipliers and response
Enter Mult — outer band. Price must travel at least this far from the baseline before an update occurs. Larger values reject more noise but increase lag.
Exit Mult — inner band for hysteresis. Keep this smaller than Enter Mult to create a hold zone that resists small re-entries.
Response — step size when outside the enter band. Higher response tracks faster; lower response is smoother.
UI settings
Show Filtered Price — plots the baseline on price.
Paint candles — colors bars by the filtered slope using your long/short colors.
How it can be used
Trend qualifier — take entries only in the direction of the baseline slope and skip trades against it.
Debounced crossovers — use the baseline as a stabilized surrogate for price in moving-average or channel crossover rules.
Trailing logic — trail stops a small distance beyond the baseline so small pullbacks do not eject the trade.
Session aware filtering — widen Enter Mult or switch to ATR mode for volatile sessions; tighten in quiet sessions.
Parameter interactions and tuning
Enter Mult vs Response — both govern sensitivity. If you see too many flips, increase Enter Mult or reduce Response. If turns feel late, do the opposite.
Exit Mult — widening the gap between Enter and Exit expands the hold zone and reduces oscillation around the threshold.
Mode choice — ATR adapts automatically; Percent keeps behavior consistent across instruments; Ticks or Points are useful when you think in fixed increments.
Timeframe coupling — on higher timeframes you can often lower Enter Mult or raise Response because raw noise is already reduced.
Concrete starter recipes
General purpose — ATR mode, atrLen=14 , atrMult=1.0–1.5 , Enter=1.0 , Exit=0.5 , Response=0.20 . Balanced noise rejection and lag.
Choppy range filter — ATR mode, increase atrMult to 2.0, keep Response≈0.15 . Stronger suppression of micro-moves.
Fast intraday — Percent mode, pctThresh=0.1–0.3 , Enter=1.0 , Exit=0.4–0.6 , Response=0.30–0.40 . Quicker turns for scalping.
Futures ticks — Ticks mode, set tickThresh to a few spreads beyond typical noise; start with Enter=1.0 , Exit=0.5 , Response=0.25 .
Strengths
Clear, explainable logic with an explicit noise budget.
Multiple threshold modes so the same tool fits equities, futures, and crypto.
Built-in hysteresis that reduces flip-flop near the boundary.
Slope-based coloring and alerts that make state changes obvious in real time.
Limitations and notes
All filters add lag. Larger thresholds and smaller response trade faster reaction for fewer false turns.
Fixed Points or Ticks can under- or over-filter when volatility regime shifts. ATR adapts, but will also expand bands during spikes.
On extremely choppy symbols, even a well tuned band will step frequently. Widen Enter Mult or reduce Response if needed.
This is a chart study. It does not include commissions, slippage, funding, or gap risks.
Alerts
DBHF Up Slope — baseline turns from down to up on the latest bar.
DBHF Down Slope — baseline turns from up to down on the latest bar.
Implementation details worth knowing
Initialization sets the baseline to the first observed price to avoid a cold-start jump.
Slope is evaluated bar-to-bar. The up and down alerts check for a change of slope rather than raw price crossings.
Candle colors and the baseline plot share the same long/short palette with transparency applied to the line.
Practical workflow
Pick a mode that matches how you think about distance. ATR for volatility aware, Percent for scale-free, Ticks or Points for fixed increments.
Tune Enter Mult until the number of flips feels appropriate for your timeframe.
Set Exit Mult clearly below Enter Mult to create a real hold zone.
Adjust Response last to control “how fast” the baseline chases price once it decides to move.
Final thoughts
Deadband plus hysteresis gives you a principled way to “only care when it matters.” With a sensible threshold and response, the filter yields a stable, low-chop trend cue you can use directly for bias or plug into your own entries, exits, and risk rules.
Martingale Strategy Simulator [BackQuant]Martingale Strategy Simulator
Purpose
This indicator lets you study how a martingale-style position sizing rule interacts with a simple long or short trading signal. It computes an equity curve from bar-to-bar returns, adapts position size after losing streaks, caps exposure at a user limit, and summarizes risk with portfolio metrics. An optional Monte Carlo module projects possible future equity paths from your realized daily returns.
What a martingale is
A martingale sizing rule increases stake after losses and resets after a win. In its classical form from gambling, you double the bet after each loss so that a single win recovers all prior losses plus one unit of profit. In markets there is no fixed “even-money” payout and returns are multiplicative, so an exact recovery guarantee does not exist. The core idea is unchanged:
Lose one leg → increase next position size
Lose again → increase again
Win → reset to the base size
The expectation of your strategy still depends on the signal’s edge. Sizing does not create positive expectancy on its own. A martingale raises variance and tail risk by concentrating more capital as a losing streak develops.
What it plots
Equity – simulated portfolio equity including compounding
Buy & Hold – equity from holding the chart symbol for context
Optional helpers – last trade outcome, current streak length, current allocation fraction
Optional diagnostics – daily portfolio return, rolling drawdown, metrics table
Optional Monte Carlo probability cone – p5, p16, p50, p84, p95 aggregate bands
Model assumptions
Bar-close execution with no slippage or commissions
Shorting allowed and frictionless
No margin interest, borrow fees, or position limits
No intrabar moves or gaps within a bar (returns are close-to-close)
Sizing applies to equity fraction only and is capped by your setting
All results are hypothetical and for education only.
How the simulator applies it
1) Directional signal
You pick a simple directional rule that produces +1 for long or −1 for short each bar. Options include 100 HMA slope, RSI above or below 50, EMA or SMA crosses, CCI and other oscillators, ATR move, BB basis, and more. The stance is evaluated bar by bar. When the stance flips, the current trade ends and the next one starts.
2) Sizing after losses and wins
Position size is a fraction of equity:
Initial allocation – the starting fraction, for example 0.15 means 15 percent of equity
Increase after loss – multiply the next allocation by your factor after a losing leg, for example 2.00 to double
Reset after win – return to the initial allocation
Max allocation cap – hard ceiling to prevent runaway growth
At a high level the size after k consecutive losses is
alloc(k) = min( cap , base × factor^k ) .
In practice the simulator changes size only when a leg ends and its PnL is known.
3) Equity update
Let r_t = close_t / close_{t-1} − 1 be the symbol’s bar return, d_{t−1} ∈ {+1, −1} the prior bar stance, and a_{t−1} the prior bar allocation fraction. The simulator compounds:
eq_t = eq_{t−1} × (1 + a_{t−1} × d_{t−1} × r_t) .
This is bar-based and avoids intrabar lookahead. Costs, slippage, and borrowing costs are not modeled.
Why traders experiment with martingale sizing
Mean-reversion contexts – if the signal often snaps back after a string of losses, adding size near the tail of a move can pull the average entry closer to the turn
Behavioral or microstructure edges – some rules have modest edge but frequent small whipsaws; size escalation may shorten time-to-recovery when the edge manifests
Exploration and stress testing – studying the relationship between streaks, caps, and drawdowns is instructive even if you do not deploy martingale sizing live
Why martingale is dangerous
Martingale concentrates capital when the strategy is performing worst. The main risks are structural, not cosmetic:
Loss streaks are inevitable – even with a 55 percent win rate you should expect multi-loss runs. The probability of at least one k-loss streak in N trades rises quickly with N.
Size explodes geometrically – with factor 2.0 and base 10 percent, the sequence is 10, 20, 40, 80, 100 (capped) after five losses. Without a strict cap, required size becomes infeasible.
No fixed payout – in gambling, one win at even odds resets PnL. In markets, there is no guaranteed bounce nor fixed profit multiple. Trends can extend and gaps can skip levels.
Correlation of losses – losses cluster in trends and in volatility bursts. A martingale tends to be largest just when volatility is highest.
Margin and liquidity constraints – leverage limits, margin calls, position limits, and widening spreads can force liquidation before a mean reversion occurs.
Fat tails and regime shifts – assumptions of independent, Gaussian returns can understate tail risk. Structural breaks can keep the signal wrong for much longer than expected.
The simulator exposes these dynamics in the equity curve, Max Drawdown, VaR and CVaR, and via Monte Carlo sketches of forward uncertainty.
Interpreting losing streaks with numbers
A rough intuition: if your per-trade win probability is p and loss probability is q=1−p , the chance of a specific run of k consecutive losses is q^k . Over many trades, the chance that at least one k-loss run occurs grows with the number of opportunities. As a sanity check:
If p=0.55 , then q=0.45 . A 6-loss run has probability q^6 ≈ 0.008 on any six-trade window. Across hundreds of trades, a 6 to 8-loss run is not rare.
If your size factor is 1.5 and your base is 10 percent, after 8 losses the requested size is 10% × 1.5^8 ≈ 25.6% . With factor 2.0 it would try to be 10% × 2^8 = 256% but your cap will stop it. The equity curve will still wear the compounded drawdown from the sequence that led to the cap.
This is why the cap setting is central. It does not remove tail risk, but it prevents the sizing rule from demanding impossible positions
Note: The p and q math is illustrative. In live data the win rate and distribution can drift over time, so real streaks can be longer or shorter than the simple q^k intuition suggests..
Using the simulator productively
Parameter studies
Start with conservative settings. Increase one element at a time and watch how the equity, Max Drawdown, and CVaR respond.
Initial allocation – lower base reduces volatility and drawdowns across the board
Increase factor – set modestly above 1.0 if you want the effect at all; doubling is aggressive
Max cap – the most important brake; many users keep it between 20 and 50 percent
Signal selection
Keep sizing fixed and rotate signals to see how streak patterns differ. Trend-following signals tend to produce long wrong-way streaks in choppy ranges. Mean-reversion signals do the opposite. Martingale sizing interacts very differently with each.
Diagnostics to watch
Use the built-in metrics to quantify risk:
Max Drawdown – worst peak-to-trough equity loss
Sharpe and Sortino – volatility and downside-adjusted return
VaR 95 percent and CVaR – tail risk measures from the realized distribution
Alpha and Beta – relationship to your chosen benchmark
If you would like to check out the original performance metrics script with multiple assets with a better explanation on all metrics please see
Monte Carlo exploration
When enabled, the forecast draws many synthetic paths from your realized daily returns:
Choose a horizon and a number of runs
Review the bands: p5 to p95 for a wide risk envelope; p16 to p84 for a narrower range; p50 as the median path
Use the table to read the expected return over the horizon and the tail outcomes
Remember it is a sketch based on your recent distribution, not a predictor
Concrete examples
Example A: Modest martingale
Base 10 percent, factor 1.25, cap 40 percent, RSI>50 signal. You will see small escalations on 2 to 4 loss runs and frequent resets. The equity curve usually remains smooth unless the signal enters a prolonged wrong-way regime. Max DD may rise moderately versus fixed sizing.
Example B: Aggressive martingale
Base 15 percent, factor 2.0, cap 60 percent, EMA cross signal. The curve can look stellar during favorable regimes, then a single extended streak pushes allocation to the cap, and a few more losses drive deep drawdown. CVaR and Max DD jump sharply. This is a textbook case of high tail risk.
Strengths
Bar-by-bar, transparent computation of equity from stance and size
Explicit handling of wins, losses, streaks, and caps
Portable signal inputs so you can A–B test ideas quickly
Risk diagnostics and forward uncertainty visualization in one place
Example, Rolling Max Drawdown
Limitations and important notes
Martingale sizing can escalate drawdowns rapidly. The cap limits position size but not the possibility of extended adverse runs.
No commissions, slippage, margin interest, borrow costs, or liquidity limits are modeled.
Signals are evaluated on closes. Real execution and fills will differ.
Monte Carlo assumes independent draws from your recent return distribution. Markets often have serial correlation, fat tails, and regime changes.
All results are hypothetical. Use this as an educational tool, not a production risk engine.
Practical tips
Prefer gentle factors such as 1.1 to 1.3. Doubling is usually excessive outside of toy examples.
Keep a strict cap. Many users cap between 20 and 40 percent of equity per leg.
Stress test with different start dates and subperiods. Long flat or trending regimes are where martingale weaknesses appear.
Compare to an anti-martingale (increase after wins, cut after losses) to understand the other side of the trade-off.
If you deploy sizing live, add external guardrails such as a daily loss cut, volatility filters, and a global max drawdown stop.
Settings recap
Backtest start date and initial capital
Initial allocation, increase-after-loss factor, max allocation cap
Signal source selector
Trading days per year and risk-free rate
Benchmark symbol for Alpha and Beta
UI toggles for equity, buy and hold, labels, metrics, PnL, and drawdown
Monte Carlo controls for enable, runs, horizon, and result table
Final thoughts
A martingale is not a free lunch. It is a way to tilt capital allocation toward losing streaks. If the signal has a real edge and mean reversion is common, careful and capped escalation can reduce time-to-recovery. If the signal lacks edge or regimes shift, the same rule can magnify losses at the worst possible moment. This simulator makes those trade-offs visible so you can calibrate parameters, understand tail risk, and decide whether the approach belongs anywhere in your research workflow.
Adaptive Valuation [BackQuant]Adaptive Valuation
What this is
A composite, zero-centered oscillator that standardizes several classic indicators and blends them into one “valuation” line. It computes RSI, CCI, Demarker, and the Price Zone Oscillator, converts each to a rolling z-score, then forms a weighted average. Optional smoothing, dynamic overbought and oversold bands, and an on-chart table make the inputs and the final score easy to inspect.
How it works
Components
• RSI with its own lookback.
• CCI with its own lookback.
• DM (Demarker) with its own lookback.
• PZO (Price Zone Oscillator) with its own lookback.
Standardization via z-score
Each component is transformed using a rolling z-score over lookback bars:
z = (value − mean) ÷ stdev , where the mean is an EMA and the stdev is rolling.
This puts all inputs on a comparable scale measured in standard deviations.
Weighted blend
The z-scores are combined with user weights w_rsi, w_cci, w_dm, w_pzo to produce a single valuation series. If desired, it is then smoothed with a selected moving average (SMA, EMA, WMA, HMA, RMA, DEMA, TEMA, LINREG, ALMA, T3). ALMA’s sigma input shapes its curve.
Dynamic thresholds (optional)
Two ways to set overbought and oversold:
• Static : fixed levels at ob_thres and os_thres .
• Dynamic : ±k·σ bands, where σ is the rolling standard deviation of the valuation over dynLen .
Bands can be centered at zero or around the valuation’s rolling mean ( centerZero ).
Visualization and UI
• Zero line at 0 with gradient fill that darkens as the valuation moves away from 0.
• Optional plotting of band lines and background highlights when OB or OS is active.
• Optional candle and background coloring driven by the valuation.
• Summary table showing each component’s current z-score, the final score, and a compact status.
How it can be used
• Bias filter : treat crosses above 0 as bullish bias and below 0 as bearish bias.
• Mean-reversion context : look for exhaustion when the valuation enters the OB or OS region, then watch for exits from those regions or a return toward 0.
• Signal confirmation : use the final score to confirm setups from structure or price action.
• Adaptive banding : with dynamic thresholds, OB and OS adjust to prevailing variability rather than relying on fixed lines.
• Component tuning : change weights to emphasize trend (raise DM, reduce RSI/CCI) or range behavior (raise RSI/CCI, reduce DM). PZO can help in swing environments.
Why z-score blending helps
Indicators often live on different scales. Z-scoring places them on a common, unitless axis, so a one-sigma move in RSI has comparable influence to a one-sigma move in CCI. This reduces scale bias and allows transparent weighting. It also facilitates regime-aware thresholds because the dynamic bands scale with recent dispersion.
Inputs to know
• Component lookbacks : rsilb, ccilb, dmlb, pzolb control each raw signal.
• Standardization window : lookback sets the z-score memory. Longer smooths, shorter reacts.
• Weights : w_rsi, w_cci, w_dm, w_pzo determine each component’s influence.
• Smoothing : maType, smoothP, sig govern optional post-blend smoothing.
• Dynamic bands : dyn_thres, dynLen, thres_k, centerZero configure the adaptive OB/OS logic.
• UI : toggle the plot, table, candle coloring, and threshold lines.
Reading the plot
• Above 0 : composite pressure is positive.
• Below 0 : composite pressure is negative.
• OB region : valuation above the chosen OB line. Risk of mean reversion rises and momentum continuation needs evidence.
• OS region : mirror logic on the downside.
• Band exits : leaving OB or OS can serve as a normalization cue.
Strengths
• Normalizes heterogeneous signals into one interpretable series.
• Adjustable component weights to match instrument behavior.
• Dynamic thresholds adapt to changing volatility and drift.
• Transparent diagnostics from the on-chart table.
• Flexible smoothing choices, including ALMA and T3.
Limitations and cautions
• Z-scores assume a reasonably stationary window. Sharp regime shifts can make recent bands unrepresentative.
• Highly correlated components can overweight the same effect. Consider adjusting weights to avoid double counting.
• More smoothing adds lag. Less smoothing adds noise.
• Dynamic bands recalibrate with dynLen ; if set too short, bands may swing excessively. If too long, bands can be slow to adapt.
Practical tuning tips
• Trending symbols: increase w_dm , use a modest smoother like EMA or T3, and use centerZero dynamic bands.
• Choppy symbols: increase w_rsi and w_cci , consider ALMA with a higher sigma , and widen bands with a larger thres_k .
• Multiday swing charts: lengthen lookback and dynLen to stabilize the scale.
• Lower timeframes: shorten component lookbacks slightly and reduce smoothing to keep signals timely.
Alerts
• Enter and exit of Overbought and Oversold, based on the active band choice.
• Bullish and bearish zero crosses.
Use alerts as prompts to review context rather than as stand-alone trade commands.
Final Remarks
We created this to show people a different way of making indicators & trading.
You can process normal indicators in multiple ways to enhance or change the signal, especially with this you can utilise machine learning to optimise the weights, then trade accordingly.
All of the different components were selected to give some sort of signal, its made out of simple components yet is effective. As long as the user calibrates it to their Trading/ investing style you can find good results. Do not use anything standalone, ensure you are backtesting and creating a proper system.
VWAP For Loop [BackQuant]VWAP For Loop
What this tool does—in one sentence
A volume-weighted trend gauge that anchors VWAP to a calendar period (day/week/month/quarter/year) and then scores the persistence of that VWAP trend with a simple for-loop “breadth” count; the result is a clean, threshold-driven oscillator plus an optional VWAP overlay and alerts.
Plain-English overview
Instead of judging raw price alone, this indicator focuses on anchored VWAP —the market’s average price paid during your chosen institutional period. It then asks a simple question across a configurable set of lookback steps: “Is the current anchored VWAP higher than it was i bars ago—or lower?” Each “yes” adds +1, each “no” adds −1. Summing those answers creates a score that reflects how consistently the volume-weighted trend has been rising or falling. Extreme positive scores imply persistent, broad strength; deeply negative scores imply persistent weakness. Crossing predefined thresholds produces objective long/short events and color-coded context.
Under the hood
• Anchoring — VWAP using hlc3 × volume resets exactly when the selected period rolls:
Day → session change, Week → new week, Month → new month, Quarter/Year → calendar quarter/year.
• For-loop scoring — For lag steps i = , compare today’s VWAP to VWAP .
– If VWAP > VWAP , add +1.
– Else, add −1.
The final score ∈ , where N = (end − start + 1). With defaults (1→45), N = 45.
• Signal logic (stateful)
– Long when score > upper (e.g., > 40 with N = 45 → VWAP higher than ~89% of checked lags).
– Short on crossunder of lower (e.g., dropping below −10).
– A compact state variable ( out ) holds the current regime: +1 (long), −1 (short), otherwise unchanged. This “stickiness” avoids constant flipping between bars without sufficient evidence.
Why VWAP + a breadth score?
• VWAP aggregates both price and volume—where participants actually traded.
• The breadth-style count rewards consistency of the anchored trend, not one-off spikes.
• Thresholds give you binary structure when you need it (alerts, automation), without complex math.
What you’ll see on the chart
• Sub-pane oscillator — The for-loop score line, colored by regime (long/short/neutral).
• Main-pane VWAP (optional) — Even though the indicator runs off-chart, the anchored VWAP can be overlaid on price (toggle visibility and whether it inherits trend colors).
• Threshold guides — Horizontal lines for the long/short bands (toggle).
• Cosmetics — Optional candle painting and background shading by regime; adjustable line width and colors.
Input map (quick reference)
• VWAP Anchor Period — Day, Week, Month, Quarter, Year.
• Calculation Start/End — The for-loop lag window . With 1→45, you evaluate 45 comparisons.
• Long/Short Thresholds — Default upper=40, lower=−10 (asymmetric by design; see below).
• UI/Style — Show thresholds, paint candles, background color, line width, VWAP visibility and coloring, custom long/short colors.
Interpreting the score
• Near +N — Current anchored VWAP is above most historical VWAP checkpoints in the window → entrenched strength.
• Near −N — Current anchored VWAP is below most checkpoints → entrenched weakness.
• Between — Mixed, choppy, or transitioning regimes; use thresholds to avoid reacting to noise.
Why the asymmetric default thresholds?
• Long = score > upper (40) — Demands unusually broad upside persistence before declaring “long regime.”
• Short = crossunder lower (−10) — Triggers only on downward momentum events (a fresh breach), not merely being below −10. This combination tends to:
– Capture sustained uptrends only when they’re very strong.
– Flag downside turns as they occur, rather than waiting for an extreme negative breadth.
Tuning guide
Choose an anchor that matches your horizon
– Intraday scalps : Day anchor on intraday charts.
– Swing/position : Month or Quarter anchor on 1h/4h/D charts to capture institutional cycles.
Pick the for-loop window
– Larger N (bigger end) = stronger evidence requirement, smoother oscillator.
– Smaller N = faster, more reactive score.
Set achievable thresholds
– Ensure upper ≤ N and lower ≥ −N ; if N=30, an upper of 40 can never trigger.
– Symmetric setups (e.g., +20/−20) are fine if you want balanced behavior.
Match visuals to intent
– Enabling VWAP coloring lets you see regime directly on price.
– Background shading is useful for discretionary reading; turn it off for cleaner automation displays.
Playbook examples
• Trend confirmation with disciplined entries — On Month anchor, N=45, upper=38–42: when the long regime engages, use pullbacks toward anchored VWAP on the main pane for entries, with stops just beyond VWAP or a recent swing.
• Downside transition detection — Keep lower around −8…−12 and watch for crossunders; combine with price losing anchored VWAP to validate risk-off.
• Intraday bias filter — Day anchor on a 5–15m chart, N=20–30, upper ~ 16–20, lower ~ −6…−10. Only take longs while score is positive and above a midline you define (e.g., 0), and shorts only after a genuine crossunder.
Behavior around resets (important)
Anchored VWAP is hard-reset each period. Immediately after a reset, the series can be young and comparisons to pre-reset values may span two periods. If you prefer within-period evaluation only, choose end small enough not to bridge typical period length on your timeframe, or accept that the breadth test intentionally spans regimes.
Alerts included
• VWAP FL Long — Fires when the long condition is true (score > upper and not in short).
• VWAP FL Short — Fires on crossunder of the lower threshold (event-driven).
Messages include {{ticker}} and {{interval}} placeholders for routing.
Strengths
• Simple, transparent math — Easy to reason about and validate.
• Volume-aware by construction — Decisions reference VWAP, not just price.
• Robust to single-bar noise — Needs many lags to agree before flipping state (by design, via thresholds and the stateful output).
Limitations & cautions
• Threshold feasibility — If N < upper or |lower| > N, signals will never trigger; always cross-check N.
• Path dependence — The state variable persists until a new event; if you want frequent re-evaluation, lower thresholds or reduce N.
• Regime changes — Calendar resets can produce early ambiguity; expect a few bars for the breadth to mature.
• VWAP sensitivity to volume spikes — Large prints can tilt VWAP abruptly; that behavior is intentional in VWAP-based logic.
Suggested starting profiles
• Intraday trend bias : Anchor=Day, N=25 (1→25), upper=18–20, lower=−8, paint candles ON.
• Swing bias : Anchor=Month, N=45 (1→45), upper=38–42, lower=−10, VWAP coloring ON, background OFF.
• Balanced reactivity : Anchor=Week, N=30 (1→30), upper=20–22, lower=−10…−12, symmetric if desired.
Implementation notes
• The indicator runs in a separate pane (oscillator), but VWAP itself is drawn on price using forced overlay so you can see interactions (touches, reclaim/loss).
• HLC3 is used for VWAP price; that’s a common choice to dampen wick noise while still reflecting intrabar range.
• For-loop cap is kept modest (≤50) for performance and clarity.
How to use this responsibly
Treat the oscillator as a bias and persistence meter . Combine it with your entry framework (structure breaks, liquidity zones, higher-timeframe context) and risk controls. The design emphasizes clarity over complexity—its edge is in how strictly it demands agreement before declaring a regime, not in predicting specific turns.
Summary
VWAP For Loop distills the question “How broadly is the anchored, volume-weighted trend advancing or retreating?” into a single, thresholded score you can read at a glance, alert on, and color through your chart. With careful anchoring and thresholds sized to your window length, it becomes a pragmatic bias filter for both systematic and discretionary workflows.
Machine Learning BBPct [BackQuant]Machine Learning BBPct
What this is (in one line)
A Bollinger Band %B oscillator enhanced with a simplified K-Nearest Neighbors (KNN) pattern matcher. The model compares today’s context (volatility, momentum, volume, and position inside the bands) to similar situations in recent history and blends that historical consensus back into the raw %B to reduce noise and improve context awareness. It is informational and diagnostic—designed to describe market state, not to sell a trading system.
Background: %B in plain terms
Bollinger %B measures where price sits inside its dynamic envelope: 0 at the lower band, 1 at the upper band, ~ 0.5 near the basis (the moving average). Readings toward 1 indicate pressure near the envelope’s upper edge (often strength or stretch), while readings toward 0 indicate pressure near the lower edge (often weakness or stretch). Because bands adapt to volatility, %B is naturally comparable across regimes.
Why add (simplified) KNN?
Classic %B is reactive and can be whippy in fast regimes. The simplified KNN layer builds a “nearest-neighbor memory” of recent market states and asks: “When the market looked like this before, where did %B tend to be next bar?” It then blends that estimate with the current %B. Key ideas:
• Feature vector . Each bar is summarized by up to five normalized features:
– %B itself (normalized)
– Band width (volatility proxy)
– Price momentum (ROC)
– Volume momentum (ROC of volume)
– Price position within the bands
• Distance metric . Euclidean distance ranks the most similar recent bars.
• Prediction . Average the neighbors’ prior %B (lagged to avoid lookahead), inverse-weighted by distance.
• Blend . Linearly combine raw %B and KNN-predicted %B with a configurable weight; optional filtering then adapts to confidence.
This remains “simplified” KNN: no training/validation split, no KD-trees, no scaling beyond windowed min-max, and no probabilistic calibration.
How the script is organized (by input groups)
1) BBPct Settings
• Price Source – Which price to evaluate (%B is computed from this).
• Calculation Period – Lookback for SMA basis and standard deviation.
• Multiplier – Standard deviation width (e.g., 2.0).
• Apply Smoothing / Type / Length – Optional smoothing of the %B stream before ML (EMA, RMA, DEMA, TEMA, LINREG, HMA, etc.). Turning this off gives you the raw %B.
2) Thresholds
• Overbought/Oversold – Default 0.8 / 0.2 (inside ).
• Extreme OB/OS – Stricter zones (e.g., 0.95 / 0.05) to flag stretch conditions.
3) KNN Machine Learning
• Enable KNN – Switch between pure %B and hybrid.
• K (neighbors) – How many historical analogs to blend (default 8).
• Historical Period – Size of the search window for neighbors.
• ML Weight – Blend between raw %B and KNN estimate.
• Number of Features – Use 2–5 features; higher counts add context but raise the risk of overfitting in short windows.
4) Filtering
• Method – None, Adaptive, Kalman-style (first-order),
or Hull smoothing.
• Strength – How aggressively to smooth. “Adaptive” uses model confidence to modulate its alpha: higher confidence → stronger reliance on the ML estimate.
5) Performance Tracking
• Win-rate Period – Simple running score of past signal outcomes based on target/stop/time-out logic (informational, not a robust backtest).
• Early Entry Lookback – Horizon for forecasting a potential threshold cross.
• Profit Target / Stop Loss – Used only by the internal win-rate heuristic.
6) Self-Optimization
• Enable Self-Optimization – Lightweight, rolling comparison of a few canned settings (K = 8/14/21 via simple rules on %B extremes).
• Optimization Window & Stability Threshold – Governs how quickly preferred K changes and how sensitive the overfitting alarm is.
• Adaptive Thresholds – Adjust the OB/OS lines with volatility regime (ATR ratio), widening in calm markets and tightening in turbulent ones (bounded 0.7–0.9 and 0.1–0.3).
7) UI Settings
• Show Table / Zones / ML Prediction / Early Signals – Toggle informational overlays.
• Signal Line Width, Candle Painting, Colors – Visual preferences.
Step-by-step logic
A) Compute %B
Basis = SMA(source, len); dev = stdev(source, len) × multiplier; Upper/Lower = Basis ± dev.
%B = (price − Lower) / (Upper − Lower). Optional smoothing yields standardBB .
B) Build the feature vector
All features are min-max normalized over the KNN window so distances are in comparable units. Features include normalized %B, normalized band width, normalized price ROC, normalized volume ROC, and normalized position within bands. You can limit to the first N features (2–5).
C) Find nearest neighbors
For each bar inside the lookback window, compute the Euclidean distance between current features and that bar’s features. Sort by distance, keep the top K .
D) Predict and blend
Use inverse-distance weights (with a strong cap for near-zero distances) to average neighbors’ prior %B (lagged by one bar). This becomes the KNN estimate. Blend it with raw %B via the ML weight. A variance of neighbor %B around the prediction becomes an uncertainty proxy ; combined with a stability score (how long parameters remain unchanged), it forms mlConfidence ∈ . The Adaptive filter optionally transforms that confidence into a smoothing coefficient.
E) Adaptive thresholds
Volatility regime (ATR(14) divided by its 50-bar SMA) nudges OB/OS thresholds wider or narrower within fixed bounds. The aim: comparable extremeness across regimes.
F) Early entry heuristic
A tiny two-step slope/acceleration probe extrapolates finalBB forward a few bars. If it is on track to cross OB/OS soon (and slope/acceleration agree), it flags an EARLY_BUY/SELL candidate with an internal confidence score. This is explicitly a heuristic—use as an attention cue, not a signal by itself.
G) Informational win-rate
The script keeps a rolling array of trade outcomes derived from signal transitions + rudimentary exits (target/stop/time). The percentage shown is a rough diagnostic , not a validated backtest.
Outputs and visual language
• ML Bollinger %B (finalBB) – The main line after KNN blending and optional filtering.
• Gradient fill – Greenish tones above 0.5, reddish below, with intensity following distance from the midline.
• Adaptive zones – Overbought/oversold and extreme bands; shaded backgrounds appear at extremes.
• ML Prediction (dots) – The KNN estimate plotted as faint circles; becomes bright white when confidence > 0.7.
• Early arrows – Optional small triangles for approaching OB/OS.
• Candle painting – Light green above the midline, light red below (optional).
• Info panel – Current value, signal classification, ML confidence, optimized K, stability, volatility regime, adaptive thresholds, overfitting flag, early-entry status, and total signals processed.
Signal classification (informational)
The indicator does not fire trade commands; it labels state:
• STRONG_BUY / STRONG_SELL – finalBB beyond extreme OS/OB thresholds.
• BUY / SELL – finalBB beyond adaptive OS/OB.
• EARLY_BUY / EARLY_SELL – forecast suggests a near-term cross with decent internal confidence.
• NEUTRAL – between adaptive bands.
Alerts (what you can automate)
• Entering adaptive OB/OS and extreme OB/OS.
• Midline cross (0.5).
• Overfitting detected (frequent parameter flipping).
• Early signals when early confidence > 0.7.
These are purely descriptive triggers around the indicator’s state.
Practical interpretation
• Mean-reversion context – In range markets, adaptive OS/OB with ML smoothing can reduce whipsaws relative to raw %B.
• Trend context – In persistent trends, the KNN blend can keep finalBB nearer the mid/upper region during healthy pullbacks if history supports similar contexts.
• Regime awareness – Watch the volatility regime and adaptive thresholds. If thresholds compress (high vol), “OB/OS” comes sooner; if thresholds widen (calm), it takes more stretch to flag.
• Confidence as a weight – High mlConfidence implies neighbors agree; you may rely more on the ML curve. Low confidence argues for de-emphasizing ML and leaning on raw %B or other tools.
• Stability score – Rising stability indicates consistent parameter selection and fewer flips; dropping stability hints at a shifting backdrop.
Methodological notes
• Normalization uses rolling min-max over the KNN window. This is simple and scale-agnostic but sensitive to outliers; the distance metric will reflect that.
• Distance is unweighted Euclidean. If you raise featureCount, you increase dimensionality; consider keeping K larger and lookback ample to avoid sparse-neighbor artifacts.
• Lag handling intentionally uses neighbors’ previous %B for prediction to avoid lookahead bias.
• Self-optimization is deliberately modest: it only compares a few canned K/threshold choices using simple “did an extreme anticipate movement?” scoring, then enforces a stability regime and an overfitting guard. It is not a grid search or GA.
• Kalman option is a first-order recursive filter (fixed gain), not a full state-space estimator.
• Hull option derives a dynamic length from 1/strength; it is a convenience smoothing alternative.
Limitations and cautions
• Non-stationarity – Nearest neighbors from the recent window may not represent the future under structural breaks (policy shifts, liquidity shocks).
• Curse of dimensionality – Adding features without sufficient lookback can make genuine neighbors rare.
• Overfitting risk – The script includes a crude overfitting detector (frequent parameter flips) and will fall back to defaults when triggered, but this is only a guardrail.
• Win-rate display – The internal score is illustrative; it does not constitute a tradable backtest.
• Latency vs. smoothness – Smoothing and ML blending reduce noise but add lag; tune to your timeframe and objectives.
Tuning guide
• Short-term scalping – Lower len (10–14), slightly lower multiplier (1.8–2.0), small K (5–8), featureCount 3–4, Adaptive filter ON, moderate strength.
• Swing trading – len (20–30), multiplier ~2.0, K (8–14), featureCount 4–5, Adaptive thresholds ON, filter modest.
• Strong trends – Consider higher adaptive_upper/lower bounds (or let volatility regime do it), keep ML weight moderate so raw %B still reflects surges.
• Chop – Higher ML weight and stronger Adaptive filtering; accept lag in exchange for fewer false extremes.
How to use it responsibly
Treat this as a state descriptor and context filter. Pair it with your execution signals (structure breaks, volume footprints, higher-timeframe bias) and risk management. If mlConfidence is low or stability is falling, lean less on the ML line and more on raw %B or external confirmation.
Summary
Machine Learning BBPct augments a familiar oscillator with a transparent, simplified KNN memory of recent conditions. By blending neighbors’ behavior into %B and adapting thresholds to volatility regime—while exposing confidence, stability, and a plain early-entry heuristic—it provides an informational, probability-minded view of stretch and reversion that you can interpret alongside your own process.
Seasonality Monte Carlo Forecaster [BackQuant]Seasonality Monte Carlo Forecaster
Plain-English overview
This tool projects a cone of plausible future prices by combining two ideas that traders already use intuitively: seasonality and uncertainty. It watches how your market typically behaves around this calendar date, turns that seasonal tendency into a small daily “drift,” then runs many randomized price paths forward to estimate where price could land tomorrow, next week, or a month from now. The result is a probability cone with a clear expected path, plus optional overlays that show how past years tended to move from this point on the calendar. It is a planning tool, not a crystal ball: the goal is to quantify ranges and odds so you can size, place stops, set targets, and time entries with more realism.
What Monte Carlo is and why quants rely on it
• Definition . Monte Carlo simulation is a way to answer “what might happen next?” when there is randomness in the system. Instead of producing a single forecast, it generates thousands of alternate futures by repeatedly sampling random shocks and adding them to a model of how prices evolve.
• Why it is used . Markets are noisy. A single point forecast hides risk. Monte Carlo gives a distribution of outcomes so you can reason in probabilities: the median path, the 68% band, the 95% band, tail risks, and the chance of hitting a specific level within a horizon.
• Core strengths in quant finance .
– Path-dependent questions : “What is the probability we touch a stop before a target?” “What is the expected drawdown on the way to my objective?”
– Pricing and risk : Useful for path-dependent options, Value-at-Risk (VaR), expected shortfall (CVaR), stress paths, and scenario analysis when closed-form formulas are unrealistic.
– Planning under uncertainty : Portfolio construction and rebalancing rules can be tested against a cloud of plausible futures rather than a single guess.
• Why it fits trading workflows . It turns gut feel like “seasonality is supportive here” into quantitative ranges: “median path suggests +X% with a 68% band of ±Y%; stop at Z has only ~16% odds of being tagged in N days.”
How this indicator builds its probability cone
1) Seasonal pattern discovery
The script builds two day-of-year maps as new data arrives:
• A return map where each calendar day stores an exponentially smoothed average of that day’s log return (yesterday→today). The smoothing (90% old, 10% new) behaves like an EWMA, letting older seasons matter while adapting to new information.
• A volatility map that tracks the typical absolute return for the same calendar day.
It calculates the day-of-year carefully (with leap-year adjustment) and indexes into a 365-slot seasonal array so “March 18” is compared with past March 18ths. This becomes the seasonal bias that gently nudges simulations up or down on each forecast day.
2) Choice of randomness engine
You can pick how the future shocks are generated:
• Daily mode uses a Gaussian draw with the seasonal bias as the mean and a volatility that comes from realized returns, scaled down to avoid over-fitting. It relies on the Box–Muller transform internally to turn two uniform random numbers into one normal shock.
• Weekly mode uses bootstrap sampling from the seasonal return history (resampling actual historical daily drifts and then blending in a fraction of the seasonal bias). Bootstrapping is robust when the empirical distribution has asymmetry or fatter tails than a normal distribution.
Both modes seed their random draws deterministically per path and day, which makes plots reproducible bar-to-bar and avoids flickering bands.
3) Volatility scaling to current conditions
Markets do not always live in average volatility. The engine computes a simple volatility factor from ATR(20)/price and scales the simulated shocks up or down within sensible bounds (clamped between 0.5× and 2.0×). When the current regime is quiet, the cone narrows; when ranges expand, the cone widens. This prevents the classic mistake of projecting calm markets into a storm or vice versa.
4) Many futures, summarized by percentiles
The model generates a matrix of price paths (capped at 100 runs for performance inside TradingView), each path stepping forward for your selected horizon. For each forecast day it sorts the simulated prices and pulls key percentiles:
• 5th and 95th → approximate 95% band (outer cone).
• 16th and 84th → approximate 68% band (inner cone).
• 50th → the median or “expected path.”
These are drawn as polylines so you can immediately see central tendency and dispersion.
5) A historical overlay (optional)
Turn on the overlay to sketch a dotted path of what a purely seasonal projection would look like for the next ~30 days using only the return map, no randomness. This is not a forecast; it is a visual reminder of the seasonal drift you are biasing toward.
Inputs you control and how to think about them
Monte Carlo Simulation
• Price Series for Calculation . The source series, typically close.
• Enable Probability Forecasts . Master switch for simulation and drawing.
• Simulation Iterations . Requested number of paths to run. Internally capped at 100 to protect performance, which is generally enough to estimate the percentiles for a trading chart. If you need ultra-smooth bands, shorten the horizon.
• Forecast Days Ahead . The length of the cone. Longer horizons dilute seasonal signal and widen uncertainty.
• Probability Bands . Draw all bands, just 95%, just 68%, or a custom level (display logic remains 68/95 internally; the custom number is for labeling and color choice).
• Pattern Resolution . Daily leans on day-of-year effects like “turn-of-month” or holiday patterns. Weekly biases toward day-of-week tendencies and bootstraps from history.
• Volatility Scaling . On by default so the cone respects today’s range context.
Plotting & UI
• Probability Cone . Plots the outer and inner percentile envelopes.
• Expected Path . Plots the median line through the cone.
• Historical Overlay . Dotted seasonal-only projection for context.
• Band Transparency/Colors . Customize primary (outer) and secondary (inner) band colors and the mean path color. Use higher transparency for cleaner charts.
What appears on your chart
• A cone starting at the most recent bar, fanning outward. The outer lines are the ~95% band; the inner lines are the ~68% band.
• A median path (default blue) running through the center of the cone.
• An info panel on the final historical bar that summarizes simulation count, forecast days, number of seasonal patterns learned, the current day-of-year, expected percentage return to the median, and the approximate 95% half-range in percent.
• Optional historical seasonal path drawn as dotted segments for the next 30 bars.
How to use it in trading
1) Position sizing and stop logic
The cone translates “volatility plus seasonality” into distances.
• Put stops outside the inner band if you want only ~16% odds of a stop-out due to noise before your thesis can play.
• Size positions so that a test of the inner band is survivable and a test of the outer band is rare but acceptable.
• If your target sits inside the 68% band at your horizon, the payoff is likely modest; outside the 68% but inside the 95% can justify “one-good-push” trades; beyond the 95% band is a low-probability flyer—consider scaling plans or optionality.
2) Entry timing with seasonal bias
When the median path slopes up from this calendar date and the cone is relatively narrow, a pullback toward the lower inner band can be a high-quality entry with a tight invalidation. If the median slopes down, fade rallies toward the upper band or step aside if it clashes with your system.
3) Target selection
Project your time horizon to N bars ahead, then pick targets around the median or the opposite inner band depending on your style. You can also anchor dynamic take-profits to the moving median as new bars arrive.
4) Scenario planning & “what-ifs”
Before events, glance at the cone: if the 95% band already spans a huge range, trade smaller, expect whips, and avoid placing stops at obvious band edges. If the cone is unusually tight, consider breakout tactics and be ready to add if volatility expands beyond the inner band with follow-through.
5) Options and vol tactics
• When the cone is tight : Prefer long gamma structures (debit spreads) only if you expect a regime shift; otherwise premium selling may dominate.
• When the cone is wide : Debit structures benefit from range; credit spreads need wider wings or smaller size. Align with your separate IV metrics.
Reading the probability cone like a pro
• Cone slope = seasonal drift. Upward slope means the calendar has historically favored positive drift from this date, downward slope the opposite.
• Cone width = regime volatility. A widening fan tells you that uncertainty grows fast; a narrow cone says the market typically stays contained.
• Mean vs. price gap . If spot trades well above the median path and the upper band, mean-reversion risk is high. If spot presses the lower inner band in an up-sloping cone, you are in the “buy fear” zone.
• Touches and pierces . Touching the inner band is common noise; piercing it with momentum signals potential regime change; the outer band should be rare and often brings snap-backs unless there is a structural catalyst.
Methodological notes (what the code actually does)
• Log returns are used for additivity and better statistical behavior: sim_ret is applied via exp(sim_ret) to evolve price.
• Seasonal arrays are updated online with EWMA (90/10) so the model keeps learning as each bar arrives.
• Leap years are handled; indexing still normalizes into a 365-slot map so the seasonal pattern remains stable.
• Gaussian engine (Daily mode) centers shocks on the seasonal bias with a conservative standard deviation.
• Bootstrap engine (Weekly mode) resamples from observed seasonal returns and adds a fraction of the bias, which captures skew and fat tails better.
• Volatility adjustment multiplies each daily shock by a factor derived from ATR(20)/price, clamped between 0.5 and 2.0 to avoid extreme cones.
• Performance guardrails : simulations are capped at 100 paths; the probability cone uses polylines (no heavy fills) and only draws on the last confirmed bar to keep charts responsive.
• Prerequisite data : at least ~30 seasonal entries are required before the model will draw a cone; otherwise it waits for more history.
Strengths and limitations
• Strengths :
– Probabilistic thinking replaces single-point guessing.
– Seasonality adds a small but meaningful directional bias that many markets exhibit.
– Volatility scaling adapts to the current regime so the cone stays realistic.
• Limitations :
– Seasonality can break around structural changes, policy shifts, or one-off events.
– The number of paths is performance-limited; percentile estimates are good for trading, not for academic precision.
– The model assumes tomorrow’s randomness resembles recent randomness; if regime shifts violently, the cone will lag until the EWMA adapts.
– Holidays and missing sessions can thin the seasonal sample for some assets; be cautious with very short histories.
Tuning guide
• Horizon : 10–20 bars for tactical trades; 30+ for swing planning when you care more about broad ranges than precise targets.
• Iterations : The default 100 is enough for stable 5/16/50/84/95 percentiles. If you crave smoother lines, shorten the horizon or run on higher timeframes.
• Daily vs. Weekly : Daily for equities and crypto where month-end and turn-of-month effects matter; Weekly for futures and FX where day-of-week behavior is strong.
• Volatility scaling : Keep it on. Turn off only when you intentionally want a “pure seasonality” cone unaffected by current turbulence.
Workflow examples
• Swing continuation : Cone slopes up, price pulls into the lower inner band, your system fires. Enter near the band, stop just outside the outer line for the next 3–5 bars, target near the median or the opposite inner band.
• Fade extremes : Cone is flat or down, price gaps to the upper outer band on news, then stalls. Favor mean-reversion toward the median, size small if volatility scaling is elevated.
• Event play : Before CPI or earnings on a proxy index, check cone width. If the inner band is already wide, cut size or prefer options structures that benefit from range.
Good habits
• Pair the cone with your entry engine (breakout, pullback, order flow). Let Monte Carlo do range math; let your system do signal quality.
• Do not anchor blindly to the median; recalc after each bar. When the cone’s slope flips or width jumps, the plan should adapt.
• Validate seasonality for your symbol and timeframe; not every market has strong calendar effects.
Summary
The Seasonality Monte Carlo Forecaster wraps institutional risk planning into a single overlay: a data-driven seasonal drift, realistic volatility scaling, and a probabilistic cone that answers “where could we be, with what odds?” within your trading horizon. Use it to place stops where randomness is less likely to take you out, to set targets aligned with realistic travel, and to size positions with confidence born from distributions rather than hunches. It will not predict the future, but it will keep your decisions anchored to probabilities—the language markets actually speak.
Dip Hunter [BackQuant]Dip Hunter
What this tool does in plain language
Dip Hunter is a pullback detector designed to find high quality buy-the-dip opportunities inside healthy trends and to avoid random knife catches. It watches for a quick drop from a recent high, checks that the drop happened with meaningful participation and volatility, verifies short-term weakness inside a larger uptrend, then scores the setup and paints the chart so you can act with confidence. It also draws clean entry lines, provides a meter that shows dip strength at a glance, and ships with alerts that match common execution workflows.
How Dip Hunter thinks
It defines a recent swing reference, measures how far price has dipped off that high, and only looks at candidates that meet your minimum percentage drop.
It confirms the dip with real activity by requiring a volume spike and a volatility spike.
It checks structure with two EMAs. Price should be weak in the short term while the larger context remains constructive.
It optionally requires a higher-timeframe trend to be up so you focus on pullbacks in trending markets.
It bundles those checks into a score and shows you the score on the candles and on a gradient meter.
When everything lines up it paints a green triangle below the bar, shades the background, and (if you wish) draws a horizontal entry line at your chosen level.
Inputs and what they mean
Dip Hunter Settings
• Vol Lookback and Vol Spike : The script computes an average volume over the lookback window and flags a spike when current volume is a multiple of that average. A multiplier of 2.0 means today’s volume must be at least double the average. This helps filter noise and focuses on dips that other traders actually traded.
• Fast EMA and Slow EMA : Short-term and medium-term structure references. A dip is more credible if price closes below the fast EMA while the fast EMA is still below the slow EMA during the pullback. That is classic corrective behavior inside a larger trend.
• Price Smooth : Optional smoothing length for price-derived series. Use this if you trade very noisy assets or low timeframes.
• Volatility Len and Vol Spike (volatility) : The script checks both standard deviation and true range against their own averages. If either expands beyond your multiplier the market confirms the move with range.
• Dip % and Lookback Bars : The engine finds the highest high over the lookback window, then computes the percentage drawdown from that high to the current close. Only dips larger than your threshold qualify.
Trend Filter
• Enable Trend Filter : When on, Dip Hunter will only trigger if the market is in an uptrend.
• Trend EMA Period : The longer EMA that defines the session’s backbone trend.
• Minimum Trend Strength : A small positive slope requirement. In practice this means the trend EMA should be rising, and price should be above it. You can raise the value to be more selective.
Entries
• Show Entry Lines : Draws a horizontal guide from the signal bar for a fixed number of bars. Great for limit orders, scaling, or re-tests.
• Line Length (bars) : How far the entry guide extends.
• Min Gap (bars) : Suppresses new entry lines if another dip fired recently. Prevents clutter during choppy sequences.
• Entry Price : Choose the line level. “Low” anchors at the signal candle’s low. “Close” anchors at the signal close. “Dip % Level” anchors at the theoretical level defined by recent_high × (1 − dip%). This lets you work resting orders at a consistent discount.
Heat / Meter
• Color Bars by Score : Colors each candle using a red→white→green gradient. Red is overheated, green is prime dip territory, white is neutral.
• Show Meter Table : Adds a compact gradient strip with a pointer that tracks the current score.
• Meter Cells and Meter Position : Resolution and placement of the meter.
UI Settings
• Show Dip Signals : Plots green triangles under qualifying bars and tints the background very lightly.
• Show EMAs : Plots fast, slow, and the trend EMA (if the trend filter is enabled).
• Bullish, Bearish, Neutral colors : Theme controls for shapes, fills, and bar painting.
Core calculations explained simply
Recent high and dip percent
The script finds the highest high over Lookback Bars , calls it “recent high,” then calculates:
dip% = (recent_high − close) ÷ recent_high × 100.
If dip% is larger than Dip % , condition one passes.
Volume confirmation
It computes a simple moving average of volume over Vol Lookback . If current volume ÷ average volume > Vol Spike , we have a participation spike. It also checks 5-bar ROC of volume. If ROC > 50 the spike is forceful. This gets an extra score point.
Volatility confirmation
Two independent checks:
• Standard deviation of closes vs its own average.
• True range vs ATR.
If either expands beyond Vol Spike (volatility) the move has range. This prevents false triggers from quiet drifts.
Short-term structure
Price should close below the Fast EMA and the fast EMA should be below the Slow EMA at the moment of the dip. That is the anatomy of a pullback rather than a full breakdown.
Macro trend context (optional)
When Enable Trend Filter is on, the Trend EMA must be rising and price must be above it. The logic prefers “micro weakness inside macro strength” which is the highest probability pattern for buying dips.
Signal formation
A valid dip requires:
• dip% > threshold
• volume spike true
• volatility spike true
• close below fast EMA
• fast EMA below slow EMA
If the trend filter is enabled, a rising trend EMA with price above it is also required. When all true, the triangle prints, the background tints, and optional entry lines are drawn.
Scoring and visuals
Binary checks into a continuous score
Each component contributes to a score between 0 and 1. The script then rescales to a centered range (−50 to +50).
• Low or negative scores imply “overheated” conditions and are shaded toward red.
• High positive scores imply “ripe for a dip buy” conditions and are shaded toward green.
• The gradient meter repeats the same logic, with a pointer so you can read the state quickly.
Bar coloring
If you enable “Color Bars by Score,” each candle inherits the gradient. This makes sequences obvious. Red clusters warn you not to buy. White means neutral. Increasing green suggests the pullback is maturing.
EMAs and the trend EMA
• Fast EMA turns down relative to the slow EMA inside the pullback.
• Trend EMA stays rising and above price once the dip exhausts, which is your cue to focus on long setups rather than bottom fishing in downtrends.
Entry lines
When a fresh signal fires and no other signal happened within Min Gap (bars) , the indicator draws a horizontal level for Line Length bars. Use these lines for limit entries at the low, at the close, or at the defined dip-percent level. This keeps your plan consistent across instruments.
Alerts and what they mean
• Market Overheated : Score is deeply negative. Do not chase. Wait for green.
• Close To A Dip : Score has reached a healthy level but the full signal did not trigger yet. Prepare orders.
• Dip Confirmed : First bar of a fresh validated dip. This is the most direct entry alert.
• Dip Active : The dip condition remains valid. You can scale in on re-tests.
• Dip Fading : Score crosses below 0.5 from above. Momentum of the setup is fading. Tighten stops or take partials.
• Trend Blocked Signal : All dip conditions passed but the trend filter is offside. Either reduce risk or skip, depending on your plan.
How to trade with Dip Hunter
Classic pullback in uptrend
Turn on the trend filter.
Watch for a Dip Confirmed alert with green triangle.
Use the entry line at “Dip % Level” to stage a limit order. This keeps your entries consistent across assets and timeframes.
Initial stop under the signal bar’s low or under the next lower EMA band.
First target at prior swing high, second target at a multiple of risk.
If you use partials, trail the remainder under the fast EMA once price reclaims it.
Aggressive intraday scalps
Lower Dip % and Lookback Bars so you catch shallow flags.
Keep Vol Spike meaningful so you only trade when participation appears.
Take quick partials when price reclaims the fast EMA, then exit on Dip Fading if momentum stalls.
Counter-trend probes
Disable the trend filter if you intentionally hunt reflex bounces in downtrends.
Require strong volume and volatility confirmation.
Use smaller size and faster targets. The meter should move quickly from red toward white and then green. If it does not, step aside.
Risk management templates
Stops
• Conservative: below the entry line minus a small buffer or below the signal bar’s low.
• Structural: below the slow EMA if you aim for swing continuation.
• Time stop: if price does not reclaim the fast EMA within N bars, exit.
Position sizing
Use the distance between the entry line and your structural stop to size consistently. The script’s entry lines make this distance obvious.
Scaling
• Scale at the entry line first touch.
• Add only if the meter stays green and price reclaims the fast EMA.
• Stop adding on a Dip Fading alert.
Tuning guide by market and timeframe
Equities daily
• Dip %: 1.5 to 3.0
• Lookback Bars: 5 to 10
• Vol Spike: 1.5 to 2.5
• Volatility Len: 14 to 20
• Trend EMA: 100 or 200
• Keep trend filter on for a cleaner list.
Futures and FX intraday
• Dip %: 0.4 to 1.2
• Lookback Bars: 3 to 7
• Vol Spike: 1.8 to 3.0
• Volatility Len: 10 to 14
• Use Min Gap to avoid clusters during news.
Crypto
• Dip %: 3.0 to 6.0 for majors on higher timeframes, lower on 15m to 1h
• Lookback Bars: 5 to 12
• Vol Spike: 1.8 to 3.0
• ATR and stdev checks help in erratic sessions.
Reading the chart at a glance
• Green triangle below the bar: a validated dip.
• Light green background: the current bar meets the full condition.
• Bar gradient: red is overheated, white is neutral, green is dip-friendly.
• EMAs: fast below slow during the pullback, then reclaim fast EMA on the bounce for quality continuation.
• Trend EMA: a rising spine when the filter is on.
• Entry line: a fixed level to anchor orders and risk.
• Meter pointer: right side toward “Dip” means conditions are maturing.
Why this combination reduces false positives
Any single criterion will trigger too often. Dip Hunter demands a dip off a recent high plus a volume surge plus a volatility expansion plus corrective EMA structure. Optional trend alignment pushes odds further in your favor. The score and meter visualize how many of these boxes you are actually ticking, which is more reliable than a binary dot.
Limitations and practical tips
• Thin or illiquid symbols can spoof volume spikes. Use larger Vol Lookback or raise Vol Spike .
• Sideways markets will show frequent small dips. Increase Dip % or keep the trend filter on.
• News candles can blow through entry lines. Widen stops or skip around known events.
• If you see many back-to-back triangles, raise Min Gap to keep only the best setups.
Quick setup recipes
• Clean swing trader: Trend filter on, Dip % 2.0 to 3.0, Vol Spike 2.0, Volatility Len 14, Fast 20 EMA, Slow 50 EMA, Trend 100 EMA.
• Fast intraday scalper: Trend filter off, Dip % 0.7 to 1.0, Vol Spike 2.5, Volatility Len 10, Fast 9 EMA, Slow 21 EMA, Min Gap 10 bars.
• Crypto swing: Trend filter on, Dip % 4.0, Vol Spike 2.0, Volatility Len 14, Fast 20 EMA, Slow 50 EMA, Trend 200 EMA.
Summary
Dip Hunter is a focused pullback engine. It quantifies a real dip off a recent high, validates it with volume and volatility expansion, enforces corrective structure with EMAs, and optionally restricts signals to an uptrend. The score, bar gradient, and meter make reading conditions instant. Entry lines and alerts turn that read into an executable plan. Tune the thresholds to your market and timeframe, then let the tool keep you patient in red, selective in white, and decisive in green.
Time-Decaying Percentile Oscillator [BackQuant]Time-Decaying Percentile Oscillator
1. Big-picture idea
Traditional percentile or stochastic oscillators treat every bar in the look-back window as equally important. That is fine when markets are slow, but if volatility regime changes quickly yesterday’s print should matter more than last month’s. The Time-Decaying Percentile Oscillator attempts to fix that blind spot by assigning an adjustable weight to every past price before it is ranked. The result is a percentile score that “breathes” with market tempo much faster to flag new extremes yet still smooth enough to ignore random noise.
2. What the script actually does
Build a weight curve
• You pick a look-back length (default 28 bars).
• You decide whether weights fall Linearly , Exponentially , by Power-law or Logarithmically .
• A decay factor (lower = faster fade) shapes how quickly the oldest price loses influence.
• The array is normalised so all weights still sum to 1.
Rank prices by weighted mass
• Every close in the window is paired with its weight.
• The pairs are sorted from low to high.
• The cumulative weight is walked until it equals your chosen percentile level (default 50 = median).
• That price becomes the Time-Decayed Percentile .
Find dispersion with robust statistics
• Instead of a fragile standard deviation the script measures weighted Median-Absolute-Deviation about the new percentile.
• You multiply that deviation by the Deviation Multiplier slider (default 1.0) to get a non-parametric volatility band.
Build an adaptive channel
• Upper band = percentile + (multiplier × deviation)
• Lower band = percentile – (multiplier × deviation)
Normalise into a 0-100 oscillator
• The current close is mapped inside that band:
0 = lower band, 50 = centre, 100 = upper band.
• If the channel squeezes, tiny moves still travel the full scale; if volatility explodes, it automatically widens.
Optional smoothing
• A second-stage moving average (EMA, SMA, DEMA, TEMA, etc.) tames the jitter.
• Length 22 EMA by default—change it to tune reaction speed.
Threshold logic
• Upper Threshold 70 and Lower Threshold 30 separate standard overbought/oversold states.
• Extreme bands 85 and 15 paint background heat when aggressive fade or breakout trades might trigger.
Divergence engine
• Looks back twenty bars.
• Flags Bullish divergence when price makes a lower low but oscillator refuses to confirm (value < 40).
• Flags Bearish divergence when price prints a higher high but oscillator stalls (value > 60).
3. Component walk-through
• Source – Any price series. Close by default, switch to typical price or custom OHLC4 for futures spreads.
• Look-back Period – How many bars to rank. Short = faster, long = slower.
• Base Percentile Level – 50 shows relative position around the median; set to 25 / 75 for quartile tracking or 90 / 10 for extreme tails.
• Deviation Multiplier – Higher values widen the dynamic channel, lowering whipsaw but delaying signals.
• Decay Settings
– Type decides the curve shape. Exponential (default 1.16) mimics EMA logic.
– Factor < 1 shrinks influence faster; > 1 spreads influence flatter.
– Toggle Enable Time Decay off to compare with classic equal-weight stochastic.
• Smoothing Block – Choose one of seven MA flavours plus length.
• Thresholds – Overbought / Oversold / Extreme levels. Push them out when working on very mean-reverting assets like FX; pull them in for trend monsters like crypto.
• Display toggles – Show or hide threshold lines, extreme filler zones, bar colouring, divergence labels.
• Colours – Bullish green, bearish red, neutral grey. Every gradient step is automatically blended to generate a heat map across the 0-100 range.
4. How to read the chart
• Oscillator creeping above 70 = market auctioning near the top of its adaptive range.
• Fast poke above 85 with no follow-through = exhaustion fade candidate.
• Slow grind that lives above 70 for many bars = valid bullish trend, not a fade.
• Cross back through 50 shows balance has shifted; treat it like a micro trend change.
• Divergence arrows add extra confidence when you already see two-bar reversal candles at range extremes.
• Background shading (semi-transparent red / green) warns of extreme states and throttles your position size.
5. Practical trading playbook
Mean-reversion scalps
1. Wait for oscillator to reach your desired OB/ OS levels
2. Check the slope of the smoothing MA—if it is flattening the squeeze is mature.
3. Look for a one- or two-bar reversal pattern.
4. Enter against the move; first target = midline 50, second target = opposite threshold.
5. Stop loss just beyond the extreme band.
Trend continuation pullbacks
1. Identify a clean directional trend on the price chart.
2. During the trend, TDP will oscillate between midline and extreme of that side.
3. Buy dips when oscillator hits OS levels, and the same for OB levels & shorting
4. Exit when oscillator re-tags the same-side extreme or prints divergence.
Volatility regime filter
• Use the Enable Time Decay switch as a regime test.
• If equal-weight oscillator and decayed oscillator diverge widely, market is entering a new volatility regime—tighten stops and trade smaller.
Divergence confirmation for other indicators
• Pair TDP divergence arrows with MACD histogram or RSI to filter false positives.
• The weighted nature means TDP often spots divergence a bar or two earlier than standard RSI.
Swing breakout strategy
1. During consolidation, band width compresses and oscillator oscillates around 50.
2. Watch for sudden expansion where oscillator blasts through extreme bands and stays pinned.
3. Enter with momentum in breakout direction; trail stop behind upper or lower band as it re-expands.
6. Customising decay mathematics
Linear – Each older bar loses the same fixed amount of influence. Intuitive and stable; good for slow swing charts.
Exponential – Influence halves every “decay factor” steps. Mirrors EMA thinking and is fastest to react.
Power-law – Mid-history bars keep more authority than exponential but oldest data still fades. Handy for commodities where seasonality matters.
Logarithmic – The gentlest curve; weight drops sharply at first then levels off. Mimics how traders remember dramatic moves for weeks but forget ordinary noise quickly.
Turn decay off to verify the tool’s added value; most users never switch back.
7. Alert catalogue
• TD Overbought / TD Oversold – Cross of regular thresholds.
• TD Extreme OB / OS – Breach of danger zones.
• TD Bullish / Bearish Divergence – High-probability reversal watch.
• TD Midline Cross – Momentum shift that often precedes a window where trend-following systems perform.
8. Visual hygiene tips
• If you already plot price on a dark background pick Bullish Color and Bearish Color default; change to pastel tones for light themes.
• Hide threshold lines after you memorise the zones to declutter scalping layouts.
• Overlay mode set to false so the oscillator lives in its own panel; keep height about 30 % of screen for best resolution.
9. Final notes
Time-Decaying Percentile Oscillator marries robust statistical ranking, adaptive dispersion and decay-aware weighting into a simple oscillator. It respects both recent order-flow shocks and historical context, offers granular control over responsiveness and ships with divergence and alert plumbing out of the box. Bolt it onto your price action framework, trend-following system or volatility mean-reversion playbook and see how much sooner it recognises genuine extremes compared to legacy oscillators.
Backtest thoroughly, experiment with decay curves on each asset class and remember: in trading, timing beats timidity but patience beats impulse. May this tool help you find that edge.
Recession Warning Model [BackQuant]Recession Warning Model
Overview
The Recession Warning Model (RWM) is a Pine Script® indicator designed to estimate the probability of an economic recession by integrating multiple macroeconomic, market sentiment, and labor market indicators. It combines over a dozen data series into a transparent, adaptive, and actionable tool for traders, portfolio managers, and researchers. The model provides customizable complexity levels, display modes, and data processing options to accommodate various analytical requirements while ensuring robustness through dynamic weighting and regime-aware adjustments.
Purpose
The RWM fulfills the need for a concise yet comprehensive tool to monitor recession risk. Unlike approaches relying on a single metric, such as yield-curve inversion, or extensive economic reports, it consolidates multiple data sources into a single probability output. The model identifies active indicators, their confidence levels, and the current economic regime, enabling users to anticipate downturns and adjust strategies accordingly.
Core Features
- Indicator Families : Incorporates 13 indicators across five categories: Yield, Labor, Sentiment, Production, and Financial Stress.
- Dynamic Weighting : Adjusts indicator weights based on recent predictive accuracy, constrained within user-defined boundaries.
- Leading and Coincident Split : Separates early-warning (leading) and confirmatory (coincident) signals, with adjustable weighting (default 60/40 mix).
- Economic Regime Sensitivity : Modulates output sensitivity based on market conditions (Expansion, Late-Cycle, Stress, Crisis), using a composite of VIX, yield-curve, financial conditions, and credit spreads.
- Display Options : Supports four modes—Probability (0-100%), Binary (four risk bins), Lead/Coincident, and Ensemble (blended probability).
- Confidence Intervals : Reflects model stability, widening during high volatility or conflicting signals.
- Alerts : Configurable thresholds (Watch, Caution, Warning, Alert) with persistence filters to minimize false signals.
- Data Export : Enables CSV output for probabilities, signals, and regimes, facilitating external analysis in Python or R.
Model Complexity Levels
Users can select from four tiers to balance simplicity and depth:
1. Essential : Focuses on three core indicators—yield-curve spread, jobless claims, and unemployment change—for minimalistic monitoring.
2. Standard : Expands to nine indicators, adding consumer confidence, PMI, VIX, S&P 500 trend, money supply vs. GDP, and the Sahm Rule.
3. Professional : Includes all 13 indicators, incorporating financial conditions, credit spreads, JOLTS vacancies, and wage growth.
4. Research : Unlocks all indicators plus experimental settings for advanced users.
Key Indicators
Below is a summary of the 13 indicators, their data sources, and economic significance:
- Yield-Curve Spread : Difference between 10-year and 3-month Treasury yields. Negative spreads signal banking sector stress.
- Jobless Claims : Four-week moving average of unemployment claims. Sustained increases indicate rising layoffs.
- Unemployment Change : Three-month change in unemployment rate. Sharp rises often precede recessions.
- Sahm Rule : Triggers when unemployment rises 0.5% above its 12-month low, a reliable recession indicator.
- Consumer Confidence : University of Michigan survey. Declines reflect household pessimism, impacting spending.
- PMI : Purchasing Managers’ Index. Values below 50 indicate manufacturing contraction.
- VIX : CBOE Volatility Index. Elevated levels suggest market anticipation of economic distress.
- S&P 500 Growth : Weekly moving average trend. Declines reduce wealth effects, curbing consumption.
- M2 + GDP Trend : Monitors money supply and real GDP. Simultaneous declines signal credit contraction.
- NFCI : Chicago Fed’s National Financial Conditions Index. Positive values indicate tighter conditions.
- Credit Spreads : Proxy for corporate bond spreads using 10-year vs. 2-year Treasury yields. Widening spreads reflect stress.
- JOLTS Vacancies : Job openings data. Significant drops precede hiring slowdowns.
- Wage Growth : Year-over-year change in average hourly earnings. Late-cycle spikes often signal economic overheating.
Data Processing
- Rate of Change (ROC) : Optionally applied to capture momentum in data series (default: 21-bar period).
- Z-Score Normalization : Standardizes indicators to a common scale (default: 252-bar lookback).
- Smoothing : Applies a short moving average to final signals (default: 5-bar period) to reduce noise.
- Binary Signals : Generated for each indicator (e.g., yield-curve inverted or PMI below 50) based on thresholds or Z-score deviations.
Probability Calculation
1. Each indicator’s binary signal is weighted according to user settings or dynamic performance.
2. Weights are normalized to sum to 100% across active indicators.
3. Leading and coincident signals are aggregated separately (if split mode is enabled) and combined using the specified mix.
4. The probability is adjusted by a regime multiplier, amplifying risk during Stress or Crisis regimes.
5. Optional smoothing ensures stable outputs.
Display and Visualization
- Probability Mode : Plots a continuous 0-100% recession probability with color gradients and confidence bands.
- Binary Mode : Categorizes risk into four levels (Minimal, Watch, Caution, Alert) for simplified dashboards.
- Lead/Coincident Mode : Displays leading and coincident probabilities separately to track signal divergence.
- Ensemble Mode : Averages traditional and split probabilities for a balanced view.
- Regime Background : Color-coded overlays (green for Expansion, orange for Late-Cycle, amber for Stress, red for Crisis).
- Analytics Table : Optional dashboard showing probability, confidence, regime, and top indicator statuses.
Practical Applications
- Asset Allocation : Adjust equity or bond exposures based on sustained probability increases.
- Risk Management : Hedge portfolios with VIX futures or options during regime shifts to Stress or Crisis.
- Sector Rotation : Shift toward defensive sectors when coincident signals rise above 50%.
- Trading Filters : Disable short-term strategies during high-risk regimes.
- Event Timing : Scale positions ahead of high-impact data releases when probability and VIX are elevated.
Configuration Guidelines
- Enable ROC and Z-score for consistent indicator comparison unless raw data is preferred.
- Use dynamic weighting with at least one economic cycle of data for optimal performance.
- Monitor stress composite scores above 80 alongside probabilities above 70 for critical risk signals.
- Adjust adaptation speed (default: 0.1) to 0.2 during Crisis regimes for faster indicator prioritization.
- Combine RWM with complementary tools (e.g., liquidity metrics) for intraday or short-term trading.
Limitations
- Macro indicators lag intraday market moves, making RWM better suited for strategic rather than tactical trading.
- Historical data availability may constrain dynamic weighting on shorter timeframes.
- Model accuracy depends on the quality and timeliness of economic data feeds.
Final Note
The Recession Warning Model provides a disciplined framework for monitoring economic downturn risks. By integrating diverse indicators with transparent weighting and regime-aware adjustments, it empowers users to make informed decisions in portfolio management, risk hedging, or macroeconomic research. Regular review of model outputs alongside market-specific tools ensures its effective application across varying market conditions.
FEDFUNDS Rate Divergence Oscillator [BackQuant]FEDFUNDS Rate Divergence Oscillator
1. Concept and Rationale
The United States Federal Funds Rate is the anchor around which global dollar liquidity and risk-free yield expectations revolve. When the Fed hikes, borrowing costs rise, liquidity tightens and most risk assets encounter head-winds. When it cuts, liquidity expands, speculative appetite often recovers. Bitcoin, a 24-hour permissionless asset sometimes described as “digital gold with venture-capital-like convexity,” is particularly sensitive to macro-liquidity swings.
The FED Divergence Oscillator quantifies the behavioural gap between short-term monetary policy (proxied by the effective Fed Funds Rate) and Bitcoin’s own percentage price change. By converting each series into identical rate-of-change units, subtracting them, then optionally smoothing the result, the script produces a single bounded-yet-dynamic line that tells you, at a glance, whether Bitcoin is outperforming or underperforming the policy backdrop—and by how much.
2. Data Pipeline
• Fed Funds Rate – Pulled directly from the FRED database via the ticker “FRED:FEDFUNDS,” sampled at daily frequency to synchronise with crypto closes.
• Bitcoin Price – By default the script forces a daily timeframe so that both series share time alignment, although you can disable that and plot the oscillator on intraday charts if you prefer.
• User Source Flexibility – The BTC series is not hard-wired; you can select any exchange-specific symbol or even swap BTC for another crypto or risk asset whose interaction with the Fed rate you wish to study.
3. Math under the Hood
(1) Rate of Change (ROC) – Both the Fed rate and BTC close are converted to percent return over a user-chosen lookback (default 30 bars). This means a cut from 5.25 percent to 5.00 percent feeds in as –4.76 percent, while a climb from 25 000 to 30 000 USD in BTC over the same window converts to +20 percent.
(2) Divergence Construction – The script subtracts the Fed ROC from the BTC ROC. Positive values show BTC appreciating faster than policy is tightening (or falling slower than the rate is cutting); negative values show the opposite.
(3) Optional Smoothing – Macro series are noisy. Toggle “Apply Smoothing” to calm the line with your preferred moving-average flavour: SMA, EMA, DEMA, TEMA, RMA, WMA or Hull. The default EMA-25 removes day-to-day whips while keeping turning points alive.
(4) Dynamic Colour Mapping – Rather than using a single hue, the oscillator line employs a gradient where deep greens represent strong bullish divergence and dark reds flag sharp bearish divergence. This heat-map approach lets you gauge intensity without squinting at numbers.
(5) Threshold Grid – Five horizontal guides create a structured regime map:
• Lower Extreme (–50 pct) and Upper Extreme (+50 pct) identify panic capitulations and euphoria blow-offs.
• Oversold (–20 pct) and Overbought (+20 pct) act as early warning alarms.
• Zero Line demarcates neutral alignment.
4. Chart Furniture and User Interface
• Oscillator fill with a secondary DEMA-30 “shader” offers depth perception: fat ribbons often precede high-volatility macro shifts.
• Optional bar-colouring paints candles green when the oscillator is above zero and red below, handy for visual correlation.
• Background tints when the line breaches extreme zones, making macro inflection weeks pop out in the replay bar.
• Everything—line width, thresholds, colours—can be customised so the indicator blends into any template.
5. Interpretation Guide
Macro Liquidity Pulse
• When the oscillator spends weeks above +20 while the Fed is still raising rates, Bitcoin is signalling liquidity tolerance or an anticipatory pivot view. That condition often marks the embryonic phase of major bull cycles (e.g., March 2020 rebound).
• Sustained prints below –20 while the Fed is already dovish indicate risk aversion or idiosyncratic crypto stress—think exchange scandals or broad flight to safety.
Regime Transition Signals
• Bullish cross through zero after a long sub-zero stint shows Bitcoin regaining upward escape velocity versus policy.
• Bearish cross under zero during a hiking cycle tells you monetary tightening has finally started to bite.
Momentum Exhaustion and Mean-Reversion
• Touches of +50 (or –50) come rarely; they are statistically stretched events. Fade strategies either taking profits or hedging have historically enjoyed positive expectancy.
• Inside-bar candlestick patterns or lower-timeframe bearish engulfings simultaneously with an extreme overbought print make high-probability short scalp setups, especially near weekly resistance. The same logic mirrors for oversold.
Pair Trading / Relative Value
• Combine the oscillator with spreads like BTC versus Nasdaq 100. When both the FED Divergence oscillator and the BTC–NDQ relative-strength line roll south together, the cross-asset confirmation amplifies conviction in a mean-reversion short.
• Swap BTC for miners, altcoins or high-beta equities to test who is the divergence leader.
Event-Driven Tactics
• FOMC days: plot the oscillator on an hourly chart (disable ‘Force Daily TF’). Watch for micro-structural spikes that resolve in the first hour after the statement; rapid flips across zero can front-run post-FOMC swings.
• CPI and NFP prints: extremes reached into the release often mean positioning is one-sided. A reversion toward neutral in the first 24 hours is common.
6. Alerts Suite
Pre-bundled conditions let you automate workflows:
• Bullish / Bearish zero crosses – queue spot or futures entries.
• Standard OB / OS – notify for first contact with actionable zones.
• Extreme OB / OS – prime time to review hedges, take profits or build contrarian swing positions.
7. Parameter Playground
• Shorten ROC Lookback to 14 for tactical traders; lengthen to 90 for macro investors.
• Raise extreme thresholds (for example ±80) when plotting on altcoins that exhibit higher volatility than BTC.
• Try HMA smoothing for responsive yet smooth curves on intraday charts.
• Colour-blind users can easily swap bull and bear palette selections for preferred contrasts.
8. Limitations and Best Practices
• The Fed Funds series is step-wise; it only changes on meeting days. Rapid BTC oscillations in between may dominate the calculation. Keep that perspective when interpreting very high-frequency signals.
• Divergence does not equal causation. Crypto-native catalysts (ETF approvals, hack headlines) can overwhelm macro links temporarily.
• Use in conjunction with classical confirmation tools—order-flow footprints, market-profile ledges, or simple price action to avoid “pure-indicator” traps.
9. Final Thoughts
The FEDFUNDS Rate Divergence Oscillator distills an entire macro narrative monetary policy versus risk sentiment into a single colourful heartbeat. It will not magically predict every pivot, yet it excels at framing market context, spotting stretches and timing regime changes. Treat it as a strategic compass rather than a tactical sniper scope, combine it with sound risk management and multi-factor confirmation, and you will possess a robust edge anchored in the world’s most influential interest-rate benchmark.
Trade consciously, stay adaptive, and let the policy-price tension guide your roadmap.
Price Exhaustion Envelope [BackQuant]Price Exhaustion Envelope
Visual preview of the bands:
What it is
The Price Exhaustion Envelope (PEE) is a multi‑factor overextension detector wrapped inside a dynamic envelope framework. It measures how “tired” a move is by blending price stretch, volume surges, momentum and acceleration, plus optional RSI divergence. The result is a composite exhaustion score that drives both on‑chart signals and the adaptive width of three optional envelope bands around a smoothed baseline. When the score spikes above or below your chosen threshold, the script can flag exhaustion, paint candles, tint the background and fire alerts.
How it works under the hood
Exhaustion score
Price component: distance of close from its mean in standard deviation units.
Volume component: normalized volume pressure that highlights unusual participation.
Momentum component: rate of change and acceleration of price, scaled by their own volatility.
RSI divergence (optional): bullish and bearish divergences gently push the score lower or higher.
Mode control: choose Price, Volume, Momentum or Composite. Composite averages the main pieces for a balanced view.
Energy scale (0 to 100)
The composite score is pushed through a logistic transform to create an “energy” value. High energy (above 70 to 80) signals a move that may be running hot, while very low energy (below 20 to 30) points to exhaustion on the downside.
Envelope engine
Baseline: EMA of price over the main lookback length.
Width: base width is standard deviation times a multiplier.
Type selector:
• Static keeps the width fixed.
• Dynamic expands width in proportion to the absolute exhaustion score.
• Adaptive links width to the energy reading so bands breathe with market “heat.”
Smoothing: a short EMA on the width reduces jitter and keeps bands pleasant to trade around.
Band architecture
You can toggle up to three symmetric bands on each side of the baseline. They default to 1.0, 1.6 and 2.2 multiples of the smoothed width. Soft transparent fills create a layered thermograph of extension. The outermost band often maps to true blow‑off extremes.
On‑chart elements
Baseline line that flips color in real time depending on where price sits.
Up to three upper and lower bands with progressive opacity.
Triangle markers at fresh exhaustion triggers.
Tiny warning glyphs at extreme upper or lower breaches.
Optional bar coloring to visually tag exhausted candles.
Background halo when energy > 80 or < 20 for instant context.
A compact info table showing State, Score, Energy, Momentum score and where price sits inside the envelope (percent).
How to use it in trading
Mean reversion plays
When price pierces the outer band and an exhaustion marker prints, look for reversal candles or lower‑timeframe confirmation to fade the move back toward the baseline.
For conservative entries, wait for the composite score to roll back under the threshold or for energy to drop from extreme to neutral.
Set stops just beyond the extreme levels (use extreme_upper and extreme_lower as natural invalidation points). Targets can be the baseline or the opposite inner band.
Trend continuation with smart pullbacks
In strong trends, the first tag of Band 1 or Band 2 against the dominant direction often offers low‑risk continuation entries. Use energy readings: if energy is low on a pullback during an uptrend, a bounce is more likely.
Combine with RSI divergence: hidden bullish divergence near a lower band in an uptrend can be a powerful confirmation.
Breakout filtering
A breakout that occurs while the composite score is still moderate (not exhausted) has a higher chance of follow‑through. Skip signals when energy is already above 80 and price is punching the outer band, as the move may be late.
Watch env_position (Envelope %) in the table. Breakouts near 40 to 60 percent of the envelope are “healthy,” while those at 95 percent are stretched.
Scaling out and risk control
Use exhaustion alerts to trim positions into strength or weakness.
Trail stops just outside Band 2 or Band 3 to stay in trends while letting the envelope expand in volatile phases.
Multi‑timeframe confluence
Run the script on a higher timeframe to locate exhaustion context, then drill down to a lower timeframe for entries.
Opposite signals across timeframes (daily exhaustion vs. 5‑minute breakout) warn you to reduce size or tighten management.
Key inputs to experiment with
Lookback Period: larger values smooth the score and envelope, ideal for swing trading. Shorter values make it reactive for scalps.
Exhaustion Threshold: raise above 2.0 in choppy assets to cut noise, drop to 1.5 for smooth FX pairs.
Envelope Type: Dynamic is great for crypto spikes, Adaptive shines in stocks where volume and volatility wave together.
RSI Divergence: turn off if you prefer a pure price/volume model or if divergence floods the score in your asset.
Alert set included
Fresh upper exhaustion
Fresh lower exhaustion
Extreme upper breach
Extreme lower breach
RSI bearish divergence
RSI bullish divergence
Hook these to TradingView notifications so you get pinged the moment a move hits exhaustion.
Best practices
Always pair exhaustion signals with structure. Support and resistance, liquidity pools and session opens matter.
Avoid blindly shorting every upper signal in a roaring bull market. Let the envelope type help you filter.
Use the table to sanity‑check: a very high score but mid‑range env_position means the band may still be wide enough to absorb more movement.
Backtest threshold combinations on your instrument. Different tickers carry different volatility fingerprints.
Final note
Price Exhaustion Envelope is a flexible framework, not a turnkey system. It excels as a context layer that tells you when the crowd is pressing too hard or when a move still has fuel. Combine it with sound execution tactics, risk limits and market awareness. Trade safe and let the envelope breathe with the market.
Fibonacci Sequence Moving Average [BackQuant]Fibonacci Sequence Moving Average with Adaptive Oscillator
1. Overview
The Fibonacci Sequence Moving Average indicator is a two‑part trading framework that combines a custom moving average built from the famous Fibonacci number set with a fully featured oscillator, normalisation engine and divergence suite. The moving average half delivers an adaptive trend line that respects natural market rhythms, while the oscillator half translates that trend information into a bounded momentum stream that is easy to read, easy to compare across assets and rich in confluence signals. Everything from weighting logic to colour palettes can be customised, so the tool comfortably fits scalpers zooming into one‑minute candles as well as position traders running multi‑month trend following campaigns.
2. Core Calculation
Fibonacci periods – The default length array is 5, 8, 13, 21, 34. A single multiplier input lets you scale the whole family up or down without breaking the golden‑ratio spacing. For example a multiplier of 3 yields 15, 24, 39, 63, 102.
Component averages – Each period is passed through Simple Moving Average logic to produce five baseline curves (ma1 through ma5).
Weighting methods – You decide how those five values are blended:
• Equal weighting treats every curve the same.
• Linear weighting applies factors 1‑to‑5 so the slowest curve counts five times as much as the fastest.
• Exponential weighting doubles each step for a fast‑reacting yet still smooth line.
• Fibonacci weighting multiplies each curve by its own period value, honouring the spirit of ratio mathematics.
Smoothing engine – The blended average is then smoothed a second time with your choice of SMA, EMA, DEMA, TEMA, RMA, WMA or HMA. A short smoothing length keeps the result lively, while longer lengths create institution‑grade glide paths that act like dynamic support and resistance.
3. Oscillator Construction
Once the smoothed Fib MA is in place, the script generates a raw oscillator value in one of three flavours:
• Distance – Percentage distance between price and the average. Great for mean‑reversion.
• Momentum – Percentage change of the average itself. Ideal for trend acceleration studies.
• Relative – Distance divided by Average True Range for volatility‑aware scaling.
That raw series is pushed through a look‑back normaliser that rescales every reading into a fixed −100 to +100 window. The normalisation window defaults to 100 bars but can be tightened for fast markets or expanded to capture long regimes.
4. Visual Layer
The oscillator line is gradient‑coloured from deep red through sky blue into bright green, so you can spot subtle momentum shifts with peripheral vision alone. There are four horizontal guide lines: Extreme Bear at −50, Bear Threshold at −20, Bull Threshold at +20 and Extreme Bull at +50. Soft fills above and below the thresholds reinforce the zones without cluttering the chart.
The smoothed Fib MA can be plotted directly on price for immediate trend context, and each of the five component averages can be revealed for educational or research purposes. Optional bar‑painting mirrors oscillator polarity, tinting candles green when momentum is bullish and red when momentum is bearish.
5. Divergence Detection
The script automatically looks for four classes of divergences between price pivots and oscillator pivots:
Regular Bullish, signalling a possible bottom when price prints a lower low but the oscillator prints a higher low.
Hidden Bullish, often a trend‑continuation cue when price makes a higher low while the oscillator slips to a lower low.
Regular Bearish, marking potential tops when price carves a higher high yet the oscillator steps down.
Hidden Bearish, hinting at ongoing downside when price posts a lower high while the oscillator pushes to a higher high.
Each event is tagged with an ℝ or ℍ label at the oscillator pivot, colour‑coded for clarity. Look‑back distances for left and right pivots are fully adjustable so you can fine‑tune sensitivity.
6. Alerts
Five ready‑to‑use alert conditions are included:
• Bullish when the oscillator crosses above +20.
• Bearish when it crosses below −20.
• Extreme Bullish when it pops above +50.
• Extreme Bearish when it dives below −50.
• Zero Cross for momentum inflection.
Attach any of these to TradingView notifications and stay updated without staring at charts.
7. Practical Applications
Swing trading trend filter – Plot the smoothed Fib MA on daily candles and only trade in its direction. Enter on oscillator retracements to the 0 line.
Intraday reversal scouting – On short‑term charts let Distance mode highlight overshoots beyond ±40, then fade those moves back to mean.
Volatility breakout timing – Use Relative mode during earnings season or crypto news cycles to spot momentum surges that adjust for changing ATR.
Divergence confirmation – Layer the oscillator beneath price structure to validate double bottoms, double tops and head‑and‑shoulders patterns.
8. Input Summary
• Source, Fibonacci multiplier, weighting method, smoothing length and type
• Oscillator calculation mode and normalisation look‑back
• Divergence look‑back settings and signal length
• Show or hide options for every visual element
• Full colour and line width customisation
9. Best Practices
Avoid using tiny multipliers on illiquid assets where the shortest Fibonacci window may drop under three bars. In strong trends reduce divergence sensitivity or you may see false counter‑trend flags. For portfolio scanning set oscillator to Momentum mode, hide thresholds and colour bars only, which turns the indicator into a heat‑map that quickly highlights leaders and laggards.
10. Final Notes
The Fibonacci Sequence Moving Average indicator seeks to fuse the mathematical elegance of the golden ratio with modern signal‑processing techniques. It is not a standalone trading system, rather a multi‑purpose information layer that shines when combined with market structure, volume analysis and disciplined risk management. Always test parameters on historical data, be mindful of slippage and remember that past performance is never a guarantee of future results. Trade wisely and enjoy the harmony of Fibonacci mathematics in your technical toolkit.
Weighted Multi-Mode Oscillator [BackQuant]Weighted Multi‑Mode Oscillator
1. What Is It?
The Weighted Multi‑Mode Oscillator (WMMO) is a next‑generation momentum tool that turns a dynamically‑weighted moving average into a 0‑100 bounded oscillator.
It lets you decide how each bar is weighted (by volume, volatility, momentum or a hybrid blend) and how the result is normalised (Percentile, Z‑Score or Min‑Max).
The outcome is a self‑adapting gauge that delivers crystal‑clear overbought / oversold zones, divergence clues and regime shifts on any market or timeframe.
2. How It Works
• Dynamic Weight Engine
▪ Volume – emphasises bars with exceptional participation.
▪ Volatility – inverse ATR weighting filters noisy spikes.
▪ Momentum – amplifies strong directional ROC bursts.
▪ Hybrid – equal‑weight blend of the three dimensions.
• Multi‑Mode Smoothing
Choose from 8 MA types (EMA, DEMA, HMA, LINREG, TEMA, RMA, SMA, WMA) plus a secondary smoothing factor to fine‑tune lag vs. responsiveness.
• Normalization Suite
▪ Percentile – rank vs. recent history (context aware).
▪ Z‑Score – standard deviations from mean (statistical extremes).
▪ Min‑Max – scale between rolling high/low (trend friendly).
3. Reading the Oscillator
Zone Default Level Interpretation
Bull > 80 Acceleration; momentum buyers in control
Neutral 20 – 80 Consolidation / no edge
Bear < 20 Exhaustion; sellers dominate
Gradient line/area automatically shades from bright green (strong bull) to deep red (strong bear).
Optional bar‑painting colours price bars the same way for rapid chart scanning.
4. Typical Use‑Cases
Trend Confirmation – Set Weight = Hybrid, Smoothing = EMA. Enter pullbacks only when WMMO > 50 and rising.
Mean Reversion – Weight = Volatility, reduce upper / lower bands to 70 / 30 and fade extremes.
Volume Pulse – Intraday futures: Weight = Volume to catch participation surges before breakout candles.
Divergence Spotting – Compare price highs/lows to WMMO peaks for early reversal clues.
5. Inputs & Styling
Calculation: Source, MA Length, MA Type, Smoothing
Weighting: Volume period & factor, Volatility length, Momentum period
Normalisation: Method, Look‑back, Upper / Lower thresholds
Display: Gradient fills, Threshold lines, Bar‑colouring toggle, Line width & colours
All thresholds, colours and fills are fully customisable inside the settings panel.
6. Built‑In Alerts
WMMO Long – oscillator crosses up through upper threshold.
WMMO Short – oscillator crosses down through lower threshold.
Attach them once and receive push / e‑mail notifications the moment momentum flips.
7. Best Practices
Percentile mode is self‑adaptive and works well across assets; Z‑Score excels in ranges; Min‑Max shines in persistent trends.
Very short MA lengths (< 10) may produce jitter; compensate with higher “Smoothing” or longer look‑backs.
Pair WMMO with structure‑based tools (S/R, trend lines) for higher‑probability trade confluence.
Disclaimer
This script is provided for educational purposes only. It is not financial advice. Always back‑test thoroughly and manage risk before trading live capital.
Kase Convergence Divergence [BackQuant]Kase Convergence Divergence
The Kase Convergence Divergence is a sophisticated oscillator designed to measure directional market strength through the lens of volatility-adjusted log return structures. Inspired by Cynthia Kase’s work on statistical momentum and price projection ranges, this unique indicator offers a hybrid framework that merges signal processing, multi-length sweep logic, and adaptive smoothing techniques.
Unlike traditional momentum oscillators like MACD or RSI, which rely on static moving average differences, KCD introduces a dual-process system combining:
Kase-style statistical range projection (via log returns and volatility),
A sweeping loop of lookback lengths for robustness,
First and second derivative modes to capture both velocity and acceleration of price movement.
Core Logic & Computation
The KCD calculation is centered on two volatility-normalized transforms:
KSDI Up: Measures how far the current high has moved relative to a past low, normalized by return volatility.
KSDI Down: Measures how far the current low has moved relative to a past high, also normalized.
For every length in a user-defined sweep range (e.g., 25–35), both KSDI_up and KSDI_dn are computed, and their maximum values across the loop are retained. The difference between these two max values produces the raw signal:
KPO (Kase Projection Oscillator): Measures directional skew.
KCD (Kase Convergence Divergence): Defined as KPO – MA(KPO) — similar in spirit to MACD but structurally different.
Users can choose to visualize either the first derivative (KPO) , or the second derivative (KCD) , depending on market conditions or strategy style.
Key Features
✅ Multi-Length Sweep Logic: Improves signal reliability by aggregating statistical range projections across a set of lookbacks.
✅ Advanced Smoothing Modes: Supports DEMA, HMA, TEMA, LINREG, WMA and more for dynamic adaptation.
✅ Dual Derivative Modes: Choose between speed (first derivative) or smoothness (second derivative) to fit your trading regime.
✅ Color-Encoded Signal Bands: Heatmap-style oscillator coloring enhances visual feedback on trend strength.
✅ Candlestick Painting: Optional bar coloring makes it easy to spot trend shifts on the main chart.
✅ Adaptive Fill Zones: Green and red fills between the oscillator and zero line help distinguish bullish and bearish regimes at a glance.
Practical Applications
📈 Trend Confirmation: Use KCD as a secondary confirmation layer after breakout or pullback entries.
📉 Momentum Shifts: Crossover and crossunder of the zero line highlight potential regime changes.
📊 Strategy Filters: Incorporate into algos to avoid trendless or mean-reverting environments.
🧪 Derivative Switching: Flip between KPO and KCD modes depending on whether you want to measure acceleration or deceleration of price flow.
Alerts & Signals
Two built-in alerts help you catch regime shifts in real time:
Long Signal: Triggered when the selected oscillator crosses above zero.
Short Signal: Triggered when it crosses below zero.
These events can be used to generate entries, exits, or trend validation cues in multi-layer systems.
Conclusion
The Kase Convergence Divergence goes beyond traditional oscillators by offering a volatility-normalized, derivative-aware signal engine with enhanced visual dynamics. Its sweeping architecture and dynamic fill logic make it especially powerful for identifying trending environments, filtering chop, and adding statistical rigor to your trading toolkit.
Whether you’re a discretionary trader seeking precision, or a quant looking to model more robust return structures, KCD offers a creative yet analytically grounded solution.
Momentum Regression [BackQuant]Momentum Regression
The Momentum Regression is an advanced statistical indicator built to empower quants, strategists, and technically inclined traders with a robust visual and quantitative framework for analyzing momentum effects in financial markets. Unlike traditional momentum indicators that rely on raw price movements or moving averages, this tool leverages a volatility-adjusted linear regression model (y ~ x) to uncover and validate momentum behavior over a user-defined lookback window.
Purpose & Design Philosophy
Momentum is a core anomaly in quantitative finance — an effect where assets that have performed well (or poorly) continue to do so over short to medium-term horizons. However, this effect can be noisy, regime-dependent, and sometimes spurious.
The Momentum Regression is designed as a pre-strategy analytical tool to help you filter and verify whether statistically meaningful and tradable momentum exists in a given asset. Its architecture includes:
Volatility normalization to account for differences in scale and distribution.
Regression analysis to model the relationship between past and present standardized returns.
Deviation bands to highlight overbought/oversold zones around the predicted trendline.
Statistical summary tables to assess the reliability of the detected momentum.
Core Concepts and Calculations
The model uses the following:
Independent variable (x): The volatility-adjusted return over the chosen momentum period.
Dependent variable (y): The 1-bar lagged log return, also adjusted for volatility.
A simple linear regression is performed over a large lookback window (default: 1000 bars), which reveals the slope and intercept of the momentum line. These values are then used to construct:
A predicted momentum trendline across time.
Upper and lower deviation bands , representing ±n standard deviations of the regression residuals (errors).
These visual elements help traders judge how far current returns deviate from the modeled momentum trend, similar to Bollinger Bands but derived from a regression model rather than a moving average.
Key Metrics Provided
On each update, the indicator dynamically displays:
Momentum Slope (β₁): Indicates trend direction and strength. A higher absolute value implies a stronger effect.
Intercept (β₀): The predicted return when x = 0.
Pearson’s R: Correlation coefficient between x and y.
R² (Coefficient of Determination): Indicates how well the regression line explains the variance in y.
Standard Error of Residuals: Measures dispersion around the trendline.
t-Statistic of β₁: Used to evaluate statistical significance of the momentum slope.
These statistics are presented in a top-right summary table for immediate interpretation. A bottom-right signal table also summarizes key takeaways with visual indicators.
Features and Inputs
✅ Volatility-Adjusted Momentum : Reduces distortions from noisy price spikes.
✅ Custom Lookback Control : Set the number of bars to analyze regression.
✅ Extendable Trendlines : For continuous visualization into the future.
✅ Deviation Bands : Optional ±σ multipliers to detect abnormal price action.
✅ Contextual Tables : Help determine strength, direction, and significance of momentum.
✅ Separate Pane Design : Cleanly isolates statistical momentum from price chart.
How It Helps Traders
📉 Quantitative Strategy Validation:
Use the regression results to confirm whether a momentum-based strategy is worth pursuing on a specific asset or timeframe.
🔍 Regime Detection:
Track when momentum breaks down or reverses. Slope changes, drops in R², or weak t-stats can signal regime shifts.
📊 Trade Filtering:
Avoid false positives by entering trades only when momentum is both statistically significant and directionally favorable.
📈 Backtest Preparation:
Before running costly simulations, use this tool to pre-screen assets for exploitable return structures.
When to Use It
Before building or deploying a momentum strategy : Test if momentum exists and is statistically reliable.
During market transitions : Detect early signs of fading strength or reversal.
As part of an edge-stacking framework : Combine with other filters such as volatility compression, volume surges, or macro filters.
Conclusion
The Momentum Regression indicator offers a powerful fusion of statistical analysis and visual interpretation. By combining volatility-adjusted returns with real-time linear regression modeling, it helps quantify and qualify one of the most studied and traded anomalies in finance: momentum.
Rolling Log Returns [BackQuant]Rolling Log Returns
The Rolling Log Returns indicator is a versatile tool designed to help traders, quants, and data-driven analysts evaluate the dynamics of price changes using logarithmic return analysis. Widely adopted in quantitative finance, log returns offer several mathematical and statistical advantages over simple returns, making them ideal for backtesting, portfolio optimization, volatility modeling, and risk management.
What Are Log Returns?
In quantitative finance, logarithmic returns are defined as:
ln(Pₜ / Pₜ₋₁)
or for rolling periods:
ln(Pₜ / Pₜ₋ₙ)
where P represents price and n is the rolling lookback window.
Log returns are preferred because:
They are time additive : returns over multiple periods can be summed.
They allow for easier statistical modeling , especially when assuming normally distributed returns.
They behave symmetrically for gains and losses, unlike arithmetic returns.
They normalize percentage changes, making cross-asset or cross-timeframe comparisons more consistent.
Indicator Overview
The Rolling Log Returns indicator computes log returns either on a standard (1-period) basis or using a rolling lookback period , allowing users to adapt it to short-term trading or long-term trend analysis.
It also supports a comparison series , enabling traders to compare the return structure of the main charted asset to another instrument (e.g., SPY, BTC, etc.).
Core Features
✅ Return Modes :
Normal Log Returns : Measures ln(price / price ), ideal for day-to-day return analysis.
Rolling Log Returns : Measures ln(price / price ), highlighting price drift over longer horizons.
✅ Comparison Support :
Compare log returns of the primary instrument to another symbol (like an index or ETF).
Useful for relative performance and market regime analysis .
✅ Moving Averages of Returns :
Smooth noisy return series with customizable MA types: SMA, EMA, WMA, RMA, and Linear Regression.
Applicable to both primary and comparison series.
✅ Conditional Coloring :
Returns > 0 are colored green ; returns < 0 are red .
Comparison series gets its own unique color scheme.
✅ Extreme Return Detection :
Highlight unusually large price moves using upper/lower thresholds.
Visually flags abnormal volatility events such as earnings surprises or macroeconomic shocks.
Quantitative Use Cases
🔍 Return Distribution Analysis :
Gain insight into the statistical properties of asset returns (e.g., skewness, kurtosis, tail behavior).
📉 Risk Management :
Use historical return outliers to define drawdown expectations, stress tests, or VaR simulations.
🔁 Strategy Backtesting :
Apply rolling log returns to momentum or mean-reversion models where compounding and consistent scaling matter.
📊 Market Regime Detection :
Identify periods of consistent overperformance/underperformance relative to a benchmark asset.
📈 Signal Engineering :
Incorporate return deltas, moving average crossover of returns, or threshold-based triggers into machine learning pipelines or rule-based systems.
Recommended Settings
Use Normal mode for high-frequency trading signals.
Use Rolling mode for swing or trend-following strategies.
Compare vs. a broad market index (e.g., SPY or QQQ ) to extract relative strength insights.
Set upper and lower thresholds around ±5% for spotting major volatility days.
Conclusion
The Rolling Log Returns indicator transforms raw price action into a statistically sound return series—equipping traders with a professional-grade lens into market behavior. Whether you're conducting exploratory data analysis, building factor models, or visually scanning for outliers, this indicator integrates seamlessly into a modern quant's toolbox.
Wavelet Filter with Adaptive Upsampling [BackQuant]Wavelet Filter with Adaptive Upsampling
The Wavelet Filter with Adaptive Upsampling is an advanced filtering and signal reconstruction tool designed to enhance the analysis of financial time series data. It combines wavelet transforms with adaptive upsampling techniques to filter and reconstruct price data, making it ideal for capturing subtle market movements and enhancing trend detection. This system uses high-pass and low-pass filters to decompose the price series into different frequency components, applying adaptive thresholding to eliminate noise and preserve relevant signal information.
Shout out to Loxx for the Least Squares fitting of trigonometric series and Quinn and Fernandes algorithm for finding frequency
www.tradingview.com
Key Features
1. Frequency Decomposition with High-Pass and Low-Pass Filters:
The indicator decomposes the input time series using high-pass and low-pass filters to separate the high-frequency (detail) and low-frequency (trend) components of the data. This decomposition allows for a more accurate analysis of underlying trends, while mitigating the impact of noise.
2. Soft Thresholding for Noise Reduction:
A soft thresholding function is applied to the high-frequency component, allowing for the reduction of noise while retaining significant market signals. This function adjusts the coefficients of the high-frequency data, removing small fluctuations and leaving only the essential price movements.
3. Adaptive Upsampling Process:
The upsampling process in this script can be customized using different methods: sinusoidal upsampling, advanced upsampling, and simple upsampling. Each method serves a unique purpose:
Sinusoidal Upsample uses a sine wave to interpolate between data points, providing a smooth transition.
Advanced Upsample utilizes a Quinn-Fernandes algorithm to estimate frequency and apply more sophisticated interpolation techniques, adapting to the market’s cyclical behavior.
Simple Upsample linearly interpolates between data points, providing a basic upsampling technique for less complex analysis.
4. Reconstruction of Filtered Signal:
The indicator reconstructs the filtered signal by summing the high and low-frequency components after upsampling. This allows for a detailed yet smooth representation of the original time series, which can be used for analyzing underlying trends in the market.
5. Visualization of Reconstructed Data:
The reconstructed series is plotted, showing how the upsampling and filtering process enhances the clarity of the price movements. Additionally, the script provides the option to visualize the log returns of the reconstructed series as a histogram, with positive returns shown in green and negative returns in red.
6. Cumulative Series and Trend Detection:
A cumulative series is plotted to visualize the compounded effect of the filtered and reconstructed data. This feature helps traders track the overall performance of the asset over time, identifying whether the asset is following a sustained upward or downward trend.
7. Adaptive Thresholding and Noise Estimation:
The system estimates the noise level in the high-frequency component and applies an adaptive thresholding process based on the standard deviation of the downsampled data. This ensures that only significant price movements are retained, further refining the trend analysis.
8. Customizable Parameters for Flexibility:
Users can customize the following parameters to adjust the behavior of the indicator:
Frequency and Phase Shift: Control the periodicity of the wavelet transformation and the phase of the upsampling function.
Upsample Factor: Adjust the level of interpolation applied during the upsampling process.
Smoothing Period: Determine the length of time used to smooth the signal, helping to filter out short-term fluctuations.
References
Enhancing Cross-Sectional Currency Strategies with Context-Aware Learning to Rank
arxiv.org
Daubechies Wavelet - Wikipedia
en.wikipedia.org
Quinn Fernandes Fourier Transform of Filtered Price by Loxx
Note on Usage for Mean-Reversion Strategy
This indicator is primarily designed for trend-following strategies. However, by taking the inverse of the signals, it can be adapted for mean-reversion strategies. This involves buying underperforming assets and selling outperforming ones. Caution: This method may not work effectively with highly correlated assets, as the price movements between correlated assets tend to mirror each other, limiting the effectiveness of mean-reversion strategies.
Final Thoughts
The Wavelet Filter with Adaptive Upsampling is a powerful tool for traders seeking to improve their understanding of market trends and noise. By using advanced wavelet decomposition and adaptive upsampling, this system offers a clearer, more refined picture of price movements, enhancing trend-following strategies. It’s particularly useful for detecting subtle shifts in market momentum and reconstructing price data in a way that removes noise, providing more accurate insights into market conditions.
