Troubleshooting

Troubleshooting

This page is organized by symptom. Find what you are seeing, then work through the likely causes.

The problems below fall into three categories, and knowing which category you are in saves time:

• Setup errors — something is misconfigured and the oscillator cannot run or produces obviously wrong output. These have a fix.

• Interpretation errors — the oscillator is working correctly but the behavior does not match your expectation. These require understanding, not a settings change.

• Expected behaviors that look like bugs — the oscillator is doing something that seems broken but is actually working as designed. These are marked "Not a bug" so you can recognize them without chasing a fix that does not exist.

The indicator will not load — runtime error

Symptom

You add the indicator to the chart and instead of the oscillator pane appearing, you see an error message in the indicator status or in the Pine error output. The message identifies a specific slot (e.g., "MA 01 timeframe cannot be lower than the chart timeframe").

Cause

The slot's configured timeframe is lower than the chart timeframe. The indicator enforces that every slot's timeframe must be at or above the chart timeframe. If your chart is on 15m and slot 01 is set to 5m, the script halts.

Fix

Either:

• Lower your chart timeframe to match or go below the lowest slot (e.g., switch to a 5m or 1m chart), or

• Open the indicator settings and increase each slot's timeframe to be at or above your chart timeframe

The default slot timeframes are 5m / 15m / 60m, which means the chart must be on 5m or lower for the defaults to work. If you want to use the indicator on a daily chart, you need to set all three slots to Daily or higher.

The oscillator is flat at zero — no movement

Symptom

All Fast lines sit at exactly zero with no oscillation. The blended lines may also be at zero or not plotted at all.

Likely causes

1. All slots are disabled. Open settings and check that at least one slot has "Enable" turned on.

1. The symbol has no data for the requested timeframe. If you request a 5m timeframe on a symbol that only provides daily data (some indices, some exotic instruments), the MA and ATR calculations return na, and the oscillator resolves to zero or na.

1. The MA length is too long for available data. If the slot's timeframe is 60m and the length is 200, the indicator needs 200 sixty-minute bars of historical data. On symbols with limited history, the data may not be available, producing na.

Fix

Check that at least one slot is enabled and that the symbol provides data at the configured timeframe. Try reducing the MA length or switching to a more data-rich symbol to test.

The blended line is missing — not plotted

Symptom

The individual slot Fast lines are visible, but the Blended Fast and Blended Slow lines are not drawn. The fill is also absent.

Likely causes

1. Plot Blended Fast/Slow is turned off. Check the Display settings. If this toggle is off, the blended lines and fill are hidden (but blended alerts still evaluate).

1. All enabled slots have weight 0. The blend requires at least one enabled slot with non-zero weight to produce a value. If every slot's weight is 0, the blend has no data to average and returns na.

1. All enabled slots have na Fast values. This can happen if the symbol has no data for any of the configured timeframes.

Fix

Turn on the blended plot if it is off. Ensure at least one enabled slot has a non-zero weight and is producing valid data.

The oscillator appears pinned at +100 or −100

Symptom

One or more Fast lines sit at or very near the bound (+100 or −100) for an extended period — sometimes dozens or hundreds of bars — without pulling back toward center.

Not a bug

This is the tanh saturation working as designed. The score approaches ±100 asymptotically when the price-to-MA distance in ATR units is large. During a strong trend, that distance stays large, and the oscillator holds near the bound.

The oscillator is not stuck. It is accurately reporting that price remains far from the average. Within the saturated zone, the oscillator cannot distinguish between "far" and "very far" — that resolution loss is a known property of the construction. See Limitations & Trust Boundaries — Saturation.

What to do

If saturation persists longer than you find useful, consider lowering the sensitivity parameter. Lower sensitivity keeps the oscillator in the high-resolution center for a wider range of actual distances, which means it takes a larger move to push the oscillator into the saturated zone. See Settings — ATR Sensitivity.

The oscillator dropped sharply but price did not reverse

Symptom

The oscillator drops from a high reading (e.g., +75) to a moderate reading (e.g., +35) over a few bars. You check the price chart and price has not reversed — it may even be slightly higher.

Not a bug

This is ATR expansion. The instrument's volatility increased (wider bar ranges, a news event, a gap), which made ATR larger. The oscillator divides the price-to-MA distance by ATR. When ATR grows and the distance stays the same, the ratio shrinks, and the bounded score drops.

The oscillator is not telling you momentum reversed. It is telling you that the same physical distance from the MA is now less remarkable in the context of the instrument's current volatility.

What to do

When the oscillator drops without a corresponding price move toward the MA, check whether volatility spiked. If it did, the drop is a normalization event, not a momentum event. This is useful information — it means the market's volatility environment changed — but it should not be interpreted as a directional signal.

The blended line does not match the visible individual slot lines

Symptom

You see two or three individual slot lines that are all in positive territory, but the blended line is lower than expected — maybe near zero, or even slightly negative.

Likely causes

1. A hidden slot is pulling the blend. The most common cause. A slot with Hide Plot on is invisible but still contributes to the blend at its configured weight. If that hidden slot is in negative territory, it pulls the blend down even though everything you can see is positive.

1. Unequal weights. Even with all slots visible, if one slot has a much higher weight and it is the slot with the lowest reading, the blend will be pulled toward that slot's value.

Fix

Open settings and check for any slot where Enable is on but Hide Plot is also on. If the hidden slot is in different territory than the visible ones, it explains the discrepancy. Either un-hide it, set its weight to zero, or disable it.

If no slots are hidden, check the weight distribution. The blend is a weighted average — a slot with weight 80 dominates two slots with weight 10 each.

An alert is not firing

Symptom

You set up an alert for one of the indicator's conditions, but it does not fire when you expect it to.

Likely causes

1. The condition is not true on a confirmed bar. All alerts are bar-close gated. The condition must be true after the bar closes, not during intrabar fluctuations. If the oscillator crosses a threshold intrabar but pulls back before bar close, the alert will not fire.

1. The trigger frequency is wrong for the condition type. If you set a continuous-state alert (like "Is Bullish") with "Only Once" trigger, it fires once and then stops — even if the condition continues to hold. If you expected repeated notifications, you need "Once Per Bar Close."

1. The slot is disabled. Disabled slots do not fire per-slot alerts. If the slot you are watching is disabled, the alert condition never evaluates.

1. The alert condition targets the blend, but the blend has na values. If all slots have weight 0 or all are disabled, the blend is na and blended alert conditions cannot evaluate.

Fix

Verify the condition is true on a completed bar (not just intrabar). Check the trigger frequency setting in TradingView's alert dialog. Confirm the relevant slot is enabled. For blended alerts, confirm at least one enabled slot has non-zero weight.

Regime color flickers rapidly

Symptom

A slot's Fast line color alternates between vivid and faded on nearly every bar, making the regime state unreadable.

Not a bug

This happens when the Fast and Slow lines are very close together and oscillating around each other. With a Slow Length of 3 (the default) and price hovering near the MA, the distance score is small and both Fast and Slow sit near zero. Tiny fluctuations push Fast above and below Slow repeatedly.

What to do

If the flicker is distracting, increase the Slow Length for that slot. A Slow Length of 5 or 7 creates more separation between Fast and Slow, which means regime flips require a more sustained change in the distance score. This reduces noise at the cost of slightly slower regime detection when genuine transitions occur.

Alternatively, recognize the flicker as information: it tells you the oscillator is near neutral for that slot. When Fast and Slow are nearly identical and both are near zero, the slot has nothing strong to say. The flicker is the visual equivalent of "no meaningful signal."

The oscillator values shift when I toggle On Bar Close

Symptom

You open settings, toggle On Bar Close from on to off (or vice versa), and the most recent bars change their oscillator values.

Not a bug

This is the expected behavior of the confirmation mechanism. See MTF & Repainting for a full explanation.

With On Bar Close on, the chart shows the previous confirmed HTF value for each lower-timeframe bar inside the requested-timeframe block. With it off, the chart uses the current HTF bar instead. Toggling can therefore change more than just the far-right bar — it changes how lower-timeframe bars inside each HTF block are painted historically.

Master smoothing makes the oscillator barely move

Symptom

You enabled master smoothing and the blended lines became very flat, barely responding to price changes. Regime flips are rare or absent.

Cause

Master smoothing adds a moving-average pass on top of the already-blended Fast and Slow lines. If the Slow Length on each slot is already long (e.g., 10) and the Master Length is also long (e.g., 10), you have double smoothing with a combined effect that dramatically flattens the output. The oscillator is being averaged so many times that the underlying distance signal is buried.

Fix

Either reduce the Master Length (try 3 for a mild effect), reduce the per-slot Slow Lengths, or disable master smoothing entirely. The tool is designed to measure distance from the average — if the smoothing makes it unable to show meaningful changes in a timeframe you care about, the smoothing has gone too far.

One slot seems much noisier than the others

Symptom

Slot 01 (5m) oscillates rapidly between wide extremes while slot 03 (60m) moves slowly and smoothly. The 5m slot looks chaotic compared to the others.

Not a bug

This is the expected consequence of different timeframes. The 5m slot reacts to shorter-term price changes — its MA and ATR update every 5 minutes, so the distance score changes frequently. The 60m slot updates every hour, so its readings are inherently smoother and slower.

If the noise on the shorter slot is distracting, you can increase its MA Length (which smooths the baseline), increase its Slow Length (which smooths the regime detection), or increase the ATR Length (which smooths the normalization). Each of these trades responsiveness for stability.

Alternatively, accept the noise as accurate information: the short-term distance score really is changing rapidly. That is what short timeframes look like.

The oscillator behaves differently on different instruments

Symptom

The same settings (sensitivity 1.0, ATR 14, EMA 20) produce an oscillator that is always saturated on one instrument but rarely leaves the center on another.

Not a bug

Different instruments have different typical distances from their MAs in ATR terms. A volatile cryptocurrency might routinely sit 3-4 ATR away from the 20-period EMA, while a low-volatility forex pair might rarely exceed 1 ATR. The same sensitivity setting maps these different distance profiles to the bounded scale differently — one saturates, the other stays centered.

What to do

Adjust the sensitivity parameter per instrument. If the oscillator is always saturated, lower sensitivity. If it never reaches the extremes, raise it. The goal is for the oscillator to use enough of the range to be informative for the instrument you are trading.

A practical starting point: look at a few weeks of data at sensitivity 1.0. If the oscillator spends most of its time above ±80 or below −80, try 0.5 or 0.7. If it rarely leaves the ±30 band, try 1.5 or 2.0. You are looking for a setting where normal market conditions use the middle portion of the scale and strong moves reach the extremes without living there permanently.

Do not assume that the sensitivity that works on one instrument transfers to another. A volatile crypto and a quiet bond ETF need different sensitivity settings the same way they need different stop distances. See Settings — ATR Sensitivity for the full explanation and For the Geeks — Sensitivity experiment for a hands-on calibration routine.