Troubleshooting

Every symptom below has a likely cause, a check you can run in under a minute, and a fix. The table up front gives you a one-screen map. The sections after walk each symptom in enough detail to actually diagnose, not just recognize.

A lot of surprise with this indicator is not a fault of the tool; it is the configuration doing something specific and the reader not having spotted it yet. These pages assume you have already read MTF & Repainting and Visuals & Logic. If you have not, most of the "likely cause" answers below will feel opaque.

A diagnostic habit that applies across every symptom: before you conclude that the tool is the problem, ask what your configuration and the current session would produce if the tool were working correctly. Most surprises land in the gap between "what I expected" and "what this configuration actually asks for." Running that question first is faster than randomly toggling inputs and often turns the symptom into a setting, not a bug.

Symptom-to-cause map

Symptom: Slot colors look unstable for a few bars after load

Likely cause. The slot is still warming up. Until the source, baseline, ATR, Fast value, and Slow value are all populated enough to be useful, early colors can be shaped by missing values or fallback behavior. With a default slow length of 3, this is a short window after the underlying values exist.

Check. Wait a few bars and see whether the color stabilizes. Look at the current slow length (Slow Length on each slot); a longer slow length takes longer to settle.

Fix. Nothing to fix. This is load behavior. If the wait is long enough to be annoying, the slow length is longer than the default.

Symptom: Pane is one color for every bar on screen

Likely cause, most often. Slot inputs all lean the same way on a strongly directional session. Same source, similar lengths, and stacked timeframes can agree for long stretches without adding much independent information.

Check. Read the source, baseline family, baseline length, slow length, and timeframe on each slot. Then scroll back across a broader session. If the pane changes color elsewhere, the current segment may simply be one-directional. If it almost never changes, the slots may be too similar or too slow for the way you are using the pane.

Fix. Differentiate at least one slot structurally, or adjust slow lengths after you know what problem you are solving. If the issue is actually a timeframe violation, the script will show the named runtime error below rather than silently collapsing the stack.

Other likely cause. Master smoothing or long slow lengths can make the visible regime change later than you expect. Check those settings before assuming the pane is stuck.

Symptom: Blend line does not draw on this bar

Likely cause. The blend is a weighted average of the enabled slots. If every enabled slot has Blended Weight = 0, the total weight is zero and the blend has nothing to average — blended fast and blended slow are na, and the plot and fill do not draw. The same happens if every enabled slot's fast value is na on that bar for any reason.

Check. Open the indicator inputs and read the three Blended Weight values. Is any enabled slot at a non-zero weight?

Fix. Set at least one enabled slot to a positive weight. The blend returns as soon as there is at least one enabled slot with a positive weight and a valid fast value.

Symptom: "MA XX timeframe cannot be lower than the chart timeframe"

Likely cause. A slot's configured timeframe is below the chart's timeframe. The script refuses this at runtime because it would require inventing higher-resolution data than the chart has.

Check. Read the error — it names the slot (MA 01, MA 02, or MA 03). Check the chart timeframe and that slot's timeframe.

Fix. Either raise the slot's timeframe to meet or exceed the chart, or blank the slot's timeframe input so it inherits the chart timeframe. The pane returns as soon as every enabled slot satisfies the rule.

Symptom: Pane pins at 0 or 100 for long stretches

Likely cause, most often. ATR Sensitivity is too high for this instrument. Pinning is the clamp doing its job — the sensitivity multiplier has pushed the ATR-normalized distance past the range the bounded transformation can resolve.

Check. Read the current sensitivity. Lower it by half on the same chart and see whether the pane resolves useful differences in the 30-to-90 band.

Fix. Lower sensitivity to a value that lets the pane spend useful time between 20 and 80 during sessions you actually trade.

Other likely cause. The instrument is genuinely extremely stretched relative to its own ATR — a gap open, a news move. In that case the pane is telling the truth and there is no fix to apply; the boundary is describing real economic behavior at the sensitivity you chose.

Symptom: Pane hovers around 50 and never reaches the reference band

Likely cause, most often. ATR Sensitivity is too low. The normalized distance is being scaled down to the point where real stretch does not push the reading out of the equilibrium zone.

Check. Raise sensitivity to 1.5 or 2.0 on the same chart for a session and watch whether the pane starts reaching 30 and 70 during moves that look meaningful on the chart.

Fix. Raise sensitivity to a value where the pane visits the reference band during moves you care about.

Other likely cause. ATR Length is long enough that the ATR reading is averaging over too much history for the instrument's regime. Shortening ATR can give the scale a tighter reference to volatility of the last few sessions.

Symptom: Higher-timeframe slot holds a flat value between its closes

Likely cause. On Bar Close? = true, which is the default. The slot is waiting for its higher-timeframe bar to close before it updates. Between closes, the slot holds the last confirmed value, which plots as a flat segment on the chart timeframe.

Check. Read On Bar Close?. On a 1m chart with a 60m slot, each flat segment is 60 chart bars long.

Fix. Nothing to fix if you want the honest behavior. If you want the slot to move between higher-timeframe closes, turn the switch off — knowing that the slot value can then change as the higher-timeframe bar evolves. See MTF & Repainting for the tradeoff.

Symptom: Slot values shift under a bar you already read

Likely cause. On Bar Close? = false, and the slot is on a higher timeframe whose bar had not closed at the time you read. The slot value was a live read of an evolving higher-timeframe bar; it has now updated.

Check. Read the switch setting and the slot's timeframe. If the switch is off and the slot timeframe is above the chart, the values at any chart bar that fell inside an uncompleted higher-timeframe bar are subject to change until that higher-timeframe bar closes.

Fix. Turn On Bar Close? = true if stability matters for your setup. If you need the faster reaction, accept that historical reads can shift under you until the higher-timeframe bar closes.

Symptom: Per-slot alerts fire from a slot you thought was off

Likely cause. The slot is enabled and only its Blended Weight is zero. Weight zero removes the slot from the blend; it does not disable the slot. The slot still fires its own per-slot alerts and participates in alignment.

Check. Read the slot's Enable toggle and its Blended Weight separately.

Fix. If you want the slot off entirely, disable it. If you want it on and observing without steering the blend (the observer-slot pattern in Workflows), leave it as is and accept that per-slot alerts will still fire.

Symptom: Alignment alerts fire all the time and do not seem to mean anything

Likely cause. The three slots are near-identical — same source, similar lengths, stacked timeframes. Alignment across three copies of the same measurement is trivially frequent during any directional session.

Check. Read the source, baseline family, baseline length, slow length, and timeframe on each slot. If they are all close in structure, alignment is not carrying information.

Fix. Differentiate at least one slot in a structural way — source, baseline family, or length. See Workflow 1 for the "differentiated three-slot stack" pattern and Limitations for why this matters.

Symptom: Pane looks calmer than it should and you cannot tell why

Likely cause. Master smoothing is on. The final smoothing pass replaces the blended fast and blended slow with smoother versions, which looks calmer and lags regime flips.

Check. Read Enable Master Smoothing.

Fix. Turn it off. Compare the two side by side on the same chart segment. If the calmer version is costing you regime flip timing you care about, leave it off. If you want the calm and can accept the lag, leave it on with a length you have audited.

Symptom: Slot line is missing but alerts still fire

Likely cause. Hide MA XX Plot is on for that slot. Hide-plot removes the line from the pane without disabling the slot.

Check. Read the Hide Plot toggle in each slot's input group.

Fix. Turn Hide Plot off if you want the line back, or disable the slot entirely if you want it fully out of the pane and out of alerts.

When a symptom does not match any row above

A few principles to use when the symptom is outside the table:

• Reload the indicator with defaults. If the symptom disappears, the cause is in your configuration. If it persists, the cause is either data (check the feed, the chart timeframe, the instrument's recent volatility) or a genuine bug to report.

• Change one input at a time and watch. Restoring a working pane by randomly changing three things hides the cause even when it finds the cure. You will hit the same symptom on a different chart with the same cause unknown.

• Ask whether the reading is consistent with the inputs you can see. A bullish blend on a red price day is not necessarily wrong; the blend is about stretch relative to baseline, not about price direction. A flat pane on a volatile day might be honest if your ATR Length is very long, which would be smoothing away the volatility change. The symptom is the one that does not fit the inputs.

• When in doubt, verify on a 1m chart on a liquid instrument. That is the environment the verification moves across this pack are written for. Reproducing a symptom there makes it cheaper to diagnose; failing to reproduce it there usually means the symptom was data-specific or regime-specific, not tool-specific.

Where to go next

• A symptom here does not match what you are seeing: Limitations & Trust Boundaries is the next place to look.

• You want to understand why a fix works the way it does: For the Geeks explains the shape of the transformation.

• You want to build a configuration deliberately rather than debug one: Workflows.