Troubleshooting

Troubleshooting

This page is organized by what you see, not by internal cause. Find the symptom, then follow the diagnosis.

The indicator shows an error instead of plots

"Timeframe cannot be lower than the chart timeframe"

What happened: One or more slots have a timeframe setting that is lower than the chart timeframe. The script enforces the rule that every slot's timeframe must be equal to or higher than the chart timeframe.

Most common scenario: You loaded the indicator on a 15-minute chart and the default Slot 01 is set to 5m.

Fix: Either lower your chart timeframe to 5m or below, or open the indicator settings and raise the offending slot's timeframe to at least match the chart. Every enabled slot must be at or above the chart timeframe.

Prevention: When you change your chart timeframe, check that no slot's timeframe falls below it. This is the most common first-use error and the most common error when switching between chart timeframes.

"Fast Length must be less than Slow Length"

What happened: An enabled slot has Fast Length >= Slow Length. The MACD calculation requires the fast MA to use a strictly shorter lookback than the slow MA.

Most common scenario: You were adjusting lengths and set Fast to 26 and Slow to 12 (inverted from the default 12/26).

Fix: Set Fast Length to a value strictly less than Slow Length. The default 12/26 is always safe.

The blended K/D lines are missing or show na

All blended lines are absent

Possible causes:

1. All enabled slots have weight 0. The blend is a weighted average — it needs at least one slot with a nonzero weight to produce a value. Slots 04–10 default to weight 0. If you disabled the first three slots and only enabled later ones without setting their weights, the blend has nothing to work with.

Fix: Set a nonzero weight on at least one enabled slot.

1. Not enough price history. Every slot's MACD needs enough bars of data for its moving average calculations. On a chart with very few bars (small timeframe on an instrument with limited history), slots with long MA lookbacks may produce na values. If all slots return na, the blend returns na.

Fix: Use a chart with more historical data, shorten the MA lengths, or lower the slot timeframes.

Blended lines appear but have gaps

Cause: One or more slots periodically produce na values (typically due to insufficient history at the start of the chart or data gaps in the feed). The blend handles this by excluding na slots from the calculation for those bars — but if all remaining contributing slots also produce na, the blend goes to na.

Fix: This is usually a data availability issue. Scroll right to more recent bars where more history is available.

A slot line is not visible

The slot is disabled

Check the "Enable MACD XX" toggle in settings. Disabled slots do not plot, do not compute, and do not contribute to anything.

The slot is enabled but hidden

Check the "Hide MACD XX Plot" toggle. When this is on, the slot's K line does not appear on the chart — but the slot still computes its MACD, still contributes to the blend (if weight > 0), and still fires alerts. This is a deliberate feature for reducing visual clutter when you want the slot's math in the blend without the visual.

Fix if unintentional: Turn off "Hide MACD XX Plot."

The slot line starts partway through the chart

Cause: The slot needs enough historical bars to compute its moving averages. Slots with longer lookback periods (high Slow Length) or slots on higher timeframes (where each bar represents more time) need more chart bars before they produce their first valid value.

This is normal. The line will appear once enough bars exist for the calculation.

Slot lines all look identical

Cause: Multiple slots share the same timeframe, source, ticker, MA settings, advanced parameters, and On Bar Close mode. They are computing the same slot logic, so their normalized values are identical and their lines overlap.

This is expected behavior given the configuration. If you see overlapping slot lines, check whether the slots differ in at least one meaningful parameter. Different timeframes, different tickers, different MA settings, different advanced parameters, or different On Bar Close modes will produce different readings.

The blended line tracks one slot exactly

Cause: Only one slot has a nonzero weight. The blend is a weighted average — with one contributor, it mirrors that contributor exactly.

Also possible: Multiple slots are enabled with nonzero weights, but all produce the same values (see "slot lines all look identical" above). The blend of identical values is the same as any one of them.

Fix if unintentional: Check the weight settings. Verify that more than one slot has weight > 0 and that the slots are configured differently enough to produce distinct readings.

The oscillator stays pinned near +100 or -100

Cause: ATR Sensitivity is set too high for the instrument and timeframe. At high sensitivity values (2.0+), even moderate momentum moves push the normalized value to the edge of the bounding curve. The oscillator saturates and loses mid-range resolution.

Fix: Lower the ATR Sensitivity. Start by reducing from the current value toward 1.0 and observe how the oscillator's range usage changes. See Settings — ATR Sensitivity for the full explanation.

Alternative cause: The instrument is in an unusually strong trend and the reading is genuinely extreme. Before adjusting sensitivity, check whether the price action supports the extreme reading.

The oscillator stays compressed near zero

Cause: ATR Sensitivity is set too low. At low sensitivity values (0.3–0.5), the oscillator barely moves and the OB/OS thresholds are almost never reached.

Fix: Raise ATR Sensitivity toward 1.0. See Settings — ATR Sensitivity.

Alternative cause: ATR Length is very long (30+) and the instrument's ATR is currently elevated due to a recent volatility spike. The high ATR denominator compresses the normalized reading even at default sensitivity. This can resolve on its own as the ATR lookback window rolls past the spike. If you suspect this is happening, temporarily shorten ATR Length to 7 and see if the oscillator's range usage changes significantly — that confirms the denominator was the issue, not the sensitivity setting.

The blended line does not react when I enable a new slot

Most likely cause: The newly enabled slot's weight is 0. Slots 04–10 default to weight 0 when first enabled. A weight of zero means the slot plots and fires alerts but does not contribute to the blend.

Fix: Set a nonzero weight for the slot. Even a small weight (10 or 20) will cause the blend to shift.

Also possible: The new slot produces values very similar to the existing slots (same timeframe, same parameters). The blend does shift, but the change is too small to see visually.

An alert fires on every bar instead of once

Cause: You set up a state alert (e.g., "MACD 01 Is Bullish" or "All MACD Slots Bullish"). State alerts fire on every confirmed bar where the condition holds — not just when the condition first becomes true.

Fix: If you only want to know about transitions:

• Use an edge alert like "Blended MACD Regime Flip" which fires once on the crossover bar.

• Or use TradingView's alert frequency option "Once Per Bar Close" or "Only Once" to limit notifications.

See Alerts for the full explanation of state vs. edge alerts.

An alert never fires

Possible causes:

1. The condition is not being met. For OB/OS cross alerts, the blended K may never reach the threshold. This is common at low ATR Sensitivity settings where the oscillator stays compressed near zero.

Fix: Either adjust ATR Sensitivity so the oscillator reaches the OB/OS levels under normal conditions, or lower the OB/OS threshold levels in settings.

1. No slots are enabled. Per-slot alerts require the slot to be enabled. Alignment alerts require at least one enabled slot.

Fix: Enable the relevant slots.

1. The alert was set on a slot that is now disabled. If you set up an alert for Slot 03 and later disable Slot 03, the alert condition cannot be true.

Fix: Re-enable the slot or remove the stale alert.

The blended K/D crossover happens later than expected

Possible causes:

1. Master Smoothing is enabled. Master Smoothing applies a post-blend moving average that delays the K/D crossover by roughly the Master Length number of bars. At Master Length = 10, the crossover is approximately 10 bars later than it would be without smoothing.

Fix: If you need faster crossovers, disable Master Smoothing or reduce the Master Length.

1. Heavy weight on a slow-moving slot. If the highest-timeframe slot carries most of the weight and it has not flipped yet, the blend will not flip until that slot's contribution allows it. This is the designed behavior of weighted averaging — the blend respects the weight distribution you chose.

Fix if unintentional: Rebalance the weights to reduce the slow slot's dominance.

History does not match what I saw live

Cause: One or more slots have On Bar Close set to OFF. Those slots were updating with unconfirmed higher-timeframe data while you watched live. When the higher-timeframe bars closed, the readings snapped to their confirmed values. The chart now shows only the confirmed values.

This is the expected behavior of On Bar Close = OFF. It is not a bug. The intrabar readings were real at the time — they reflected the developing HTF bar's actual state. But they were provisional. When the HTF bar closed, the final confirmed value replaced them. This can be disorienting the first time you notice it, especially if you made a decision based on a reading that no longer appears on the chart.

What to check: Open the indicator settings and look at the On Bar Close toggle for each active slot. If any are set to OFF, those slots were showing intrabar data that has since been overwritten. This does not mean the tool gave you false information — it means it gave you a real-time reading of a bar that was still forming, and the bar's final value turned out differently.

Fix: If you need history to match what you saw live, set On Bar Close to ON for the relevant slots. If you deliberately use On Bar Close OFF for responsiveness, understand that the chart will not serve as a reliable journal of what you saw intrabar. See MTF and Repainting for the full explanation and a live verification walkthrough.

The oscillator looks different from a standard TradingView MACD

This is expected and it is the first thing that trips up traders who know standard MACD well. The Axiom oscillator normalizes the raw MACD by ATR and bounds the result to -100/+100. A standard MACD is unbounded and denominated in price units. The shapes may look similar, but the magnitudes are unrelated. A standard MACD might read +2.5 while the Axiom oscillator reads +45 for the same momentum condition. The numbers will never match because the scales are fundamentally different.

What should match: the direction, but only in a like-for-like comparison. If you set the slot to the chart timeframe and temporarily turn On Bar Close OFF, then when the standard MACD line is above its signal line the Axiom slot should be in its bright (bullish) color, and when below, faded. If the directions disagree, check that both are using the same timeframe, the same MA types (EMA for both fast/slow and signal), and the same lengths (12/26/9). If On Bar Close is ON on a same-timeframe slot, the Axiom slot is intentionally one confirmed bar behind.

To run this comparison: add TradingView's built-in MACD with 12/26/9 EMA to the same chart, set the Axiom slot to the chart timeframe, and temporarily set On Bar Close to OFF for that slot. Compare regime direction, not levels. See Quick Start for the step-by-step procedure.

Per-slot Power User settings seem to have no effect

Cause: The Power User parameters (ALMA Offset, KAMA Fast/Slow, Jurik Phase/Power, etc.) only activate when the corresponding MA type is selected. If the slot's MACD MA Type is EMA (the default), all Power User parameters for the MACD side are ignored. If the Signal MA Type is EMA, all "D" Power User parameters are ignored.

Fix: These settings only matter when you deliberately select a non-EMA MA type. If you are using EMA, the Power User settings are correctly inactive.