Troubleshooting

Troubleshooting

This page organizes problems by what you see on the chart. Each entry covers the symptom, the most likely cause, what to check, and how to fix it. If the problem is actually the tool working as designed (and several of them are), the entry says so and points you to the page that explains why.

Runtime errors (the indicator will not load)

"Error: Timeframe must be >= chart timeframe" or similar

What happened: A slot's timeframe is set lower than the chart timeframe. This is a hard constraint — the indicator cannot request data at a finer resolution than the chart provides.

Fix: Open the indicator settings and check the timeframe for every enabled slot. Each must be at or above the chart's timeframe. If you recently changed the chart timeframe (e.g., from 1m to 15m), any slot set to a timeframe between 1m and 15m will now violate the constraint.

"Error: Lower TF Precision must be < slot timeframe" or similar

What happened: A slot's Lower TF Precision is set equal to or higher than its slot timeframe.

Fix: Set Lower TF Precision to a timeframe strictly below the slot's timeframe. If the slot is 15m, valid Lower TF Precision values include 1m, 5m, etc. — anything below 15m.

"Error: Window must be >= slot timeframe" or similar

What happened: A slot's Window setting is lower than its slot timeframe.

Fix: Set the Window to at least the slot timeframe. If the slot is 1h, the window must be at least 60m. For Session mode, a daily window (D) is the most common choice.

The indicator loads but the pane is empty

Likely causes:

• All slots are disabled. Check that at least one slot has "Enable CVD XX" turned on.

• The chart symbol has no volume data. The oscillator requires volume. Some instruments (certain forex pairs, synthetic indices) may report zero or no volume, in which case the oscillator has nothing to estimate from.

• A ticker override points to an invalid or unsupported symbol. Check Optional Ticker on any enabled slots.

Visual confusions

All slot lines jump to zero at the same time

What it is: A session reset. All default slots use Session mode with a daily window, so they all reset at the daily session boundary.

Is this a problem? No. This is the oscillator working as designed. In Session mode, the CVD accumulation restarts at each window boundary. See Quick Start for what normal post-reset behavior looks like.

The 1h slot (or any higher-TF slot) barely moves

What it is: On Bar Close is ON and the slot timeframe is much higher than the chart timeframe. A 1h slot on a 5m chart only updates once every 12 chart bars. Between updates, the line holds flat.

Is this a problem? No. The slot is waiting for the next confirmed HTF bar before updating. This is the cost of non-repainting data. If you want faster updates, you can turn On Bar Close off for that slot — but the values will repaint until the HTF bar closes. See MTF and Repainting.

The blended line does not match the visible average of the slot lines

Most likely cause: A hidden slot is contributing to the blend. Check whether any slots have "Hide CVD XX Plot" turned on — a hidden slot still computes and contributes at full weight to the blended calculation. Its line is invisible but its numbers are in the blend.

Second possibility: A slot that appears disabled (no visible line) is actually enabled with an unusual weight. Open the settings and check the Enable and Blended Weight status of all 10 slots.

Third possibility: A slot has its weight set to zero. It draws on the chart but does not contribute to the blend. The visual average of all visible lines will not match the blend because the weight-zero slot's values are excluded from the calculation.

Readings look extreme (+80 or higher) on low volume

What it is: The normalization range is narrow, likely because the session just reset or the rolling window covers very few bars. A small amount of directional volume can push the normalized reading to an extreme because the range has not had time to expand.

Is this a problem? Only if you treat the reading with the same confidence as a mid-session extreme. The number is mathematically correct — it is at the top of the current range. But the current range is narrow. Give the session 15-30 minutes of data before weighting the readings.

Power User settings appear to do nothing

What it is: The Power User parameters (ALMA Offset, KAMA Fast/Slow, Jurik Phase, etc.) only activate when the corresponding MA type is selected for that slot's CVD Type or Signal Type. If you are using SMA or EMA, the advanced parameters are inert by design.

Fix (if you expected them to work): Check whether the CVD Type or Signal Type for the relevant slot is set to the MA type that uses those parameters. ALMA params work only when ALMA is selected. KAMA params work only when KAMA is selected. And so on.

Interpretive confusions

The oscillator shows bullish but price is falling

What it means: The participation model is classifying the candle structure as leaning bullish — more bars are classified as having bullish participation than bearish — but price is declining. This can happen when:

• The decline is on low volume while the prior rise was on higher volume. The accumulated delta from the bullish bars still outweighs the bearish contribution.

• The candle structure during the decline has mixed characteristics (long lower wicks, close in the upper half of the range) that the model classifies as weakly bullish or carry rather than bearish.

• The reading is lagged by the MA smoothing. The CVD line is reflecting bars from a few periods ago, and the decline has not yet worked through the smoothing.

What to do: Check two things. First, is the CVD line declining even though it is still above the Signal? If so, you are looking at bullish regime with bearish momentum — the label has not caught up to what the line is doing. The direction of the CVD line is the earlier signal; the regime label is the slower confirmation. Second, check the candle structure on the chart. Are the declining candles mixed — red bodies but with lower wicks, closes in the mid-range? The participation model may be classifying them as weakly bearish or even carry rather than strongly bearish, which means the accumulated delta is changing slowly. This is the model being conservative, not wrong. If the decline is on clean, decisive bearish candles, the CVD line should catch up and the regime will flip. If the decline is on ambiguous structure, the model is telling you the pressure is weaker than the price chart looks.

CVD near zero for a long time — is the indicator working?

What it means: Estimated buying and selling pressure are close to equal within the active window. This is a valid market state, not a malfunction. Volume can be high while net directional pressure is balanced.

What to check: Confirm the slot is enabled and running (its line should be visible and near zero, not absent). If you are in Rolling mode, check whether old directional bars have aged out of the window, leaving a neutral-looking read.

One slot reads very differently from the others

Possible causes:

• Different timeframes, which is the point. A 5m slot and a 1h slot should read differently most of the time.

• Different window modes. A Session-mode slot and a Rolling-mode slot on the same timeframe will produce different readings because they normalize against different data ranges.

• Different Pressure Sensitivity or Wick Weight. In the Pro variant, these are per-slot. If one slot has sensitivity at 0.50 and another at 3.0, their classifications of the same candle data will differ substantially.

• Different ticker overrides. A slot reading BTCUSDT and a slot reading ETHUSDT are reading different markets entirely.

What to do: Check the settings for each slot. If the difference is intentional, it is working as designed. If it is not intentional, the most common accidental cause is a leftover ticker override or a window mode that was changed during testing and not reverted.

Alerts fire much more often than expected

Most likely cause: You are using a per-slot state alert, which fires every confirmed bar that the condition is true. A 30-minute bullish run on a 1-minute chart produces 30 alerts.

Fix: Switch to TradingView's "Once Per Bar Close" frequency setting, or use the Blended CVD Regime Flip alert (which fires only at the crossing, not every bar). See Alerts for the full catalog of state vs. edge alerts.

The indicator looks different after changing the chart timeframe

What happened: The relationship between the chart timeframe and the slot timeframes changed. Slots that were previously on a higher timeframe may now be at or equal to the chart timeframe, changing their update frequency and resolution. A 1h slot that updated every 12 bars on a 5m chart updates every bar on a 1h chart.

What to check: After changing the chart timeframe, review the slot timeframes. Ensure each is still at or above the new chart timeframe and that the timeframe spread still makes sense for the new chart resolution.

The oscillator looks different on Monday morning compared to Friday's close

What happened: If you are using Session mode with a daily window, the weekend or holiday gap introduced a new session boundary. The CVD accumulation reset, and the normalization range started fresh with Monday's data. The Friday close reading was based on a full day's data with a wide normalization range. Monday's opening reading is based on the first few bars of a new session with a narrow range.

If you are using Rolling mode, the gap may have caused bars from Friday's session to age out of the window without replacement during the weekend. The rolling range may have contracted, which means the same absolute CVD value produces a different normalized reading than it would have on Friday.

What to check: Confirm which window mode each slot uses. For Session-mode slots, the reset at Monday's open is expected — treat the first 15-30 minutes the same way you would any session-open thin-data period. For Rolling-mode slots, be aware that the range may need time to re-establish itself with new data. Neither case is a malfunction.