Troubleshooting

This page maps the symptoms most readers report to the actual causes underneath. Most of the items below are not defects — they are settings interacting with other settings, or the consequences of timing choices that the trim makes explicit. The framing matters: a misunderstanding has a different fix from a setup error, which has a different fix from a real product limit.

Use the three categories below to locate your symptom quickly. Inside each item, the structure is the same: what you see, the most likely cause, the verification step, the fix.

A useful frame before you start

When something looks wrong on the pane, sort it into one of three buckets:

• Setup error. A specific setting is invalid for your chart or for another setting. Almost always reproducible. Almost always fixable in the dialog.

• Misunderstanding. A deliberate behavior of the indicator that does not match a reader's expectation. The fix is to understand the behavior and either accept it or reconfigure the surrounding settings to align with what you actually want.

• Genuine limit. A behavior the indicator cannot do. The fix is to recognize the limit and stop trying to bend the indicator into something it is not.

Several items below are explicitly tagged so you know which bucket you are in. The buckets matter because each one has a different posture — a setup error demands action, a misunderstanding demands study, and a genuine limit demands either acceptance or a different tool. Trying to "fix" a misunderstanding through settings usually produces a second problem layered on top of the first.

The pane is empty or the blend is missing

Symptom: nothing draws at all

Likely cause (setup error): every slot is disabled. Verification: open the settings dialog and check the Enable RSI NN flags. If all ten are off, the pane has nothing to draw. Fix: enable at least one slot.

Likely cause (setup error): every weight is zero. Verification: check the Blended Weight values for the enabled slots. Fix: set at least one weight to a non-zero value if you want the blend to draw, or accept that with all-zero weights only the slot lines (not the blend) can appear.

Symptom: slot lines draw but the blend does not

Likely cause (setup error): every enabled slot has weight zero. Verification: sum the weights of the enabled slots. If the total is zero, the blend has no contributors. Fix: raise at least one slot's weight above zero.

Likely cause (setup error): you have set Plot Blended RSI/Signal to off. Verification: check the input under the General section. Fix: turn the blended pair plot back on.

Likely cause (misunderstanding): the contributing slots are still in their warm-up window and have not produced values yet. Verification: scroll to a more recent bar and see whether the blend appears once the slots warm up. Fix: wait for the slots to warm up; do not over-read what is happening during the first handful of bars after a chart load.

Symptom: blend draws but no slot lines

Likely cause (setup error): every slot has Hide RSI NN Plot set to on. Verification: check the hide flags on enabled slots. Fix: unhide at least one slot if you want to see per-slot behavior.

A runtime error names a slot

Symptom: TradingView shows a runtime error message that names one of the slots

Likely cause (setup error): the named slot's timeframe is below your chart timeframe. Verification: open settings, find the slot named in the error, compare its TimeFrame value to the chart timeframe. Fix: raise the slot's timeframe to a value greater than or equal to the chart timeframe, or lower the chart timeframe so it sits at or below the slot's timeframe.

Why the script does this (frame: this is not a defect): a higher-resolution series cannot be honestly served down into a lower-resolution chart bar without TradingView inventing values. The script declines to invent and refuses the request. Naming the offending slot in the error message is intentional, so you do not have to hunt across ten slots to find which one disagrees with the chart.

Alerts behave in unexpected ways

Symptom: an alert fires every confirmed bar instead of only on changes

Likely cause (misunderstanding): every alert in the script is a state alert, not an edge-triggered cross alert. State alerts fire on every confirmed bar where the named state holds. Verification: read Alerts, particularly the "state alerts versus cross alerts" section. Fix: if you wanted cross behavior, you build it externally — either by composing your own conditions in TradingView's plot-based alert dialog, or by routing through automation that suppresses repeats.

Symptom: All RSI Slots Bullish fires when only one slot is bullish

Likely cause (misunderstanding): only one slot is enabled and available. Alignment counts enabled slots with available RSI values — when only one slot is available, "all" is satisfied trivially whenever that one slot is in the named state. Verification: count the slots that have Enable set to on. Fix: if you want the alignment to mean "agreement across multiple slots that you actually intend to count," enable the slots you want to count and disable the rest.

Symptom: All RSI Slots Bullish does not fire even though the slots that "matter" are all bullish

Likely cause (misunderstanding): a weight-zero slot is bearish and is silently breaking the alignment. Alignment counts enabled slots with available RSI values regardless of weight. Verification: look for an enabled slot whose weight is zero. Check its bullish/bearish state. Fix: disable (rather than zero-weight) the slots you do not want counted in alignment, or accept that your alignment includes them.

Symptom: there is no alert for "RSI crosses overbought" or similar

Likely cause (genuine limit): the script does not expose alert conditions on the overbought, oversold, or midline guides. The 70/30/50 lines are visual references only. Verification: read Alerts for the full list of exposed conditions. Fix: build the alert externally using TradingView's plot-based alert dialog with the indicator's blended RSI or a slot's RSI as the source.

Timing and higher-timeframe slots

Symptom: a slot appears to be lagging behind the others

Likely cause (misunderstanding): On Bar Close? is on for that slot. With the switch on, the slot returns the previous confirmed higher-timeframe bar's value, which is by definition lagged by one higher-timeframe bar. Verification: check the slot's On Bar Close? setting in the per-slot power-user section. Fix: if you want responsiveness, turn the switch off — and accept that the slot's value can move during the higher-timeframe bar's formation. See MTF and Repainting.

Symptom: a slot's historical value changes when the chart reloads

Likely cause (misunderstanding): the slot has On Bar Close? off. With the switch off, the slot returns the live higher-timeframe value during bar formation; once the higher-timeframe bar closes, the historical value at that point settles to the closed-bar value rather than the live one. Reloading the chart shows the settled historical values. Verification: the previous-versus-live test described in MTF and Repainting. Fix: if you want stable historical reads, turn the switch on. If you want responsive live reads, accept that historical values will settle.

Symptom: a slot is "frozen" for hours

Likely cause (misunderstanding): the slot has an Optional Ticker set to a symbol whose market is currently closed. The slot is honestly reporting that no new bar has formed for the foreign symbol. Verification: check whether the foreign symbol's market is open. Turn the optional ticker off temporarily and confirm the slot starts updating. Fix: none required if you want the cross-asset behavior; this is the expected reading. If you do not want the foreign session pause, choose a different symbol or remove the optional ticker.

Cross-asset and cross-session

Symptom: the blended pair seems to describe the wrong instrument

Likely cause (setup error or misunderstanding): one or more enabled slots have an Optional Ticker set, and those slots have non-zero weight. The blend now incorporates the foreign symbol into a composite that is no longer "an RSI of the chart." Verification: check Optional Ticker on every enabled slot. Note which slots have non-zero weight. Fix: either accept that the blend is a multi-context composite (and read it that way deliberately), or set the cross-asset slot's weight to zero so it provides context without steering the blend. See Workflows for the cross-asset routine.

Symptom: alerts fire repeatedly during a session where I did not expect them

Likely cause (misunderstanding): a cross-asset slot's session does not align with the chart's. The alert is honestly evaluating on every confirmed chart bar; the foreign slot has not changed state because its market has not produced new data; the slot's prior state is still active and the alert is still firing. Verification: check whether the slot driving the alerts is on a foreign symbol with a closed market. Fix: decide on purpose whether you want alerts to keep firing during off-hours mismatches. If not, gate them externally or restructure your slots so the alerts run on slots that align with your chart's session.

Visual and color confusion

Symptom: a slot color "pops" briefly at chart load

Likely cause (misunderstanding): during warm-up, before the slot's calculated signal value exists, the color is driven by an RSI-vs-50 fallback rather than by the RSI-vs-signal comparison. Once the signal warms up, the color settles into its normal behavior. Verification: scroll back to the beginning of the indicator's history on the chart and confirm the early bars use the fallback. Fix: none required; treat the first handful of bars as warm-up.

Symptom: the blend lags visibly behind the slots

Likely cause (misunderstanding): master smoothing is on. The post-blend smoothing pass adds latency to the composite by design. Verification: check Enable Master Smoothing and Master Length. Fix: if you wanted responsiveness, lower the master length or turn master smoothing off entirely. If you wanted calm, the lag is the cost.

Likely cause (misunderstanding): a heavily-weighted higher-timeframe slot is pulling the blend toward its slower pace. Verification: look at the weight distribution and the slot timeframes. Fix: rebalance weights, or disable master smoothing, or read the slot plots when you want the live story.

Symptom: the blended pair stops reaching the overbought/oversold guides

Likely cause (misunderstanding): the master smoothing length is producing a smoothed series that no longer extends to the default thresholds. Verification: turn master smoothing off temporarily; if the unsmoothed blend reaches the guides, smoothing is the cause. Fix: widen the overbought and oversold guides (75/25, 80/20, etc.) or lower the master length so the smoothed series reaches the guides again.

Symptom: a line pinned at 0 or 100

Likely cause (misunderstanding): the value reached or exceeded the clamp boundary. The clamp is doing its job; the line is not measuring an extreme of the underlying series, only reporting "at or beyond the frame." Verification: check the unsmoothed slot lines for the actual story. Fix: none required, but read pinned values as boundary, not as quantified extreme.

Performance, refresh, and edge cases

Symptom: the chart feels slow when the indicator is loaded

Likely cause (misunderstanding): you have many slots enabled with heavy MA types and tuning. Each enabled slot computes whether or not it draws. Verification: disable slots you are not actively using and watch the chart's behavior. Fix: disable slots you do not need rather than just hiding their plots; hidden slots still compute. Choose simpler MA types where possible.

Symptom: the indicator behaves differently after a TradingView reload

Likely cause (misunderstanding): live values you saw before the reload have settled to confirmed values. Slots with On Bar Close? off are most affected. Verification: see MTF and Repainting for the live-versus-confirmed walkthrough. Fix: none required; the behavior is honest. If consistency across reloads matters more than responsiveness, turn the switch on.

Symptom: I cannot find a setting I expect from a different RSI tool

Likely cause (genuine limit): this trim does not expose every input that other RSI implementations expose. It exposes the inputs that the design supports — slot configuration, MA selection, weights, on-bar-close, master smoothing, overbought, oversold, line widths, and the per-slot power-user knobs. Verification: read Settings for the full surface. Fix: if the missing setting matters, the trim may not be the right tool for that workflow.

When the symptom does not match anything here

A short routine for getting unstuck.

1. Reset the indicator to defaults (the dialog has a "Defaults" option). Confirm whether the symptom is present at defaults.

2. If the symptom is present at defaults, the cause is in your chart or your TradingView environment, not in your configuration. Check chart timeframe, symbol, and any saved layouts.

3. If the symptom disappears at defaults, you are dealing with a configuration interaction. Re-introduce your changes one at a time until the symptom reappears. The change that reintroduces it is the cause. This is slower than it sounds if you had twenty changes; do it anyway. The alternative is guessing.

4. If the symptom persists at defaults and on a freshly loaded chart, you have likely found a real defect. Capture the following before you report it:

• The chart symbol and the chart timeframe.

• The time of day and the TradingView timezone setting.

• The indicator's published version (visible in TradingView's indicator info).

• A screenshot that includes the pane, the relevant settings dialog section, and the chart context (at least a handful of bars around the moment).

• A short written description of what you expected versus what you saw. The written description matters more than it seems — it is what lets a support reviewer reproduce the read on their own chart.

Report through the support channel that came with your subscription. Skipping the capture step because the symptom "seems obvious" is the most common reason a defect report goes unresolved for longer than it needed to.

Where to go next

• For settings details that may resolve a configuration issue, go to Settings.

• For timing-related causes, go to MTF and Repainting.

• For honest limits that are not bugs, go to Limitations and Trust Boundaries.