Limitations and Trust Boundaries

This page is the honest-limits page. It is here so that the parts of the indicator most likely to be misread under pressure have names you can reach for. None of these limitations are surprises hiding in the source. They are the visible consequences of choices the indicator makes on purpose, and your job, as a reader of the pane, is to keep them in mind when the chart starts feeling persuasive.

There is no setting in this trim that removes the items below. They are the boundary of what the instrument can honestly tell you. You can work with them; you cannot work around them.

What the composite is, and what it is not

The blended RSI line and the blended signal line are the visible centerpiece of the pane. They are easy to read at a glance and easy to over-trust at the same glance. Be precise about what they actually are.

They are a weighted average of the enabled slots that have non-zero weight and usable RSI values, with weights you chose, clamped into 0..100.

That definition has consequences:

• The blend is exactly as honest as your weighting. If you weighted on purpose, the composite reflects intent. If you accepted defaults and added more slots without revisiting weights, the composite reflects whatever the defaults happened to do.

• The blend treats your enabled, non-zero-weight slots with usable RSI values as the universe of opinion. It cannot tell you whether you chose the right slots — it can only average the ones you let steer it.

• The blend is silent about why the slots agree or disagree. A clean, all-bright pane is not evidence that the read is correct; it is evidence that your composite is internally consistent.

What the blend is not:

• Not a "consensus across timeframes." It is an average of correlated views.

• Not an "objective read." Every input was your decision.

• Not a "confirmation" of the slot pattern. It is the same information arithmetically combined.

If you find yourself describing the blended pair to someone else as "the indicator says X," substitute "my weighted average of these slots says X." You will catch yourself recognizing how much of the read you authored.

Autocorrelation across timeframes on a single symbol

The single most common misread of this trim, by a margin: treating multi-timeframe alignment as multi-source confirmation.

When you run RSI on the same symbol at 5m, 15m, and 60m, you are not seeing three independent opinions. You are seeing three smoothed views of the same price path. The 60m slot sees what the 5m and 15m slots saw, plus more bars; the three RSIs are derived from overlapping data windows and will agree with each other most of the time by construction.

Concretely:

• If you fire on All RSI Slots Bullish with a typical multi-timeframe stack on one symbol, you are firing on a state that is much more common than "three independent indicators agree."

• A high alignment rate is not evidence of confluence. It is evidence that the slot lines are tracking the same underlying series.

• The information that the slots disagree is more interesting than the information that they agree. A fractured stack tells you about regime transition or about timeframe-specific behavior; a coherent stack tells you about your slot configuration.

The way to recover real independence is to add slots that are genuinely independent — typically a cross-asset slot on an instrument with low correlation to the chart symbol, or slots that read different price series (close vs. typical price), or slots that use very different smoothing types. Doing this without thinking about it produces a different composite, not a more reliable one.

A blunt heuristic worth holding: if you cannot name, specifically, what information a slot adds that the other slots do not already carry, that slot is probably contributing autocorrelation rather than independence. The right question is not "how many slots should I enable" but "what question is each slot answering that no other slot answers." If two slots answer the same question, one of them is noise dressed as confirmation.

The category error in cross-asset slots

If any slot has an Optional Ticker set and a non-zero weight, the blended pair is no longer "an RSI of the chart symbol." It is a composite that includes one or more outside markets. A zero-weight cross-asset slot can still draw and still participate in alignment counts, but it does not steer the blend.

This is a category error if you forget you turned it on. The blend's lime/red still draws above the chart, the alignment alerts still fire on confirmed bars, and the pane still looks like a tool that describes the chart. It does not. It describes a portfolio of RSI reads across whatever symbols you loaded, weighted by whatever weights you set.

Honest framing for a setup with cross-asset slots:

• The composite is a multi-context read. Treat it that way explicitly. Do not let yourself hear the lime/red as "the chart symbol is bullish."

• If a foreign symbol's market is closed during the chart's hours, that foreign slot is reporting a stale value. The composite is being driven by whichever slots are still updating.

• If a foreign symbol's correlation to the chart symbol changes (which it will, regularly), the composite's meaning changes too. There is no internal mechanism that detects this drift.

A good rule of thumb: if a colleague glanced at your pane and assumed the blend was about your chart symbol, would you be comfortable with that? If not, lower the cross-asset slot's weight or move it to weight zero so it stays as context only.

What weight zero actually does

A slot with weight zero is silent in the blend, but not silent everywhere. Knowing the asymmetry prevents the most consistent surprise on this page.

A weight-zero slot:

• Does not contribute to the blended RSI or the blended signal.

• Does still draw on the pane (unless the slot's Hide Plot is also set).

• Does still compute, which means it still consumes the script's compute budget.

• Does still fire its own per-slot bullish/bearish alerts.

• Does still count in the alignment alerts (All RSI Slots Bullish / All RSI Slots Bearish).

The last item is the one people miss. The alignment check is over enabled slots with available RSI values, not over weighted slots. A weight-zero context slot you set up so it would not "vote" still votes for alignment purposes once it has a usable RSI value. If that is not what you wanted, disable the slot rather than zero-weighting it.

What every-weight-zero produces

If every enabled slot has weight zero, the blended pair is NA. The blended RSI and the blended signal do not draw, and the blend's bullish/bearish alerts do not fire.

This is the script doing exactly what you asked it to. There is no implicit "fall back to equal weight" — if you told the indicator that no slot should steer the blend, the blend has nothing to compute. The slots themselves still draw and still fire their per-slot alerts.

If you see a missing blend on the pane, the cause is almost always one of:

• Every slot disabled.

• Every weight set to zero.

• A configuration where every contributing slot is currently NA (typically warm-up).

Master smoothing and the regime-change comfort trap

Master smoothing trades responsiveness for calm. That trade is honest, and it is a useful one when you want a slower regime read. It is also where readers most often quietly blind themselves.

Long master-smoothing lengths can hold the blended pair on one side of the midline for several bars after the underlying slots have already flipped. The smoothing pass is doing exactly what you asked. What it is doing is hiding the moment that just happened.

When this matters most:

• During regime transitions. The slot plots will show the flip; the smoothed blend will not, until the smoothing window catches up.

• When you have set alerts on the blended pair. The state alerts will keep firing in the prior direction during the lag.

• When you have set the overbought and oversold guides at default and the smoothed blend rarely reaches them anymore. The guides are now visual references for a series that does not extend that far in either direction.

The honest move when you suspect a transition: cross-check against the raw slot plots. They are the source. The blended pair, smoothed or not, is your composite of them.

Pinned values mean "at or beyond the frame," not "extreme"

The slot plots and the blended pair are clamped into 0..100. A line pinned at 0 or 100 is at or beyond the frame the indicator will report — not a measured extreme of the underlying series.

In practice, the slot lines rarely pin (RSI is by construction in 0..100). The blended pair can pin in unusual configurations — for example, a heavy weighting that pushes the weighted average outside the frame momentarily, then gets clamped back in. When you see a pin:

• Do not read it as "RSI hit X." Read it as "the value the indicator can report has run out of room."

• Look at the unsmoothed slot lines for the actual story.

• Consider whether your weighting is doing something you did not intend.

Things this indicator does not measure

It is worth listing these directly, because the pane reads as if it is reporting on the symbol.

• It does not measure price. RSI is a momentum derivative. The pane can be deeply lime while price is in a downtrend, and deeply red while price is in an uptrend. Whatever your composite says, the price chart is the rest of the picture.

• It does not measure volume, order flow, or liquidity. Whatever you are doing with these reads, you remain responsible for the rest of the market context.

• It does not measure the correlation between cross-asset slots. A composite of two correlated assets is different from a composite of two uncorrelated ones, and the indicator does not warn you when correlations drift.

• It does not measure session quality. A slot that updates only during a foreign symbol's open hours has nothing to say about the rest of the chart's day.

• It does not measure your weighting choices. You did. The composite is honest about reflecting them, not about whether they were sensible.

Things this indicator cannot alert on

The script exposes twenty-four alert conditions, and they cover slot bullish/bearish, blended bullish/bearish, and alignment. They do not cover:

• Crosses (the alerts are state alerts, not edge-triggered crosses).

• Threshold touches against overbought or oversold (the levels are visual guides, not alert sources).

• Midline crosses or directional changes by themselves.

• Anything intra-bar (alerts evaluate on confirmed bars only).

If you need any of these, you build them externally — either by composing alert conditions in TradingView's plot-based dialog using the indicator's plotted series, or by routing through automation that converts state into events.

Verification routines worth keeping

Three small habits that keep your trust in the pane calibrated. Each takes only a few minutes.

Once per setup change: cross-check the blend against a single-timeframe RSI on the chart. Add a stock RSI of comparable length and source to the chart's own pane. The blended RSI should not be wildly out of line with what a careful read of the stock RSI would tell you. If it is, your weighting or your slot configuration is producing a composite that no longer maps to a reasonable RSI of the symbol.

Once per session change: re-confirm On Bar Close? settings after holiday or session-boundary days. Slot rhythms can look different when bars come at different intervals; you want to reach for the switch from a remembered state, not a discovered one.

Once per major regime move: re-check whether master smoothing is still serving you. The length that produced a calm read in a flat regime can hide the next transition. If the slots are flipping and the master-smoothed blend is not, decide whether you want to update the smoothing or read the slots.

Posture when the pane feels persuasive

The most expensive trades you take with this tool are the ones where the pane looks clean and you do not pause to ask why it does. The composite is yours. The slots are yours. The weighting is yours. The on-bar-close switch is yours. The pane is reporting what you built, not what is true.

When the picture feels confident:

• Look at the visible slots and ask whether they are drifting together because the symbol is doing one thing or because they are correlated by construction.

• Check whether the alignment is a real disagreement between independent slots or an alignment of overlapping windows.

• Notice whether the master smoothing is doing the calming or the symbol is.

• Notice whether any cross-asset slot is doing the work that you might be reading as "the chart is doing the work."

• Ask how long it has been since you last changed a setting. A pane that has been persuasive since you configured it is a pane you have not stress-tested; markets change regimes, and configurations that were honest in one regime quietly drift into something else.

The point of the workbench is to give you an instrument honest enough that the doubt is also honest. If the pane is persuasive, the doubt is the next instrument you should use.

The trust boundary in one sentence

What you can trust on this pane is arithmetic — the slots are what they claim to be, the blend is the weighted average of the contributing slots, the clamp holds. What you cannot trust on this pane is meaning — whether the arithmetic is relevant to the decision you are about to make is your problem, not the instrument's, and no setting will change that.

Where to go next

• For the timing concepts that drive most of the limitations on this page, go to MTF and Repainting.

• For concrete routines that respect these boundaries, go to Workflows.

• For the underlying pipeline that produces the composite, go to For the Geeks.