Limitations & Trust Boundaries

This is the page you read when you are tempted to trust the pane more than you should. It is organized by damage potential, leading with the single misread most likely to cost a careful trader money and attention. Everything on this page is a truth about how the tool works. None of it is a disclaimer — the limits below are part of the instrument, not asterisks next to it.

The alignment trap, first

All MA Osc Slots Bullish and All MA Osc Slots Bearish sound like breadth. They are not breadth. They report that every enabled slot is currently on the same side of its own fast-vs-slow relationship. The claim that alignment makes is about your configuration, not about independent evidence.

Consider the shipped defaults. Three slots, all reading close on the chart's instrument, all using EMA(20) baselines and EMA(3) slow lines, at timeframes of 5m, 15m, and 60m. When the instrument is trending, the three slots will align almost by construction — they are three sampling cadences of the same measurement on the same price series. Treating their alignment as three independent votes is treating one observation as three. The statistical word for this situation is that the effective sample size is closer to one than to three; an alignment alert does not know the difference, but the reader has to.

A concrete way to see this: on a strong trend day on a liquid index, watch how often the shipped defaults fire All MA Osc Slots Bullish. The answer is "most of the session," and that frequency should bother you. Breadth alerts that fire constantly are not carrying breadth information. They are confirming the trivial fact that three sampling cadences of the same measurement agree with the measurement.

With three slots (and Base has exactly three), each slot is a larger share of the alignment claim than a slot is in CTX with more slots available. Fewer slots does not make this trap smaller; it makes every slot carry more weight in the claim. The honest fix is to differentiate the slots. Change source on one. Change baseline family or length on another. Consider giving one slot a materially different timeframe. Alignment across genuinely different measurements carries weight that alignment across three copies of the same measurement does not.

The practical mitigation before you route alignment alerts anywhere:

• If slots 01, 02, and 03 all use close with similar lengths, mentally treat the alignment reading as a single slot's reading, not as three.

• Differentiate at least one slot in a structural way — source, baseline family, baseline length — so that the slots have real reason to disagree sometimes.

• Run the configuration through a session where the instrument does something unusual (a gap, a spike, a midday news move) and see whether the slots disagreed at least sometimes. If they never did, alignment is not carrying information.

This trap sits at the top because a reader who internalizes this one limit will avoid most of the expensive mistakes the indicator can enable.

Boundary pinning is not intensity

The pane clamps into 0..100 by design. Readings that pin at 0 or 100 for bars at a time are the clamp doing its job.

A reading of 70 is not "70% stretched." A reading of 100 is not "twice as stretched as 50." The scale is a bounded transformation; differences between 70 and 100 do not scale linearly with underlying distance. Once the reading reaches a boundary, the tool is telling you that distance has grown beyond the sensitivity you asked it to resolve. That is useful — it means the instrument is deeply stretched relative to its own recent volatility. It does not mean the instrument is twice as stretched as it was a few bars ago at 70.

The failure mode is a reader who pushes ATR Sensitivity up until the pane pins more often, then reads the frequent pinning as frequent high-conviction stretch. The pane is pinning because the clamp has been asked to bite earlier, not because the instrument is more stretched. This misreading scales in both directions: a reader who pushes sensitivity down until the pane hovers near 50 will then read the flat reading as "calm market," when in fact the dial has been set below the range that would resolve any real stretch at all.

The audit move: if you see long periods of pinning, compare the pane's behavior at a lower sensitivity on the same instrument. If the pane at a lower sensitivity resolves useful information where your current sensitivity just pins, your current sensitivity is wrong for this instrument.

Sensitivity is not calibration to truth

ATR Sensitivity = 1.0 is the multiplier that leaves the ATR-normalized distance unchanged. It is a neutral position in the math; it is not a "calibrated" or "correct" sensitivity for any particular instrument. Different symbols have different volatility structures and different stretch regimes. What is a useful sensitivity on a major index future is often wrong on a low-float equity.

The reader has to audit. Pick a sensitivity that lets the pane spend useful time between 20 and 80 during the sessions you trade on the instruments you trade. If the pane rarely leaves the 45-55 band, sensitivity is too low to resolve anything. If the pane is at a boundary more often than not, sensitivity is too high. The right number is instrument-specific and session-specific. The tool cannot pick it for you without lying.

Weight zero is not "slot off"

Blended Weight = 0 on an enabled slot does not turn the slot off. The slot still computes. The slot still plots (unless Hide MA Plot is also on). The slot still fires its per-slot alerts. The slot still participates in the alignment alerts. All weight zero does is remove the slot from the weighted average that produces the blend fast and blend slow lines.

This matters because the observer-slot pattern in Workflows uses weight zero deliberately, and a reader who misreads weight zero as "off" will either be surprised by alerts or misconfigure the pane when they revisit it. If you want the slot off — no plot, no alerts, no contribution — use the slot's Enable toggle.

Master smoothing trades noise for lag

Master smoothing adds a final MA pass on the already-blended fast and slow lines. It makes the blend calmer. It also makes the blend later to tell you a regime flipped.

Neither effect is hidden. Both are real. A reader who enables master smoothing and then builds confidence because the pane "looks cleaner" is buying calm at the cost of responsiveness, without knowing how much responsiveness they bought. The cost is specific to the length: short lengths are subtle delays that show up mostly on fast reversals; longer lengths can delay regime flips enough to be visible in hindsight.

The specific failure mode worth naming: a reader enables master smoothing with a long length on a choppy session, watches the pane, likes the calmer reading, and then gets hit on the first clean regime change because the smoothed blend was still calling the old regime. Compare smoothed vs unsmoothed side by side before deciding whether the tradeoff is worth it for your setup.

Chart-timeframe guardrail

When the chart timeframe sits above any enabled slot's configured timeframe, the slot does not resolve upward to the chart timeframe. The script raises a named runtime error and stops drawing until the invalid slot is fixed. With defaults of 5, 15, and 60, a 1H chart allows the 60-minute slot but rejects the 5-minute and 15-minute slots. That hard stop is deliberate.

This is not a bug in the tool; it is the tool refusing to silently fake a lower-timeframe read from a higher-timeframe chart. The honest move is to reconfigure: either go back to a lower chart timeframe where the configured slots actually stack, or raise every enabled slot timeframe so it is at or above the chart. Defaults of 5, 15, 60 want a chart at 1m or 5m. A 4H chart wants slots at something like 240, D, W if you want meaningful separation.

Symptom-wise, this looks like the named MA 0N timeframe cannot be lower than the chart timeframe error. Troubleshooting has the symptom-to-cause framing.

Warm-up behavior is honest and brief

Until the Fast and Slow values have enough history to be useful, early bars can include na values or fallback behavior. The important point is simpler than the internals: those first bars are warm-up, not the settled Fast-vs-Slow regime the pane uses once values are populated. EMA(3) settles quickly; longer slow lengths take longer.

This matters mostly for readers who load the indicator and immediately start interpreting the first few seconds. Let the pane settle.

Overbought and oversold are zones, not triggers

The 30 and 70 reference lines tell you the reading is in a part of the range where stretch is usually meaningful. They do not mean reverse. The script does not fire alerts when they are crossed. Treating them as thresholds — buying when the reading crosses 30 from below, selling when it crosses 70 from above — is transferring RSI habits onto a pane that is not RSI.

The honest use of the reference band: use it as a framing device that tells you where in the range the reading currently lives. The regime information you want is in the fast-vs-slow relationship, not in the position of either line relative to 30 or 70.

The tool cannot know your intent

A few limits that exist because the pane describes a state, not a plan:

• It cannot tell you whether a stretch will mean-revert or extend.

• It cannot tell you how old the current regime is.

• It cannot tell you whether you should size up, size down, or pass.

• It cannot replace your own rules about when to act and when to sit.

These are not failings. They are what "stretch gauge, not signal service" means in practice. The page is naming them so the pane does not quietly get handed decisions it cannot make.

When to set the pane aside

Part of using an instrument well is knowing the conditions under which it stops earning its place on the screen. For this one:

• During the first handful of bars after load. The Fast and Slow values may still be warming up. Whatever color the pane shows you is not the mature read yet.

• When a configured slot timeframe is below the chart timeframe. The pane will stop with a named runtime error. Either move the chart lower or reconfigure every enabled slot so it is at or above the chart timeframe.

• Immediately after you changed a setting. The knobs that change pane shape (ATR Sensitivity, master smoothing, slot weights) take a session of observation before you can tell what the new configuration is doing. Do not draw conclusions in the first half hour after a change.

• When the instrument just gapped or absorbed a news print. The ATR window has a sudden new value in it. The normalization is going to take a while to stabilize. The pane is not wrong during that window; it is reporting honest but fast-changing math. Trust it less during the first session after a major volatility shift.

Knowing when not to read the pane is a different skill than knowing how to read it, and it takes longer to develop. This section names the obvious cases so you can spend your attention on the subtle ones.

Honest pointers beyond this page

• If a limitation above is a gap for you and is addressed by more configuration surface: CTX has ten slots, per-slot symbol overrides, per-slot repaint, and Power User parameter blocks.

• If a limitation above is about how you are interpreting the pane rather than what the pane is doing: Visuals & Logic and For the Geeks are where the mental model lives.

• If a limitation above is causing a specific symptom on your chart: Troubleshooting.

Where to go next

• You need to understand how the tool is actually building the pane: For the Geeks.

• You need to configure around one of the limits above: Settings and Workflows.

• You need a quick diagnostic for a specific symptom: Troubleshooting.