Limitations and Trust Boundaries

This is the page the rest of the pack circles back to. Every honest indicator has places where it will mislead a careless reader, and a wider workbench like this one has more of them, not fewer. The point of this page is not to discourage trust. The point is to earn trust by being specific about where the trust has to stop — so that the reader who commits real money on the basis of what they see is doing so with their eyes open.

If you read one page in this pack when the work feels serious, read this one.

The trust boundary, stated plainly

The blended K and blended D are a weighted average across enabled slots that have non-zero weight and a usable K value. The weights are yours. The slot roster is yours. The source, the timeframe, the Optional Ticker on any slot are yours. The composite is a picture the reader draws with the knobs. It is not a consensus. It is not an oracle. It is a summary you designed.

That one paragraph is the root of every limitation below. Each over-trust risk on this page is a specific way readers forget that boundary under pressure. Naming them out loud does not make the boundary go away; it makes the boundary easier to hold when a live chart is making it inconvenient to remember.

Over-trust risk 1 — "The blended pair is a consensus"

This is the most common misread in the category, and the one the wider workbench makes easier to fall into. Because the blend combines several slots into a single line, the line looks authoritative. It is not. It is a weighted average of numbers the reader chose to combine, and it carries only as much meaning as that choice supports.

Three specific consequences worth holding in working memory:

• Equal weights across correlated slots do not produce consensus. Three slots of the same symbol on 5m, 15m, and 60m are three overlapping windows on one price series. Their agreement overstates what is happening in the market, not because any slot is wrong, but because they are autocorrelated: each slot is reading a moving slice that includes most of the bars the others are reading. "All three agree" in this configuration means "the same symbol agrees with itself at three scales." That is often the case; treating it as three independent witnesses to the same event is the error.

• One heavy-weighted slot steers the blend. If slot 01 has weight 80 and slots 02–03 have weight 10 each, the blend is mostly slot 01 with two smaller corrections. A reader who watches the blended color without watching the weights can quietly reduce their composite to one dominant slot and not know it.

• Cross-asset slots change what the blend means. A slot with Optional Ticker set is a different market entirely. The blend that includes it is a multi-market composite. See Over-trust risk 3 for the full framing.

The mature read on every blended value is: "what is the composition of weight and roster right now, and which slots are doing the talking?" If you cannot answer that question about your own pane in under five seconds, the blend is steering you and you are not steering the blend. The fix is never to read harder; the fix is to know the weights well enough that the steering slots are audible at a glance. One way to build that reflex: write your current weights on a sticky note and keep it by the monitor for two weeks. Most readers find they no longer need the note after the first few sessions; the reflex has taken hold. The readers who still need it after two weeks are telling themselves something useful about how well they know their own configuration.

There is a second-order trap here worth naming. A reader who knows the blend is a weighted average will sometimes reason: "I know the blend is my weighting, so it is honest for me, which means I can act on it freely." This is wrong in a subtle way. The blend is honest to your weighting; that does not make it a good weighting for what you are trying to do. A weighting you chose three months ago may no longer match the market regime or the trading horizon you are currently working in. The trust boundary is not "the composite is mine, therefore the composite is right." It is "the composite is mine, therefore I am responsible for whether it currently matches what I am trying to read." Owning the composite includes owning its staleness.

Over-trust risk 2 — "The on-bar-close switch is a fix for repainting"

It is not. There is no fix for repainting. A live higher-timeframe value moves because the bar is forming. The switch exposes that tradeoff per slot and asks the reader to choose:

• Switch on, take the lag. The slot returns the confirmed previous higher-timeframe bar. The value does not repaint. The cost is that the slot is one higher-timeframe bar behind the market.

• Switch off, take the movement. The slot returns the live higher-timeframe bar. The value is responsive. The cost is that the slot will be different the next time you look at the same bar.

Turning On Bar Close? off across every slot to chase responsiveness and then acting on slot values that are still moving is the specific way this risk hurts people. The switch was built to make that choice uncomfortable; do not bypass the discomfort.

For a full treatment of what each side looks like on a real chart, see MTF and Repainting.

Over-trust risk 3 — "The blend describes the chart symbol"

Not if any slot has Optional Ticker set. When Optional Ticker is populated on a slot, the slot reads its entire stochastic context — the source, the high, and the low — from the chosen symbol. That slot is no longer a read of the chart; it is a read of the other market.

The blend then is not a stochastic of the chart symbol. It is a weighted mixture of stochastic reads across different markets. The pane will not tell you that. The color rule will still be lime for blended K above blended D and red otherwise, and the 0..100 frame will still hold, and everything will look numerically correct. What will be wrong is the reading.

The honest frame for every Optional Ticker slot:

• Write down why it is there. A cross-asset slot for benchmark context. A slot for a correlated pair. A slot for a sector proxy. Whatever the reason, name it.

• Know the market it represents. If you cannot say what the other symbol's session, typical intraday range, and current regime are, the slot is adding mystery, not context.

• Read the blend as multi-market. Every time. The blend line does not reshape itself to make this obvious; the reader has to.

Over-trust risk 4 — "Alignment means multiple slots agree, weighted by how much they matter"

It does not. The alignment alerts — All Stoch Slots Bullish, All Stoch Slots Bearish — count every enabled slot regardless of weight, and they require strict K-versus-D direction. A slot with weight zero still counts toward alignment. A slot with weight 99 still counts exactly once. A slot whose K equals D is neither bullish nor bearish for alignment.

Three specific sharp edges:

• A single enabled slot satisfies alignment trivially. All Stoch Slots Bullish is true whenever that one slot is bullish. If you rely on alignment as a filter, enabling fewer than two slots makes the filter a pass-through.

• Weight-zero slots break alignment. If you keep a zero-weighted slot enabled for visual reference while the other slots are the ones you want the alignment to track, a disagreement from the zero-weighted slot will break the alignment. The alert will not fire even when the slots you actually care about agree.

• Enable is the mechanism for alignment membership, not Weight. Disable the slot if it should not count. Zero-weighting is the wrong tool for this particular job.

For the full framing, Alerts carries the state-versus-cross and enabled-versus-weighted distinctions.

Over-trust risk 5 — "A pinned line is a strong signal"

A slot K or a blended K that sits pinned at 0 or 100 for one or more bars is not a quantified extreme. It is the clamp holding the native 0..100 stochastic frame. The reading is "at or beyond the frame this computation can represent," not "the most extreme reading this instrument has ever seen."

Two consequences:

• A line pinned at 100 is not stronger than a line at 95. The pane cannot tell you whether "at or beyond the frame" is 101 or 250. It can only draw 100.

• Pinned lines are more likely with aggressive smoothing choices (short K length, hot MA tuning) than with default settings. If you are seeing pinning often, the pinning is as much about your tunings as about the market.

Over-trust risk 6 — "Master smoothing makes the composite cleaner"

Master smoothing makes the composite calmer, which is not the same as cleaner. The two senses diverge during regime changes, and the divergence is where the harm happens.

• A long master length holds the blended pair on one side of its midline for several bars after the underlying slots have already turned. On a smooth pane that looks like fidelity. During an actual regime change, it means the composite is lying about the slots. The slots are the ground truth; the composite is a processed picture of the slots.

• Alerts configured against the blended state will stay silent across exactly the regime-change bars they were wired to catch. The alerts are not malfunctioning; the master pass is holding the state. The effect is the same.

• Any recommendation in this pack that mentions master smoothing will pair it with this risk. If you do not see the pairing, the recommendation is incomplete.

The working discipline: when master smoothing is on at any meaningful length, cross-check the raw slot lines every time the blend feels sluggish. If the slots have turned and the blend has not, trust the slots.

Over-trust risk 7 — "The 80/20 guides are triggers"

They are visual landmarks. The numeric values on the Overbought Level and Oversold Level inputs draw dashed guide lines on the pane. Nothing changes in the indicator's output when a line crosses a guide.

Specifically:

• No slot line changes color when it crosses 80 or 20. The color rule is entirely about K-versus-D.

• The blended K does not change color when it crosses 80 or 20. Same rule.

• No alert fires on a 80/20 cross. The 24 alert conditions are state alerts on K-versus-D (per slot, blended, or aligned across slots). There is no overbought alert or oversold alert in the script.

If you want an 80 or 20 notification, you need to build one externally against the plotted values. Inventing one in the manual that the script does not actually fire would be a generous-looking form of lying.

Over-trust risk 8 — "The midline at 50 means something special"

It is a visual reference for "the middle of the stochastic range." It is not a zero-crossing signal in the way a MACD histogram might be; the stochastic is not constructed around a zero line. It is a useful landmark because human eyes read the upper half versus lower half of a 0..100 range quickly, and 50 is the landmark that makes that read easy.

It does not color any line. It does not feed any alert. If you treat it as a trigger, you will be wrong more than you are right.

Over-trust risk 9 — "The tool will produce the same reading for everyone"

No. Because the tool is a workbench, different readers produce different composites from the same pane. Two traders with the same slot roster but different weightings will see different blended lines. Two traders with the same weighting but different K and D MA choices on a slot will see different slot colors, and sometimes different alignment counts.

This is the point. An instrument you can calibrate to your own process produces a reading that reflects your process. A reading that reflects your process is more useful than a reading that is "the same for everyone," because "the same for everyone" is usually also "optimized for nobody."

The trust boundary that comes with this: you are responsible for your configuration. If you took a screenshot of someone else's pane and copied their settings, you are reading their weighted composite, not yours. That may or may not be what you intended.

Over-trust risk 10 — "There is a correct set of settings"

There is not. The defaults are a restrained starting point. Every other configuration is a tradeoff. "No best settings" is not a slogan; it is the structural truth about a workbench whose correct configuration depends on what the reader is building with it.

Practical consequences:

• Do not ship a configuration until you have used it long enough to form an expectation of its behavior on the market you are actually reading.

• Do not borrow someone else's settings file without first reading what each knob does on Settings. A file that works for one trader on one market and one timeframe is not a default.

• Do not freeze a configuration once you find one you like. Markets drift, regimes change, your own reading matures. Revisit the configuration periodically.

The tool is not going to decide for you. That is not laziness on the product's part. It is the product's honest answer to the question "what should I use?" The answer is "whatever you can defend — and be prepared to change it when the conditions you defended it against change."

What the instrument can drift on

A short, honest list of things the pack cannot remove, only name:

• TradingView data. The higher-timeframe values come from TradingView. If TradingView is late, or delivers a corrected bar, the slot is late or corrected with it.

• Session calendars for cross-asset slots. Changes in a foreign symbol's session hours (for example, an exchange changing its schedule) propagate into the slot's behavior.

• MA library evolution. The shared Axiom MA library is versioned. If a new version of the library changes the behavior of a specific MA family — for example, refining a warm-up period — a configuration that worked identically before a library bump may behave slightly differently afterward. The pack will note meaningful library-level changes in release notes when they occur.

• Extreme market conditions. A security with a halt, a gap, or an unusual session can push stochastic inputs into edges the smoothing chain will handle correctly but may represent with pinned values or unusual settling behavior. The clamp will still hold; the reading will still be honest. The reader should be aware that the pane under extreme conditions will look different from the pane during normal tape.

None of these is a defect in the indicator. All of them are constraints any instrument of this kind lives inside.

What the instrument will not do

• Predict the future. The 0..100 frame is about the past; it cannot tell you the next tick.

• Filter noise for you. The smoothing choices are yours; noise rejection is a consequence of your tuning.

• Prevent you from enabling a dangerous configuration. The On Bar Close? switch is clearly labeled; the Optional Ticker is clearly labeled; no combination is blocked because the tool assumes the reader wanted to try it.

• Suppress redundant alerts. State alerts fire every bar the state holds; the downstream alerting service is where suppression belongs.

The minimum verifications before committing a configuration

If you are about to depend on a configuration in live trading, these four verifications are the cheapest insurance the pack can offer. Do them once, per configuration, before you wire live alerts or size up positions on the reading.

1. Push one slot's K Length to an extreme (2, then 50). Confirm the line stays inside 0..100 under both settings. Return to 14.

2. Toggle On Bar Close? for the slowest higher-timeframe slot. Watch one full higher-timeframe bar on both sides. Confirm the behavior matches the descriptions on MTF and Repainting.

3. Set one slot's weight to zero temporarily and note how the blend reshapes. Return to your configured weight. You now know which slot is steering how much.

4. If any slot has Optional Ticker set, toggle it off and on once, and observe how the slot's K moves. Confirm the slot is behaving according to the other symbol's price action, not the chart's.

Four minutes. Every configuration. Forever.

A second pass through this page, one quarter from now

The limitations on this page are not things a reader learns once and owns forever. They are things a reader learns, internalizes, and then quietly forgets as the pane becomes familiar and the trades become routine. The familiarity is the exact condition under which the limitations begin to bite again. A pane you have read for a month stops feeling like a composite you authored and starts feeling like "the reading" — which is how over-trust risk 1 comes back without announcing itself.

Set a reminder for three months out. Re-read this page then. You will find at least one risk that has quietly drifted from "I remember this" to "I do not remember when I last thought about this." That is not a failure. That is the texture of how any working tool becomes a working habit, and the work of staying honest with the tool is the work of periodically dragging the limitations back into view.

If that reminder never gets set, no amount of careful wording on this page will hold the boundary for you. A manual can describe a trust boundary; only a reader's habit can keep the boundary intact.

Where to go next

• For the mechanism behind each limitation, return to Visuals and Logic and MTF and Repainting.

• For the pipeline the limitations rest on, see For the Geeks.

• For routines that respect these boundaries, see Workflows.

• For the questions a reader is most likely to confuse with defects, see Troubleshooting.