Limitations & Trust Boundaries

This page is a map of what you are choosing when you put Axiom MA on a chart. It is not a legal disclaimer and it is not a softer sales pitch. The goal is to name, precisely, what the tool is reliably good at, what it cannot tell you, and the handful of places readers habitually give the tool more credit — or less credit — than it has earned.

Read it once before you trust anything here in production. Read it again when something on the chart surprised you.

What this tool is reliably good at

Axiom MA is a compact instrument that does a small number of things carefully.

• It stacks up to three moving averages from different timeframes on one overlay chart, with their computations routed through a real higher-timeframe context rather than a re-sampling of the chart bar.

• It gives you a single switch that decides whether every slot reports the last confirmed higher-timeframe bar or the live higher-timeframe bar, and it tells you which one you are choosing in plain terms.

• It refuses to render if a slot's timeframe is below the chart's, and it names the offending slot rather than failing silently.

• It combines the enabled, non-zero-weight slots into one weighted composite line, with a declared averaging rule and a declared trend-voting rule.

• It exposes ten alert conditions that fire on closed chart bars, and two hidden count plots that let you enrich alert messages with live slot counts.

Within that shape, the tool is dependable. The numbers on the chart match the rules above. The alerts fire on the conditions above. The runtime behaviour matches the documented behaviour.

What this tool cannot tell you

The short version: it cannot tell you anything it was not designed to compute.

• It cannot tell you whether price is about to turn.

• It cannot tell you whether the current move is "real," "tradable," or "a trap."

• It cannot tell you whether your timeframe choice, your length choice, your Trend Length choice, or your weights are appropriate for the instrument, your account size, your session, or your process.

• It cannot separate the signal from the noise in your own chart; that is the job of the reader, the process, and the broader methodology the reader is running on top of the chart.

• It cannot compensate for a faulty mental model of what "higher-timeframe trend" actually means to you. If you do not know what you expect the 60-minute slot to tell you before you read it, the slot cannot tell it to you.

None of these are bugs. They are the edges of the instrument.

Blend honesty — what you are reading when you read the blended line

The blended line is the most over-read element on this tool. The rule is disclosable; the failure modes of reading it carelessly are also disclosable.

• Hidden-plot contribution. A slot with Hide MA NN Plot set to true is not drawn on the chart but is still in the blend — its value is accumulated, its trend state is counted in the vote. If you want a slot out of the blend, set its Blended Weight: to zero. See Visuals & Logic and Settings.

• Weight-zero visibility. A slot with Blended Weight of zero is skipped by the blend accumulator but still renders on the chart and still fires its own per-slot trend alert. Visibility and blend participation are independent.

• Weight-majority trend rule. The blended colour is decided by a vote, not by the slope of the blended line. Sum the voting weights on each side; the heavier side wins. A flat-looking blended line can read uptrend or downtrend because the line's slope is not what decides the colour.

• Tie-break to uptrend. A tied vote resolves to uptrend. This is an asymmetry in the rule and it is disclosed on purpose — if you are expecting "tied means no read," you will misread tied bars as bullish.

• Blend colour does not fade. The blended line is lime or red. There is no middle, dimmed, or neutral state. A pale-looking blend is a chart-display artefact, not a state the tool is reporting.

• No usable positive blend weight is not a read. If the blend is enabled while every enabled slot is weight 0, or if unusual negative weights leave no usable positive total weight, the current source can leave the blend at its default 0/down state. That is configuration failure, not market information.

• The blend is not "smarter" than its components. It is a weighted composite under a declared rule. It cannot see anything the slots do not see. It can, however, mask disagreement between the slots — which is why the page you are reading keeps asking you to read the slots underneath.

A practical reframing that has held up in real sessions: the blend tells you where your weights agree, nothing more. Not where the market is, not where the regime is, not where you should enter. Where your weights agree — and only on the voting rule the tool documents.

Alert honesty — what a fired alert does and does not confirm

The Alerts page carries the detail. A condensed set of trust boundaries worth re-stating here because they are a common source of over-trust:

• Bar-close-gated on the chart timeframe. Every alert fires only on closed chart bars. It will not fire intra-bar.

• HTF state may be live under OFF. A chart-bar-confirmed alert can still reflect a live higher-timeframe bar when On Bar Close? is OFF. The chart-bar gate is not a higher-timeframe-bar gate; do not infer the latter from the former.

• Alignment counts hidden slots once they have a value. The alignment alerts require every enabled slot with a non-na MA value to agree, including any hidden-plot slot. If you want a hidden slot excluded from alignment once valid, disable it instead of hiding it.

• Trend state is not a trade signal. Per-slot and blended trend alerts report computed smoothed-state classifications on specific timeframes. They are context notifications, not triggers. Wiring them into execution as triggers is a misuse pattern documented in Workflows.

• Configuration-dependent. Every alert is only as meaningful as the settings it is actually running with. Do not assume a saved TradingView alert changed just because you changed chart inputs later; verify or recreate the alert when settings matter.

The version surface — Base is narrower by design

Axiom indicator families use a Base → CTX → STR progression when the full set exists. Base is the default, untagged version that this manual documents. The other versions exist because different reads want different surface area:

• CTX expands Base into a broader context tool: ten slots, per-slot repaint posture, cross-ticker reads, per-slot colour inputs, and the Pro-library MA palette.

• STR sits after CTX where it exists: it keeps the stronger per-slot control pattern and spends extra processing on structure around the blended output, which is why STR carries fewer slots than CTX.

Base does not ship those features. That is a version decision, not a defect. The Base surface is narrower on purpose so the instrument is faster to read and harder to miscalibrate. If your read genuinely needs per-slot repaint posture or cross-ticker, Base is not the tool; see the MA CTX pack. If three slots, one global switch, the Lite MA palette, and fixed colours fit the work you actually do, Base is not a "stripped-down" CTX — it is the right fit.

A misread worth naming: moving from Base to CTX does not make the read automatically better; it makes the surface wider. If a CTX user is not actually using the per-slot posture or cross-ticker, they are paying attention costs for features that are not doing work. Choose the version on purpose, in both directions.

The defaults — what they are, what they are not

The shipped configuration — three slots at 5 / 15 / 60 minutes, SMA-20 on each, Trend Length 3, equal 33.3 weights, line widths 2 / 2 / 2 / 3 — exists because it loads cleanly on a 1-minute chart of a liquid regular-session symbol and demonstrates the tool's mechanics in plain terms.

• The defaults are not an endorsement of 5 / 15 / 60 for any instrument or trading style.

• The defaults are not optimised. Nothing was back-tested to produce them.

• The defaults are not ranked. Equal weights means every slot gets an equal vote in the blend — that is a statement, not a neutral posture.

• The defaults are the first chart. They are meant to be a known starting point. Your real configuration is the one you choose.

If a page in this pack ever seems to drift toward framing the defaults as "recommended," treat it as a drafting error and ignore the framing. The reader-facing product language this pack is bound to explicitly refuses to rank defaults.

Practical trust boundaries, stated directly

A list to keep handy. Each item is a place the reader is most likely to over-trust, paired with the accurate framing.

• "The 60-minute slot is accurate because it is on a higher timeframe." The slot is as accurate as its source, its length, its Trend Length, the On Bar Close? posture, and whether the 60-minute bar is live or confirmed. Accuracy on a moving average is a function of inputs and posture, not of timeframe.

• "The blended line captures what three timeframes agree on." Only when the weights are equal and the trend states agree. Under heavy-weight or disagreement, the blend's colour follows the weighted majority, not the unweighted majority.

• "Under On Bar Close? OFF, the 5-minute slot is still confirmed — I only wanted to loosen the 60m." Under OFF, every slot is live-HTF-exposed, including the 5-minute slot. The switch's scope is global.

• "All MA Slots Uptrend did not fire — the tool must be missing the alignment." The alignment alert counts every enabled slot with a non-na MA value. If a hidden valid slot is disagreeing, the alert will not fire. If a slower slot is still na, it is not in the active set yet, which can make the count smaller than the number of enabled slots.

• "A per-slot trend alert firing is a turn." It is not. Per-slot trend alerts fire on every closed chart bar while the slot is in that trend state — they report a state, not a transition.

• "Hiding a slot takes it out of the alignment read." It does not. Hidden is visual-only. Disabling is total.

• "Weight zero is the same as disabling." No. Weight zero removes the slot from the blend; the slot is still drawn (unless you hide it), still computes its trend, and still fires per-slot alerts.

• "The tool is repaint-proof under ON." Under ON, the slot does not move on the live higher-timeframe bar. That is narrower than "repaint-proof." Outside the ON posture, live HTF reads can drift until the HTF bar closes. See MTF & Repainting for the full posture discussion.

• "Absence of cross-ticker / per-slot repaint / colour inputs is a defect." Those features live in CTX. Base holds them back on purpose. See the version note above.

If you are unsure whether to trust a specific read

A sequence of checks that tends to resolve most moments of ambiguity:

1. Open TradingView's Data Window (Alt+D) and check the values of MA 01, MA 02, MA 03, MA Blended, Active Uptrend Count, and Active Downtrend Count at the bar you are reading.

2. Confirm On Bar Close?'s state in the inputs dialog. Confirm every slot's TimeFrame:, Length:, Trend Length:, and Blended Weight:.

3. If the blend's colour surprises you, compare the slots' trend states and weights. A lime blend with two down-voting slots means one heavy-weight slot is carrying the vote; a red blend at tied weights is not possible (tie goes to up).

4. If an alignment alert did not fire when you expected it to, check whether a hidden valid slot is still counted and disagreeing. If a slot is na on early bars, it is not counted yet, so compare the active counts instead of assuming all enabled slots are in the vote.

5. If a higher-timeframe line is flickering, check On Bar Close?. If it is OFF, the slot is reading live HTF; that is the posture's cost.

Troubleshooting maps the most common symptom-to-cause cases directly.

Where to go next

• Visuals & Logic for how the blend's rule looks on an actual chart.

• MTF & Repainting for the repaint posture's full mechanism and verification walkthrough.

• Workflows for the routines and anti-patterns that put these boundaries into practice.