Visuals & Logic

This page is about reading the chart. Everything on it — every line, every color, every fill, every state transition — is a computed condition about a lookback you configured, not a claim about what the market is going to do. The job of this page is to make every element specific enough that you can look at a chart mid-session, name what you are seeing, and know what a change in that element actually means.

Treat this as the page that teaches the grammar. Workflows teaches sentences; this page teaches what each word can and cannot say on its own.

Every element on the chart, named

When Axiom DC is running cleanly with the shipped defaults, the chart carries seven kinds of visible element.

Slot upper (per slot). The highest high across the slot's Length bars, read at the slot's chosen timeframe on the chart symbol. Drawn in the slot's color at 50% opacity. Not smoothed. A fresh swing high on the slot's timeframe moves this line directly when the slot's bar acknowledges it (immediately under On Bar Close? = OFF; on the next higher-timeframe close under ON).

Slot lower (per slot). The mirror of upper. Lowest low across Length, same timeframe, same color-at-half-opacity, also raw.

Slot basis (per slot). The midpoint of upper and lower, smoothed through the slot's Type MA at Basis MA Length. Drawn in the slot's color at full opacity — the slot's basis is the "darker" of the three same-color lines. At Basis MA Length = 1 with the default SMA, the basis is effectively the raw midpoint because a one-sample MA is the sample itself. Above 1, the length-based basis choices lag the midpoint in the way MA smoothing lags its input. SWMA is the Lite-library exception because it uses TradingView's fixed SWMA behavior rather than the user length.

Blended upper. A red line drawn by the blended channel. Equals the weighted average of the enabled, non-zero-weight slots' uppers. Line width is controlled by the DC Blended group.

Blended lower. A red line, mirror of blended upper. Same weighting rule.

Blended fill. A translucent red shaded region between blended upper and blended lower. Purely a visual orientation cue — a quick way to see the extent of the composite band. It carries no regime meaning and is not a "zone" in the structural sense.

Blended basis. A lime-green line drawn by the blended channel. Equals the weighted average of the enabled, non-zero-weight slots' bases. The most common use is a quick read on where the weighted middle of the stack sits, not as a price target.

The color split, and why it matters

Each slot owns a color — teal for Slot 01, blue for Slot 02, purple for Slot 03 — and each color is drawn twice: at half opacity for the channel bounds and at full opacity for the basis. The blended channel uses red for the bounds and lime for the basis to stand apart from any single slot.

When the stack is tight and the slots are close to each other, the color split is how you pick out a basis from an outer bound without squinting. If you cannot tell a basis from a bound on your chart, raise the line widths or disable a slot until the stack reads cleanly.

The smoothing asymmetry — the most commonly missed structural fact

Only the basis is smoothed. The upper and lower channel lines on every slot are raw Donchian levels — the highest high and the lowest low across Length. Basis MA Length and Basis MA Type do not touch them.

Why this matters in practice: you cannot smooth your way out of a bad Length:. If the outer channel looks ragged, the cause is the Donchian window, not the basis smoothing, and no amount of raising Basis MA Length will calm the bounds. Conversely, if the bounds look fine but the basis is too jittery for your read, raising Basis MA Length will calm the basis without changing the channel width at all.

Stepping Basis MA Length from 1 to 5 to 20 on a chart, with Length: held fixed and a length-based MA selected, shows this directly: the basis curve becomes progressively calmer while the upper and lower lines stay exactly where they were. Try it once. It will save you from reaching for the wrong knob later.

State changes that matter

Not every pixel that moves on the chart is a meaningful state change. The ones that are, for this indicator:

• Basis trend state (per slot). The slot's basis is rising or falling, where "rising" and "falling" are defined by the Basis Trend Length lookback on the smoothed basis at that slot's timeframe. This is what the per-slot basis-trend alerts fire on. A flip is one visible kind of state change, but the alert condition is the state itself, not only the transition.

• Price crossing a slot basis. A slot's basis sits somewhere in the band. Price moving from below to above it is a common read for that slot's near-term bias. The tool does not fire an alert on this crossing; you read it from the chart.

• Price touching or piercing a slot's upper or lower. A touch of the raw Donchian extreme on the slot's timeframe. Again, no alert — it is a visual read.

• Blended basis trend state. The blended basis direction is resolved by a weight-majority vote across enabled, non-zero-weight slots. The blended alerts fire while that state matches the named direction on a confirmed chart bar; they are not limited to the transition bar.

• Alignment event. Every enabled slot's basis trend agrees on direction. This fires the alignment alert. Hidden-plot slots still count here — see below.

Reading the blended channel: agreement vs cancellation

The blended channel is a weighted composite of whichever slots the reader enabled with non-zero weight, once those slots have valid values. It averages agreement and averages disagreement by the same rule. During early warmup, missing slot values can be zero-filled to keep the composite numeric, so the blend is not a trustworthy read until the contributing slot lines have come in. A narrow blended band can mean two very different things:

1. Agreement. Every contributing slot's band happens to be tight, or the slots' bands happen to overlap in a narrow region. The narrow blend is a real signal that the timeframes agree on a tight range.

2. Cancellation. One contributing slot is wide, another is tight, and the weighted average lands inside the tight slot. The blend looks narrow because the narrow slot carries the heavier weight, not because the whole stack agrees.

These two cases read the same on the blended channel alone. They read very differently on the slots underneath.

A worked example

Take the shipped defaults on a 1-minute chart: Slot 01 at 5m with weight 40, Slot 02 at 15m with weight 35, Slot 03 at 60m with weight 25.

Suppose Slot 01's band is very tight — the last twenty 5-minute bars have not ranged much. Slots 02 and 03 are wider. The blended upper and lower are:

• Blended upper = (40 × Slot01.upper + 35 × Slot02.upper + 25 × Slot03.upper) / 100

• Blended lower = (40 × Slot01.lower + 35 × Slot02.lower + 25 × Slot03.lower) / 100

With 40 weight on the tight slot, the blend will sit closer to Slot 01 than to either of the wider slots. The blended channel may look narrow. A reader glancing only at the red band might conclude "low volatility." The fifteen- and sixty-minute slots are saying something different, and they are visible on the chart — you just have to look at them.

The lesson is simple and load-bearing: the blended channel is a weighted summary of what you told it to summarize, at the weights you gave it. It is not a consensus and it is not a regime classifier. Treat it as a composite you authored, not as the indicator's answer.

A reading habit that prevents the agreement-vs-cancellation mistake: any time the blend draws your attention, glance at the three slot colors underneath before you act on the blend. Three questions are enough. Are the slots agreeing on direction? Are their bands overlapping in the region the blend is drawing? Is the weight doing what you thought it was doing? Thirty seconds of that check will catch most of the cancellation traps before the next bar closes.

Hidden plots still count — the quietest trap on the indicator

Hide DC NN Plot suppresses a slot's three lines from the chart. That is all it does. The slot still computes, still contributes to the blended channel at its current weight, and still participates in alignment alerts.

This is a trap because the blend can move or an alignment alert can fire, and the reader cannot point at a slot on the chart that "caused" it. Nothing caused it on-chart — the cause is on-slot, just not on the plot.

If the intent is to take a slot out of the blend:

• Set Blended Weight to 0 to keep the slot visible but remove it from the vote. The slot's three lines stay drawn; the blended channel recomposes without it.

• Turn Enable DC NN off to remove the slot entirely. No computation, no plot, no blend contribution, no participation in alerts.

Hiding a slot's plot is for the reader who wants a cleaner chart without changing the math. Know that you are choosing the cleaner chart, not the cleaner math.

The tie-to-rising rule on a three-slot stack

The blended basis-trend direction is decided by weight, not by count. The tool sums the total weight voting rising and the total weight voting falling, and reports rising when the rising total is greater than or equal to the falling total. A tie resolves as rising.

On a three-slot stack, ties are realistic rather than theoretical. A few common configurations that can produce one:

• Equal weights (for example 20 / 20 / 20) with a 1-versus-1 split on basis trend and a third slot at na. The two contributing slots tie; the rule resolves rising.

• Unequal weights that happen to match. Slot 01 at 40 voting rising, Slot 02 at 40 voting falling, Slot 03 at 20 voting rising. Uptrend weight = 60, downtrend weight = 40, so this resolves rising on the majority — not a tie, but worth pausing to confirm the vote. Change the third slot to falling and the weights become 40-vs-60 falling; the blend reports falling.

The >= in the tie-break is a deterministic convention. It is there so the blended basis-trend state always has a value — the indicator cannot leave the state "undefined." It is not an opinion about what the market is doing when the vote is split. If you want strict-majority behavior, read the slots directly and do not lean on the blended trend alert for tie cases.

Ambiguous states and how to reason through them

A few situations that read cleanly on the chart but require a second look:

The chart looks quiet because the slots disagree, not because the market is quiet

Narrow blended band, wide slot bands. Agreement-vs-cancellation applies. Read the slots.

The basis looks clean but the channel is jagged

The basis is smoothed; the bounds are not. A tidy basis inside a jagged channel is usually a short Length: producing noisy extremes rather than a broken indicator. Either the Length is doing what you wanted or it is not — Basis MA Length will not fix it either way.

An alignment alert did not fire even though the direction "obviously" agrees

"Obvious" agreement on the chart can still be 2-of-3, not 3-of-3. Alignment requires every enabled slot with a valid basis to agree on direction. Hidden valid slots count. If Slot 03 is enabled, hidden, and disagreeing, the alignment alert stays silent and the alert is working correctly.

The blend flipped direction but only one slot looks like it moved

On the Base trim this is usually a weight effect. If the slot that "moved" is the highest-weighted one and the other two are close to a tie, a small move on the heavy slot can change the majority vote on its own. Also check for hidden slots — a hidden slot with non-zero weight is doing work without showing it.

A slot's lines are missing on the earliest bars

Early-bar warmup. Before a slot has accumulated enough higher-timeframe bars to fill its Length window, its values can be na and the lines do not render. The blend can still draw because missing slot values are zero-filled internally, which means the early composite can sit somewhere useless, including near zero. Scroll forward until the slot has enough history and the lines come in before reading the blend or its alerts.

A note on the blended fill

The red fill between the blended upper and lower lines is there to make the composite band easy to pick out at a glance. It is a visual orientation cue and nothing more. The fill is not a zone, not a range target, not an area that the chart has "defined as important." The band behind it is what matters, and the band is composed the way you composed it.

Reading the fill as a "zone" is a common habit, and a habit worth interrupting. The composite is not a declaration about the chart. It is a summary you authored. Use it that way.

A reading routine that holds up under pressure

When the chart is calm it is easy to read it carefully. Under stress the reading collapses into "what is the red band doing," and the red band is the part of the chart most vulnerable to misreading. A working routine for the stressed moment, in four steps:

1. Name the slots first, not the blend. Look at the three slot colors and ask whether each one is rising, falling, or flat. Do this before you look at the red band. The blend's state will make sense as a summary of the slots; start from the summary and you are reading the picture you authored without checking its sources.

2. Locate the basis on each slot. The fully opaque line is the basis; the half-opacity lines are the raw bounds. Confirm that the basis is where you think it is. The most common mid-session misread on a tight stack is treating an outer bound as a basis.

3. Ask where the slots are stepping. Under ON, each slot sits still between its higher-timeframe closes. If a slot is moving on every chart bar and you did not intend that, something is under OFF that you thought was under ON.

4. Only then read the blend. A narrow blend is agreement or cancellation; an aligned-trend blend is every slot moving the same way; a flipped blend is often a weight effect on a close vote. You will not misread the blend if you have already read the slots.

Four steps, maybe twenty seconds once the routine is habit. It is worth practicing while the chart is quiet so it is available when the chart is not.