Visuals & Logic

This page is the mental model for what you are looking at on the pane. Not "how to configure" — that is Settings. Not "how to avoid misreading" — that is Limitations & Trust Boundaries. This page answers the more fundamental question: when you open the pane and look at it, what is each line, each column, each color, and each reference level telling you, and how should you read them together?

The goal is a frame rich enough that you can describe, out loud, why a slot is where it is and what the blend is doing with its voices — without handwaving. Once you have that, everything else in the pack (setup, tuning, alerts, workflows) sits on top of it cleanly.

The centered-momentum frame

Start with the most important number on the pane: 50. The midline is the equilibrium.

The way MACD works under the hood — before any normalization — is that MACD equals fast MA minus slow MA, the signal line equals a smoother of MACD, and the histogram equals MACD minus signal. Three numbers. Each can be positive or negative. Zero means "no difference." The textbook MACD pane has a zero line running through the middle and plots all three above and below it.

This pane keeps the same story but moves the zero line to 50 and bounds the pane to 0 and 100. Raw MACD, raw signal, and raw histogram each get divided by ATR, scaled by ATR Sensitivity, and pushed through a centered sigmoid. The shape of the transformation is such that a raw value of zero maps exactly to 50; positive raw values map into 50 to 100; negative raw values map into 0 to 50. Nothing about "up means bullish, down means bearish" has changed — the pane is still a centered momentum read. The midline just lives at 50 instead of at zero, and the extremes are bounded instead of open-ended.

Carry that frame. When you see a slot line near 50, the slot is telling you its underlying MACD is close to its own zero line: fast MA and slow MA are close together after ATR scaling. When you see a slot line well above 50, the underlying MACD is materially positive in ATR units. Well below 50, the mirror. The slot's color, not its height, tells you whether that MACD is above or below its signal. The 70/30 guides are just that — guides. They are not thresholds.

Why the pane is bounded at all

Bounding is how the indicator makes cross-timeframe and cross-instrument slots readable on the same scale. Raw MACD on a 60-minute SPY chart runs through bigger numbers than raw MACD on a 5-minute SPY chart, and raw MACD on SPY has nothing in common, unit-wise, with raw MACD on BTCUSDT. If you want all of those slots drawn on one pane without scale tricks, you need a normalization that removes price scale entirely. That is what the ATR-sigmoid pair does.

The cost of the bounding is that the pane cannot distinguish "stretched" from "extremely stretched" once a slot saturates near 0 or 100. A slot at 98 and a slot at 94 are both very stretched against their ATR; the pane will not tell you which move would have been bigger in price units. That limitation is a property of the bounded frame, not a defect. For the reader who needs raw magnitude, the pane is not the tool. For the reader who needs cross-context comparability, it is.

The three families of plot on the pane

When you look at the pane you are looking at three kinds of plot, stacked on each other:

• Slot lines. One line per enabled, un-hidden slot. Each slot's line is its ATR-normalized MACD — the slot's fast value — drawn in that slot's color.

• Blend lines. Two lines, the blended fast and the blended slow, drawn on top of the slot lines.

• Blended histogram. Columns rooted at 50, drawn under everything else.

Each family carries a different piece of information. You read them together. Learning to move between them deliberately is most of the skill with this pane.

The slot lines — the evidence

Each enabled slot plots one line, colored by a two-state rule with a fallback. The rule:

• If the slot's normalized signal (the slot's slow) exists (is not na), color the line by whether the slot's fast is above or below the slot's slow. Above → full-opacity in the slot's color. Below → half-opacity in the same color.

• If the slot's slow has not formed yet (first bars after load, or after a settings change that resets the MA state), fall back to whether the slot's fast is above or below 50. Above 50 → full-opacity. Below 50 → half-opacity.

The default slot colors are:

The slot line tells you, for that slot specifically, where the underlying MACD sits relative to zero — expressed in bounded 0-to-100 terms. A teal line at 62 running full-opacity means MACD 01 (the 5-minute slot by default) is moderately positive against recent 5-minute ATR, and the color says it is above its own signal. A blue line at 42 running faded means MACD 03 (the 60-minute slot by default) is mildly negative against recent 60-minute ATR, and the color says it is below its own signal. Two slots. Two contexts. Same unit.

The fallback to a 50-comparison during warmup matters most on the higher-timeframe slots right after a load. You may briefly see the 60-minute slot's color come from the fallback rule, then snap to the proper slot-fast-vs-slot-slow rule once the signal MA has enough history to form. Under defaults this is a few hourly bars. On a fresh chart with limited backscroll, let the pane run a few hours before you lean on the color of a higher-timeframe slot.

The blend lines — the summary

When Plot Blended K/D is on (the default), two more lines draw on the pane:

• Blended fast — the weighted average of every enabled slot's normalized fast, across slots with non-zero weight. Lime when the blended fast is above the blended slow; red when below. If the blended slow has not yet formed, falls back to whether the blended fast is above 50.

• Blended slow — the corresponding weighted average of the slot slow values. Drawn in gray.

Between the blended fast and blended slow, a translucent fill draws, carrying the same regime color as the blended fast. That fill is deliberate — it gives you the blend's current posture at a glance without having to trace the exact positions of two thin lines.

The blend is the summary of the slots. It is one voice that speaks for the voices you weighted. It inherits the alignment or disagreement of those voices: when every slot leans the same way, the blend leans decisively too; when slots disagree, the blend lands somewhere in the middle, colored by whichever side the weighted average tips to. That middle can be misleading on its own — a mild blend can reflect genuine consensus on "not much happening" or a genuine disagreement between slots that cancel out. The correct read in those moments is to look at the slot lines and decide which interpretation the evidence supports. Limitations & Trust Boundaries covers the mild-blend-with-split-slots case explicitly.

The blended histogram — the convergence speedometer

When Plot Blended Histogram is on (the default), columns draw rooted at 50. Each column's height represents the weighted average of the enabled, non-zero-weight slots' normalized histogram values for that bar. Each slot histogram is MACD-minus-signal pushed through the same normalization before the blend is averaged. Columns above 50 mean the contributing slot histograms lean positive in the weighted average; columns below 50 mean the mirror; a flat row at 50 means the weighted histogram blend is at equilibrium.

The column's color encodes two independent facts at once — which side of 50 it sits on, and whether it is rising or falling versus the prior bar's column. The four-state color code:

This four-state code is the single most useful visual shortcut on the pane. It lets you read the blend's current momentum in one glance, before looking at the lines at all.

A shift from full-green to faded-green is the first sign that the positive histogram blend is losing thrust. Often that happens before the blended fast crosses below the blended slow, but it is not a guarantee that a crossover is coming. Similarly, a shift from full-red to faded-red is the first sign the negative histogram blend is losing downside pressure. Readers who watch only the lines tend to notice crossovers; readers who watch the histogram notice deceleration earlier. Deceleration often precedes crossover, and sometimes it resolves without one.

The symmetric logic across the 50 midline is deliberate. A faded-green bar and a faded-red bar both mean "the story is losing thrust" — one into a top, one out of a bottom. Two halves of the same rule.

The reference levels

Five horizontal guides draw across the pane regardless of which slots or blend features are on:

• 100 — red line. The upper boundary. Nothing draws above it.

• Overbought — dashed gray line, at Overbought Level (default 70). Pure visual bracket.

• Mid Line — solid gray line at 50. The equilibrium midline.

• Oversold — dashed gray line, at Oversold Level (default 30). Pure visual bracket.

• 0 — green line. The lower boundary. Nothing draws below it.

The color choice on the boundary lines (green at 0, red at 100) is a deliberate visual cue that extreme saturation is present, not a claim about the direction of any trade. A line pegged at 100 is pegged because its underlying MACD is extremely stretched against recent ATR and the chosen ATR Sensitivity is mapping that stretch into the top of the pane. A line pegged at 0 is the mirror. What you do with that information depends on the rest of the context and is not a conclusion the reference colors are making for you.

The 70/30 reference lines are not alert triggers. No native alert fires on a crossing of either one. They are visible landmarks so the eye has something to anchor to in the bounded pane. If you want an alert on a slot crossing 70, you build it with TradingView's generic alert wiring on the slot's plot, not through this indicator's alert conditions.

How the plot changes — causality in four sentences

Because the pane mashes several moving pieces together, it helps to have a compact mental model of what causes what.

1. Source price moves. The raw MACD and raw signal on each slot shift. The slot's fast and slot's slow shift with them, in 0-to-100 units. The slot's line moves.

2. ATR changes. The yardstick each slot uses for normalization shifts. A volatility expansion enlarges the yardstick, shrinking the same raw MACD into a smaller normalized read. A volatility contraction shrinks the yardstick, inflating the same raw MACD into a larger normalized read. Slots can move without any MACD change if ATR moved. See Limitations & Trust Boundaries for when that matters.

3. Settings changes. Adjusting Fast Length, Slow Length, MACD MA Type, Signal Length, Signal MA Type, or any Power User parameter changes the underlying MACD and signal the slot is normalizing. Adjusting ATR Length or ATR Sensitivity changes the normalization itself. Adjusting Blended Weight changes the blend, not the slot. Adjusting Hide MACD NN Plot changes visibility only.

4. Higher-timeframe bar state. When a slot's On Bar Close? is true (default), the slot holds its value inside the chart bars that sit inside the slot's forming HTF bar, then steps at the HTF bar close. When false, the slot reports the live forming HTF bar's value and may revise as the bar develops. See MTF & Repainting for the full picture.

Four causes, four kinds of motion. When you see the pane doing something you did not expect, working through those four sentences usually locates the reason in under a minute.

Reading the pane — three concrete patterns

Once the frame is in place, a few concrete reading patterns are worth practicing until they feel natural.

Pattern 1: Slot alignment and disagreement

Look at the slot lines together. Are they all pointing in the same direction, or are they split? On the default three-slot setup:

• All three full-opacity, on the same side of 50. The 5-minute, 15-minute, and 60-minute MACDs all sit above their signals by color, and all three are on the same side of the MACD zero line by position. The three cadences agree on both axes. The blend will be on the same side decisively.

• Split across the midline. The 5-minute slot is above 50 and the 60-minute slot is below 50. The fastest context is doing something different from the slowest. The blend will land near the middle, and the blend's color depends on whose voice is weighted more.

• All on the same side but disagreeing about fast vs slow. All three slots are above 50, but one is full-opacity (its fast above its slow) and two are faded (their fasts are below their slows). The MACD-family story is bullish across cadences at the stretch level, but the convergence story is mixed. The histogram will usually be close to 50.

Each of these reads is different information. The blend flattens them into one number. The slots show you what the blend hid.

Pattern 2: Histogram shifting before blend lines cross

When a bullish blend (blended fast above blended slow, lime line above gray) is about to lose its lead, what changes first is often the histogram's rising-vs-falling state. The histogram goes from full-green to faded-green while the blend lines are still on the same side of each other. The blend has not yet crossed its signal; the histogram is telling you that the weighted normalized histogram is moving back toward 50.

That faded-green state is usually the earliest of the three warnings the pane offers — earlier than the blend crossing its signal, earlier than a slot flipping color. Not every faded-green moment is followed by a crossover; plenty revert to full-green without anything changing. Readers who watch the histogram often see pressure changes before readers who watch only the lines.

Pattern 3: Slot pinned, blend calm, histogram near 50

A single slot saturates near a boundary. The other two slots are well inside the pane. The blend is near 50, and the histogram is flat at 50.

This usually means one of two things. Either the pinned slot is reading a moment of unusual stretch that the other slots do not share — real dispersion across timeframes, which is legitimate context — or the pinned slot has crossed into ATR-collapse territory where a small raw move is normalizing into a large number. The tell is on the chart: are the bars on the pinned slot's timeframe visually large (real stretch) or visually quiet (ATR collapse)? Limitations & Trust Boundaries walks ATR-collapse pinning in full and explains the fix (reduce ATR Sensitivity, lengthen ATR Length, or both).

Pattern 4: Slot color versus slot position

One separation worth drilling into. A slot's color (full vs. faded) encodes fast-vs-slow on that slot; the slot's position on the pane (near 0, near 50, near 100) encodes how stretched the fast line is in ATR units right now. Those two facts are independent, and the pane shows them simultaneously without fusing them.

A slot line at 60 running faded is a slot whose fast is moderately above the 50 equilibrium but below its own slow line — the underlying MACD is above zero, but the signal has caught up and is now above MACD. A slot line at 40 running full-opacity is a slot whose fast is below the 50 equilibrium but above its own slow — underlying MACD is below zero, but MACD has pulled back above its signal from a more bearish position.

Neither of those is a contradiction. Full-opacity-below-50 and faded-above-50 are normal states, and readers who expect color and position to tell the same story end up confused whenever the two pieces of information stop agreeing. They are not supposed to agree; they are two axes of the same slot, and the pane is showing them both so you can read both.

Common misreads this page is trying to prevent

• "The line is above 70, so it is overbought." The line is above 70 because its underlying MACD is materially positive in ATR units. Its color tells you whether that MACD is above or below its signal. It is a momentum read, not a mean-reversion read. Acting on a 70 crossing as if it were RSI is an anti-pattern.

• "The blend is flat, so nothing is happening." The blend is flat when the weighted voices cancel. The slots can be loud and disagreeing, producing a calm-looking blend that hides real dispersion. The read is to look at the slots, not at the blend.

• "The histogram is decoration." The histogram is the convergence/divergence speedometer. The four-state color code is where the pane's earliest information sits. Hiding the histogram to reduce clutter costs more than it saves.

• "All three slots align, so the blend is confirmed." If the three slots share source, MA family, and lengths and differ only in timeframe, their agreement is mostly trend persistence on the underlying instrument, not independent confirmation. The alignment trap lives in Limitations & Trust Boundaries; it is serious enough that the pack names it in several places.

Where to go next

• For the full repaint and multi-timeframe treatment, MTF & Repainting.

• For the 24 alert conditions and what they do and do not confirm, Alerts.

• For the alignment trap and the ATR-pin trap in full, Limitations & Trust Boundaries.

• For the shape of the ATR-sigmoid mapping explained without the recipe, For the Geeks.