Visuals and Logic

The pane is built to carry a specific mental model: the convergence and divergence of MACD lines, stacked across three timeframes, projected into a bounded 0-to-100 frame so the three can sit on one axis honestly. This page teaches that model, then names every element on the pane and where it sits in the reading order.

Read this page once end-to-end. Come back to the four-state histogram table — it is the part that most rewards a second look.

The mental model in one paragraph

For each slot, three raw quantities live in price units: MACD, signal, and histogram. The slot's MACD is the fast MA minus the slow MA. The slot's signal is a smoothing of the slot's MACD. The slot's histogram is MACD minus signal. All three of those numbers are then projected into a bounded 0-to-100 frame built around ATR on the slot's own timeframe and a sigmoid that puts 50 at the center. Fifty is equilibrium for the series being mapped. For the normalized MACD line, 50 means the raw MACD is zero, so the fast MA and slow MA are equal. For the normalized signal value, 50 means the raw signal value is zero. For the normalized histogram, 50 means MACD equals signal, so the raw histogram is zero. 0 and 100 are asymptotes, not reversal thresholds.

The blend is a weighted average of the slots' normalized MACD lines, signal lines, and histograms. It is a summary, not a separate indicator. The histogram column color is a four-state code of the blended normalized histogram's above/below-50 state crossed with rising/falling, and it is the single most informative visual on the pane for catching the convergence story accelerating or unwinding.

The reading order

This is the most important section of the page. Get this habit right and most misreads of the pane resolve themselves.

Histogram first. Slots second. Blend last.

Histogram first

The four-state column color answers a question in one glance: is the convergence story bullish or bearish right now, and is it accelerating or losing thrust? That is the speedometer. A reading above 50 and rising is bullish acceleration; above 50 and falling is bullish deceleration. Below 50 and falling is bearish acceleration; below 50 and rising is bearish deceleration.

The transitions between those states (bright green → faded green → faded red → bright red) often telegraph a regime change before the slot lines cross their own signals, because the histogram is tracking whether the blended normalized MACD-vs-signal gap is expanding or contracting around the midline. The slot lines measure where the MACD side is; the histogram measures whether that gap is gaining or losing force. When those two disagree — a slot line still reading bullish while the histogram has turned faded green for several bars — the pane is warning that thrust is leaving the current side before the line state has fully flipped.

Slots second

Once the histogram has told you the tempo, the slots tell you whose story it is. Each slot colors full-tone when its fast is above its slow and faded when its fast is below its slow. Look for:

• All three in full tone — all three timeframes are reading bullish on their own terms. Whether that is meaningful depends on whether you differentiated the slots (see Limitations and Trust Boundaries on the alignment trap).

• All three faded — all three bearish on their own terms. Same qualifier.

• A ladder — slow slot full-tone, middle mixed, fast faded. That is a higher-timeframe baseline holding while the intra-day is rolling over, and it reads differently from a clean disagreement at all three cadences.

Blend last

The blended fast and blended slow pair is a weighted summary of what the slots did. Treat it as a headline that comes after the evidence. When the blend disagrees with the slot picture, your eye should go back to the slots and figure out which slot's weight is pulling the blend, not trust the blend because it is the loudest element.

If you lead with the blend, you hide the deceleration beat inside the histogram and you stop noticing when the slots are telling different stories at different cadences. The blend exists to save you looking at three slot lines in a hurry, not to replace the habit of looking at them.

The pane, element by element

From top to bottom, what is on the pane and what it means.

The 100 rail (red)

An asymptote, not a level. The slot values and the blended values are clamped at 100 as a defensive step, and a reading pinned near 100 is telling you the sigmoid has hit its flat region for this sensitivity-and-ATR pairing. It is not saying "twice as stretched as 80." It is saying "the dial and the volatility have combined to park the reading up here."

The overbought guide (dashed gray, default 70)

A user-configurable reference zone. It is a visual bracket. It is not a threshold that fires an alert, and it is not a reversal trigger. If you want to read it as a reversal zone, do so with the understanding that the mean-reversion interpretation is an RSI habit transferred onto a different kind of pane.

The 50 midline (solid gray)

Equilibrium. On the normalized MACD line, this is where raw MACD is zero. On the normalized signal line, this is where the signal value is zero. On the histogram, this is where MACD equals signal and the raw histogram is zero. Most of the interesting transitions happen as lines cross 50 or as the histogram shifts its four-state color around it.

The oversold guide (dashed gray, default 30)

Symmetric to overbought. Same posture: a reference zone, not a trigger.

The 0 rail (green)

The lower asymptote. Same semantics as 100.

The blended fast line (lime or red, thicker)

The weighted combination of the enabled, non-zero-weight slots' normalized fast lines. Lime when the blended fast sits above the blended slow, red when below. Its thickness is controlled by Blended Line Width.

If every enabled slot sits at weight zero, or every contributing slot is still warming up (its MA history is insufficient), the blend is na and nothing from the blend layer draws.

The blended slow line (gray, thicker)

The weighted combination of the slots' normalized slow lines. Runs alongside the blended fast and carries the tinted fill between the two.

The tinted fill

A faded version of the blend regime color — lime-tinted when the blended fast is above the blended slow, red-tinted when below. It gives the blend posture a read-from-across-the-room surface.

The slot lines (teal, aqua, blue)

One line per enabled slot, each slot's normalized fast. Slot 01 is teal, slot 02 is aqua, slot 03 is blue. Full tone when that slot's fast is above its own slow; faded tint when below.

One subtlety: in the earliest bars after a load, before the slot's slow line has accumulated enough history, the slot's slow value can be na. In that case the slot falls back to coloring by whether its fast sits above or below 50. This first-bar fallback is a nicety to keep the line from being colorless; it is not a hidden smoothing rule, and it settles to the standard fast-vs-slow coloring as soon as the slow line has enough bars to form.

The blended histogram columns (four colors, rooted at 50)

This is the pane's speedometer for the convergence story. The column color combines two facts at once: whether the blended histogram sits above or below 50, and whether the current column is larger or smaller than the prior column. The four resulting states:

Two things to notice about that encoding. First, the color transitions are the early-warning surface for regime change — bright green to faded green is the first visible sign that the bullish thrust is decelerating, and it frequently shows up before the slot lines cross their own signals. Second, a reader who scans only "green or red" and not "bright or faded" throws away half the information the histogram is carrying.

On the very first bar after a fresh load, when there is no prior bar to compare against, the rising/falling state falls back to the above-50 side, so the first column is colored by its side of the midline only. Once there is a prior bar, the full four-state logic takes over.

Slot colors in detail

The slot color rule:

• Slot 01 draws in teal. Full tone when slot 01's normalized fast is above its own normalized slow. Half-opacity teal when below. Fallback to "fast vs. 50" when the slow line has not formed yet.

• Slot 02 draws in aqua. Same rule.

• Slot 03 draws in blue. Same rule.

The design intent: each slot tells its own timeframe's story in its own color, and you can read agreement or disagreement across the three at a glance without having to translate.

Blend colors in detail

The blend color rule:

• Lime when the blended fast sits above the blended slow. The tinted fill matches.

• Red when the blended fast sits below the blended slow. The tinted fill matches.

• When either the blended fast or the blended slow is na, the line and the fill disappear from the pane. The pane is telling you the current configuration produced no blend on this bar — either every enabled slot is at weight zero or every contributing slot's value is na — rather than telling you the tool has failed.

If you see the blend disappear and it surprises you, run the Troubleshooting check on weights and warm-up first.

What moves the plot

Knowing what causes the plot to move is how you troubleshoot a surprise from it. In order, the causes:

• Source price moves. The raw MACD gap on one or more slots changes, the normalized K and downstream D and H shift, and the blend and the histogram follow.

• ATR on a slot's timeframe changes. The normalization scale changes. A volatility expansion can drag a slot's reading toward 50 on steady raw MACD because the normalization bucket got bigger. A volatility contraction can push a slot toward a boundary on small underlying moves.

• ATR Sensitivity changes. The dial that scales every slot's reading before the sigmoid. Higher pins faster; lower flattens. There is no "best" sensitivity — there is a trade between expressiveness and saturation.

• ATR Length changes. The yardstick itself. Longer steadier, shorter jumpier.

• Slot length or MA-family changes. Changes the evidence a slot is normalizing. This is the per-slot tuning surface.

• Slot Source changes. Changes which series the slot reads.

• Slot TimeFrame changes. Changes which higher-timeframe context the slot is reading.

• Slot Blended Weight changes. Changes the blend; does not change that slot's own line or its per-slot alerts.

• Enable Master Smoothing toggles or retunes. Changes the blend's lag profile; does not change any slot line or per-slot alert. Blend-based alerts fire on the smoothed blend when smoothing is on.

• On Bar Close? flips globally. Changes every slot's repaint behavior at once. On returns the confirmed previous slot-timeframe bar values per slot. Off returns the live intra-bar values per slot. Detail conversation lives on MTF and Repainting.

• Reference-line level changes. Only the dashed guides move. The oscillator values do not change.

If a pane movement does not match one of the causes above, you are either looking at a warm-up artifact (the MA and ATR histories have not filled), a chart-timeframe collapse (the chart sits at or above the slowest slot), or a first-bar color fallback settling. Troubleshooting walks each case.

Two interpretations to avoid

The bounded pane is not RSI

Fifty is equilibrium. Seventy and thirty are reference zones you set. The pane does not fire on them and the reading is not a mean-reversion gauge. A reading of 85 on a quiet instrument and a reading of 85 on a violent instrument do not describe the same economic event — they describe different raw MACD gaps that happen to have the same position in their own ATR-normalized units. Treating 70 as "sell" and 30 as "buy" is the single most common RSI habit transferred onto this pane, and it is the misread Limitations and Trust Boundaries fences off.

The histogram is a speedometer, not a direction indicator

Bright green is not "up." Bright green is "above 50 and rising." The color is a two-dimensional reading compressed into one glance, and scanning it as one-dimensional throws away exactly the deceleration information the four-state code was built to surface.

The missed state almost everyone fails at first is faded green. It sits next to bright green on the pane and looks almost the same at a distance. A reader who has trained their eye on red-or-green sees faded green and files it as "still up." What the pane is actually telling them is that the bullish thrust is decelerating — the blended normalized histogram is narrowing back toward 50, even though the current bar has not yet crossed. That is the earliest visible warning the pane produces, and mis-scanning it costs precisely the informational advantage the four-state code was built for. The symmetric case is faded red on the bearish side: "still down" is the wrong translation; "thrust leaving the bearish side" is the right one.

Where to go next

• Knobs as decisions — Settings.

• The repaint switch and the slot-timeframe rule — MTF and Repainting.

• The ten alert conditions and what they do not confirm — Alerts.

• The alignment trap (three slots at identical defaults mistaken for three independent votes) and the rest of the honest limits — Limitations and Trust Boundaries.

• The mental-model shape of the ATR-sigmoid mapping — For the Geeks.