Visuals and Logic

Visuals and Logic

This page explains what you see in the oscillator pane, what each element means, and — just as importantly — what the chart can lead you to believe that is not quite right. Most misreads do not happen because the oscillator is broken. They happen because the reader sees a number, a color, or a shape and draws a conclusion the data does not actually support. If you want to get better at reading this tool over time, this is the page to return to.

What is in the pane

Slot lines (CVD 01, CVD 02, CVD 03)

Each enabled slot draws a single line in its assigned color (teal, aqua, and blue at defaults). The line represents the slot's normalized CVD value — where the estimated net directional pressure sits within the -100 to +100 range for the current window.

Each line is colored by its own regime state: one shade when the slot's CVD is above its Signal MA, another shade when below. This coloring tells you which side of its own moving average the slot is currently on. It does not tell you whether the reading is strong, weak, or trustworthy — just the directional relationship between the CVD line and its smoothed reference.

Blended CVD and Signal

When Plot Blended CVD/Signal is on, two additional lines appear:

• Blended CVD — a weighted composite of all enabled slots' CVD values, drawn in a distinct color.

• Blended Signal — a weighted composite of all enabled slots' Signal values, drawn in gray.

A shaded fill between the blended CVD and Signal lines shows green tint when CVD is above Signal (blended bullish) and red tint when below (blended bearish).

The blend is a summary. It tells you roughly where the combined reading sits. It does not tell you whether the underlying slots agree or disagree — you need to check the individual lines for that.

Session reset markers

When a Session-mode slot's window resets (e.g., a new daily boundary begins), the indicator draws a vertical dashed line in that slot's color spanning the full pane height. If all three slots are in Session mode with the same window, you will see three overlapping dashed lines at each boundary.

Reset markers tell you that the accumulation window just rolled over. The running CVD sum and normalization range restart there, but the carry state used for neutral-bar classification can still echo the prior classified direction until new bars overwrite it.

Reference lines

Five horizontal lines provide context:

How to read the states that matter

Single-slot regime

A slot is bullish when its CVD line is above its Signal line, bearish when below. This is a simple moving-average crossover classification applied to the normalized CVD.

What it tells you: the direction of the recent trend in estimated pressure within that slot's timeframe and window.

What it does not tell you: whether the reading is near zero or near the extremes, whether the regime just flipped or has held for fifty bars, or whether the underlying volume was high or low. Regime color is a binary state — it carries less information than it appears to.

Regime transitions happen when the CVD line crosses its Signal line. Because both lines are smoothed by their respective MAs, regime transitions are lagged relative to the raw CVD shift. A slot's CVD may have already started reversing before the regime color changes. If you are watching for early shifts, the CVD line's direction matters more than the regime color.

Multi-slot agreement and disagreement

When all enabled slots share the same regime (all bullish or all bearish), estimated pressure is aligned across timeframes. When they disagree, at least one timeframe is seeing different conditions than the others.

Agreement during a trend is expected, not exceptional. If the market has been moving cleanly in one direction, the 5m, 15m, and 60m slots will almost certainly agree — they are all measuring the same dominant move at slightly different resolutions. Three green lines during a strong uptrend is not three independent confirmations. It is one trend seen from three overlapping viewpoints.

The real value of multi-slot stacking shows up during transitions and ambiguity. When the 5-minute slot flips bearish while the 60-minute slot remains bullish, that tells you something worth paying attention to: short-term estimated pressure has shifted, but the larger-timeframe read has not. Is the short-term move a counter-trend pullback that the longer timeframe will absorb? Or is it the leading edge of a reversal the higher timeframe has not yet reflected? The oscillator does not answer that question. It surfaces it.

What to do with that information depends on how you trade. If you are following the higher-timeframe trend, a short-term slot flipping against you is worth noticing but not necessarily acting on — the higher slot still holds. If you are trading the shorter timeframe, the flip is direct information about conditions on your execution horizon, and the higher slot's stubbornness tells you the broader picture has not changed yet. The key insight is that disagreement between slots is information, not noise. When you see it, the right move is almost always to investigate further — check price structure, check the candles the short-term slot is reacting to — rather than either ignoring the flip or treating it as a confirmed reversal.

Blended CVD near zero while slots disagree

This is one of the most important states to recognize. The blend is a weighted average. If Slot 01 is at +60 and Slot 03 is at -40 (with equal weights and Slot 02 somewhere in between), the blend can land near zero. On the chart, that looks like neutrality — the blended line hovering around the center, the fill barely visible.

But neutrality and conflict are not the same thing. A blend near zero from unanimous slots near zero is genuinely neutral. A blend near zero from opposing slots is a masked disagreement. The only way to distinguish them is to check the individual slot lines.

If you use the blend-only view (individual slots hidden), you lose the ability to make this distinction. The blend will look calm while the underlying readings are at war with each other. This is the single most important reason to check the individual slots when the blend's behavior seems ambiguous or surprisingly flat.

Readings near the extremes

A reading near +100 or -100 means the slot's CVD accumulation has reached (or is near) the most extreme value within its current normalization window. It does not mean buying or selling pressure is at some universal maximum.

This distinction matters most in two situations:

Early in a Session-mode window, when the normalization range is narrow, modest volume can push a reading to +80 or +90. The number looks extreme, but the range it is being measured against is small. These early readings are mechanically correct — the CVD is near the top of its range — but the range itself is thin. Wait for the session to build out before treating extreme values as contextually significant.

Late in a session, when the normalization range has expanded, a +80 represents more accumulated directional volume than the same +80 read earlier. The normalization is relative to the window's history, so the same number can mean different things depending on when it appears within the window.

CVD divergence from price

Sometimes the oscillator's estimated pressure moves in one direction while price moves in the other. The CVD is rising but price is falling, or vice versa.

This is a classic CVD divergence pattern, and it can mean two very different things — and the way you think through the ambiguity matters more than the divergence itself.

Scenario 1: Absorption. The candle structure shows net buying participation (more sub-bars classified as bullish), but price is not moving up because selling is absorbing that flow. Concretely: you see green-bodied candles with buying wicks, the model classifies them as bullish, CVD rises — but price stays flat or drops slightly because each buyer is being met by a seller at or near the current level. This can precede a breakout if the buying eventually overwhelms the selling. It can also be distribution where buyers are being fed inventory by larger sellers who want out. The oscillator sees the buying-side candle structure. It does not see the seller standing behind the wall.

Scenario 2: Estimation ambiguity. The candle structure is genuinely mixed — long wicks on both sides, small bodies, dojis — and the participation model is classifying sub-bars as weakly directional or carrying the prior direction forward through ambiguous bars. The divergence may not reflect genuine opposing forces at all. It may reflect the model reaching for a classification that the data does not strongly support, especially if Pressure Sensitivity is set high enough that neutral bars carry substantial directional weight.

How to think through it: Check the candle structure on the divergence bars directly. If the candles have clean bodies and clear direction but price is not moving, the divergence is more likely to reflect genuine absorption — something interesting is happening that the oscillator surfaced. If the candles are messy, mixed, or thin-bodied, the divergence may be the model over-reading ambiguous structure. Also check your Pressure Sensitivity: at high values, the carry logic through neutral bars can create a directional drift in CVD that does not reflect new commitment. Lowering sensitivity during a suspected divergence and watching whether it persists is a reasonable diagnostic.

The oscillator cannot tell you which scenario is true. What it can do is surface the divergence so you investigate rather than assuming the trend is uncontested.

Shallow versus mature reading

There is a meaningful gap between what this oscillator looks like at a glance and what it is actually telling you once you understand the mechanics. The difference is worth naming directly.

The shallow read

"The line is positive, so buyers are winning." "All three slots are green, so the trend is confirmed." "The blend crossed above zero, so I should look for longs." "+80 means very strong buying pressure."

These are not wrong in some loose directional sense, but they skip past the parts that determine whether the reading is actually informative. They treat the oscillator as if it were reporting a fact when it is reporting an estimate — an estimate that is relative to its own window, smoothed by its own MAs, and based on candle-structure classification rather than order-book data.

The mature read

A reader who has spent time with this tool understands that:

• The CVD values are estimates from candle structure, not measurements from the order book. A positive reading means the participation model classified more sub-bar volume as buyer-aligned than seller-aligned. It does not mean literal buyers outweighed literal sellers.

• A slot's reading is relative to its own normalization window. A +70 in a low-volatility session with a narrow range may represent less absolute volume commitment than a +40 later when the range has expanded. The number tells you where you are within the window, not how much actual volume was committed.

• Regime color is a lagging classification. The CVD line may have already reversed before the regime color flips. The lag depends on the Signal MA length.

• Session mode's cold-start bars are thin-data artifacts. The first few bars after a reset are operating on very little history, and the normalized values during that window can swing to extremes on modest volume.

• In Rolling mode, the oscillator can change direction because old bars left the window, not because new bars were directional. A decline in the CVD reading does not always mean new selling pressure arrived — it may mean old buying pressure aged out.

• Multi-slot agreement is only as meaningful as the independence between the timeframes. Slots on adjacent timeframes measuring the same dominant move will agree by correlation, not by independent confirmation.

None of this means the oscillator is unreliable. It means the readings become more useful once you understand what shapes them.

States worth pausing on

These are the situations where a quick glance is most likely to mislead and a closer look is most likely to pay off: