Quick Start

This page gets the indicator onto a chart and gives you a working read without asking you to understand the whole pack first. The goal is a configuration that is not wrong, a short sanity check, and enough orientation...

Written By Axiom Admin

Last updated 22 days ago

Quick Start

This page gets the indicator onto a chart and gives you a working read without asking you to understand the whole pack first. The goal is a configuration that is not wrong, a short sanity check, and enough orientation that the first few minutes in front of the live pane are spent learning the instrument rather than wondering if it is broken.

There is a temptation, on a first session with any new tool, to start tuning. Resist it. The defaults on this trim are not optimized for your instrument β€” nothing could be β€” but they are deliberately positioned so that a serious reader can see the pane behave honestly within a few minutes and form their own questions about what they would like to change. Tuning before you can describe what you are correcting is the fastest way to convince yourself the tool is broken when nothing is wrong.

If a step here raises a question you want answered immediately, follow the cross-link. Otherwise finish the setup, watch the pane for a while, and come back for depth when you have a specific question in mind.

Before you add it

Pick a chart you already have real context on. The pane will read as abstract on an instrument you have never looked at, because every line on it is relative to the window's own observed high and low. On a chart you know, you can tell a genuinely quiet stretch from a quiet window that normalization is amplifying. On an unfamiliar chart, you cannot β€” and the first read will be harder to calibrate than it needs to be.

A 5-minute chart of a liquid instrument β€” a major futures contract, a highly-traded equity, a large-cap crypto pair during active hours β€” is the easiest first canvas. The default slot stack was designed against that kind of canvas. It will still work elsewhere, but you will be doing a little more translation while you learn.

Have the TradingView Alerts dialog available in another tab or window. The Alerts verification steps near the end of this page expect you to be able to see the alert list without restarting.

The shortest correct setup

  1. Add the indicator to the chart. Leave every input on its default the first time you load it.

  2. Confirm the pane opens beneath price. You should see three colored lines (teal, aqua, blue), a lime-or-red blended line, a gray blended signal, and a translucent fill between the two blended lines. Horizontal guides should sit at 0, 20, 50, 80, and 100. If any of these is missing, stop and read Troubleshooting before changing anything else.

  3. Confirm every visible line stays inside the 0-to-100 pane. This is a mechanical sanity check, not a value judgment about the read. The lines are clamped by the normalization; if something is drawing outside the guides, the script is not doing what it should and the pack wants to know.

  4. Leave the default slot stack as-is for now. 5m / 15m / 60m at equal weights. Session window mode anchored on D. CVD Length 3 SMA, Signal Length 3 SMA. The defaults are a reasonable starting point for intraday reading. They are not a prescription β€” the Workflows page walks through the reasons you might change them β€” but they are legitimate while you calibrate your eye.

  5. Leave Master Smoothing OFF. It is OFF by default, and you want to learn the raw blend first. Master smoothing adds lag in exchange for less noise; applying it before the stack has been understood makes the raw character harder to feel.

  6. Leave On Bar Close ON. It is ON by default. ON is the settled, one-slot-bar-lagged read, and it is the honest default for first contact. Toggling it OFF is on the checklist below, once the rest of the pane is stable.

You should now be looking at a working pane that a serious reader would consider correctly configured for a first session. The next several minutes are the most valuable time you will spend with this indicator β€” and the easiest time to teach yourself things that are hard to unlearn later.

A sanity check, once the pane is up

Do each of these at least once before you trust anything the pane is saying.

Scroll the chart to recent bars and look at the 5-minute slot. It should move visibly during the day β€” up into the upper third on stretches of one-sided buying, down into the lower third on stretches of one-sided selling, through the middle during balanced conditions. If it is pinned at 0 or 100 for the entire visible window, something in the window configuration or the instrument is off; check the Settings page section on Window Mode and Window before assuming the read is telling you about the market.

Check that the 15-minute and 60-minute slots move more slowly than the 5-minute slot. A higher timeframe slot should draw in slower arcs β€” fewer peaks, longer travel, less noise. If the 60-minute slot looks as jumpy as the 5-minute slot, the chart is probably not showing enough history for the higher-timeframe slot to do its job yet. Scroll back or wait; the read should calm down as the slot sees more of its own bars.

Watch a session rollover. If your chart is inside regular trading hours on a session-traded instrument, you should see a dashed vertical line in each slot's color land at the start of the next daily window. If your chart is 24-hour (crypto, some FX), the reset will still land at the daily boundary the exchange uses, which may not match your local day. Either is correct; what you are confirming is that the reset is visible and in the slot's color.

Toggle On Bar Close once. Go to the Inputs dialog, find the On Bar Close? control under PU Settings, and turn it off. Watch the lines during the next bar on your chart timeframe. They should move as the slot bars form, rather than waiting for the slot bar to close. Turn it back on and watch the lines stabilize. This is the single cheapest way to feel the tradeoff the MTF page describes β€” ON is settled, OFF is live but can repaint until the slot bar closes.

Wire one per-slot alert. Open the TradingView Alerts dialog, add an alert on the indicator, pick CVD 01 Is Bullish as the condition, and set it to fire on every bar close. Watch what happens over the next several bars. You should see the alert re-fire on every confirmed bar while the state holds β€” not once on the flip. That repeat behavior is how the pack's alerts are designed to work, and the Alerts page is explicit about why.

If any of those five checks produces something that does not match the description above, the Troubleshooting page is organized by symptom and will tell you whether it is a setup error, a misread, or documented behavior.

What a healthy pane looks like on first load

It looks quiet but alive. The 5-minute slot moves the most and carries most of the short-term character. The 15-minute slot moves in slower arcs. The 60-minute slot drifts across longer stretches. The blend sits somewhere inside the three and moves with the weight of the majority β€” when the three slots lean the same way, the blend confirms cleanly; when they disagree, the blend sits closer to the middle and does not color as confidently.

Session reset dashes land at the start of each new daily session. The fill between the blend and its signal intensifies when the blend pulls away from the signal and thins when they come back together. Color flips happen at every slot's crossings, and some of those flips will be small β€” a hair-crossing on one slot while the other two hold their state is common and does not, by itself, mean much.

A short way to know the pane is honest with you on first load: when nothing is happening on price, the slots should be moving slowly and quietly, the blend should not be staking out an extreme, and the fill should not be screaming a state. When something happens on price, the relevant slot should respond with visible character within a few bars. If the pane is loud during dead air or silent during decisive moves, something in the configuration is off and the rest of this section will not save you β€” go look at Settings.

You will not be able to tell in the first session whether the default stack is the right configuration for your actual process. That is normal, and the Settings and Workflows pages walk through the reasons to change it. What you are after right now is a calibration of your own eye against a legitimate baseline.

First traps

These are the things new readers get wrong in the first hour. Naming them up front is cheaper than watching you learn them the hard way.

  • Reading a color flip as a signal. It is a state descriptor. When the slot's CVD crosses its signal, the color flips, regardless of how far apart the two lines are. A hair-crossing and a strong crossing look the same on the color alone. Weight your attention by the gap between the lines, not by the fact that a flip happened.

  • Treating the blend as the verdict. The blend is a weighted mean of three slots. When the slots agree, the blend is a clean summary. When they disagree, the blend averages the disagreement β€” and a blend at 50 with two slots leaning hard up and one leaning hard down is not the same read as a blend at 50 with three slots sitting flat. Check the stack before you lean on the blend.

  • Treating the session reset dash as a trade event. It is an accounting event. The slot's window rolled over, and the next cumulative period started. Nothing about that implies price is about to do anything specific.

  • Treating a slot pinned at 100 as a strong-buy verdict. Pinning means the slot's current cumulative is near the upper end of what its window has observed. If the window has been quiet, a small move can pin the line. The Visuals and Limitations pages both walk through this; the short version is that normalization tells you about the window as much as it tells you about the market.

  • Hiding a slot and expecting it to stop contributing to the blend. Hide Plot hides the drawing. The math keeps running, the slot keeps feeding the blend, and the per-slot alerts keep firing. If you want a slot's influence on the blend removed, set its Blended Weight to 0. Settings explains the tradeoff.

  • Pushing Pressure Sensitivity or Wick Weight to the edges before you understand them. Both tunables change the character of the slot line in ways that are easy to misread as the market changing character. The Settings page walks you through what each edge of the range does and what it costs.

  • Running the defaults on an instrument they were not designed for. A session-traded equity out of hours, a thinly-traded altcoin, an overnight futures session β€” these will each want something else. The defaults are a starting point, not a universal configuration. Workflows walks through the reading conventions for matching slots to context.

Before you change anything

Once the sanity check passes, sit with the default configuration on a familiar instrument for at least one full session. Not because the defaults are universal β€” they are not β€” but because the only honest way to know what you want to change is to watch the unchanged version through real conditions. A reader who walks into Settings on minute four and starts moving Pressure Sensitivity has nothing to compare against. A reader who watches a full session first can describe, in a sentence, what character of read they wish the line had. That sentence is what Settings is for.

If you do nothing else after the setup above, do that one thing.

Where to go from here

  • You want to change slot timeframes, lengths, or weights. Start with Settings. The controls are described in configuration order, not dialog order.

  • You want to understand what you are looking at on the pane. Visuals and Logic. Includes the blend-at-50 ambiguity walkthrough.

  • You want to use the alerts in a real process. Alerts, then Workflows for the triage pattern.

  • Something on the chart looks wrong. Troubleshooting.

  • You want to know how much weight the estimator deserves. Limitations and Trust Boundaries and For the Geeks are the two honest pages on that question.