Market breadth indicators: compare participation on TradingView

Choose the measure by the question and available data

Start with the population you want to study: a named index, all eligible listings on an exchange, or an explicitly maintained basket. The matrix compares indicator families rather than reproducing their formulas. The linked A/D Line and McClellan guides cover calculation and initialization details.

One universe, a rising index and four falling stocks

This hypothetical six-stock universe has beginning-of-session capitalization weights totaling 100%. Assume unchanged membership, shares and float, no dividends or corporate actions, and synchronized closing prices. Index return is the sum of each beginning weight multiplied by its price return.

• The capitalization-weighted index gains 1.50 + 0.50 - 0.10 - 0.10 - 0.10 - 0.10 = 1.60%. An index starting at 1,000 closes at 1,016.
• Only 2 of 6 stocks advance: 33.33% advancing, with 4 of 6, or 66.67%, declining. Coverage is complete and no stock is unchanged.
• An equally weighted portfolio of these same stocks returns (3 + 2 - 1 - 2 - 2 - 2) / 6 = -0.33% for the session. Equal weighting measures return magnitude; the advancing percentage measures direction only.
• The disagreement identifies concentrated index gains. It does not tell you when the index will fall, whether it will fall, or whether the four declining stocks will recover.

Check what TradingView is actually measuring

TradingView documents breadth on its Indices and World Stocks tables, country pages and individual index pages. Country views describe local markets, while individual index views describe their components. For chartable breadth series, its official tutorial suggests searching terms such as Highs, Lows and Above Moving Average, then selecting the relevant market.

• Read the series description and provider before adding it to Supercharts. Record the full symbol, intended universe, eligible security types, interval and update time. A similar symbol name does not establish identical coverage.
• Do not compare NYSE exchange breadth with a Nasdaq index and call the difference constituent divergence. It can be a cross-market comparison, but label it that way.
• Check whether funds, preferred shares, multiple share classes or other listings are included. McClellan documents why composite exchange data and filtered operating-company data can differ.
• For a custom basket, retain the membership list used on each observation date. Applying today's surviving members to the past changes the research question and can introduce survivorship bias.
• A breadth figure shown on a web page does not establish that the same series is available to Pine. Confirm an accessible chart symbol or the required constituent inputs before specifying a script.

Report coverage before interpreting a percentage

Suppose a fixed universe contains 100 securities, but only 80 have valid prices and enough history for a 200-session average. If 48 are above the average, the observed breadth is 48 / 80 = 60%, with 80% coverage. Calling this 60% of all 100 securities would be incorrect.

• The 20 unknown classifications leave full-universe breadth between 48 / 100 = 48% and (48 + 20) / 100 = 68%. A claim that a majority is above the average cannot be established from these inputs.
• Do not silently count missing observations as below the average or remove them without reporting the reduced denominator. If nothing is valid, the percentage is unavailable, not zero.
• Separate unchanged issues from missing issues. A confirmed unchanged close is an observation; an absent or stale quote is not evidence that a stock was flat.
• Check whether fresh listings, suspended stocks, holiday differences or inconsistent adjustments explain missing classifications. Different indicator families can have different valid samples even on the same date.

Write an observable rule before reviewing its outcome

A research flag can be precise without claiming predictive power: at the accepted session close, flag a positive index return when fewer than half of the same universe advanced. Require complete valid coverage for this example. A lower coverage threshold would be a separate policy to define and test, not a reason to hide missing members.

• With the six-stock example, the rule flags +1.60% index return and 33.33% advancing participation. If F has no valid closing observation, this complete-coverage rule produces unavailable rather than true or false.
• Define the comparison close, timezone and release timing. If breadth is published or corrected after the index close, the earliest usable decision time is after that information arrives.
• When daily inputs are used on an intraday chart, yesterday's completed breadth remains yesterday's observation. Do not relabel it as fresh intraday participation or substitute the final daily value into earlier bars.
• Give Pineify's Pine Script AI Coding Agent the verified symbols or basket, timing rule and missing-data policy to draft a chart aid or alert. Inspect the generated requests and compare sample outputs with the source data before using notifications.
• Save the original hypothesis, universe version, coverage, values and a screenshot URL in Trading Journal notes. Add the later review separately. Neither a script nor a journal can supply a missing historical observation or make a hindsight label available in real time.