Positioning vs Sentiment vs Breadth vs Volatility

Article

Why do these four terms get mixed up?

Market commentary uses "sentiment" as a catch-all. A fall in the VIX, a lopsided options ratio, a crowded futures trade, and a narrow rally all get described as "bullish sentiment" or "bearish sentiment." That shorthand is convenient and costly. It hides which question the data actually answers, and the answer changes what you can reasonably conclude.

This page treats the four ideas as separate evidence families. Think of them as four instruments on a dashboard. One gauge shows fuel, another shows speed, another shows engine temperature, and another shows the road ahead. A driver who reads only one of them, or who reads all four as "how the car is doing," will misjudge the trip.

If you want the narrower comparison of opinion versus participation, read Sentiment vs. Breadth: Measuring Opinion Versus Participation. That page focuses on two of these families. This page widens the frame to four, adds positioning and volatility, and asks how the four fit together and where they quietly count the same thing twice. For the wider topic map, start at the Market Sentiment Analysis hub.

What is the difference between the four, in one table?

The four families can tell different stories about the same day without any of them being wrong. That is the central idea of this page.

What is positioning?

Positioning is the inventory of exposure that already exists. It measures what traders and investors hold, not what they say they believe. Someone can describe themselves as cautious while holding a large long position, or call themselves optimistic while being mostly hedged. Positioning data looks at the holdings, not the commentary.

Where does positioning data come from?

A classic public source is the Commitments of Traders (COT) program run by the U.S. Commodity Futures Trading Commission (CFTC). The CFTC's Commitments of Traders page describes several report families. The one aimed at financial contracts is the Traders in Financial Futures report, which sorts reportable traders into four groups: Dealer/Intermediary, Asset Manager/Institutional, Leveraged Funds, and Other Reportables.

Two details matter for interpretation. First, the reports are weekly, and the figures describe positions as of an earlier day, so the data is always somewhat old when you read it. Second, the CFTC's explanatory notes explain that a trader appears in the detailed categories only when its position reaches a reporting level set by the agency. Positions below those levels are shown together as "nonreportable," calculated by subtraction from total open interest. So COT is a structured view of large futures positions, not a census of every investor's portfolio.

For a full treatment of the report, see Commitments of Traders. For the individual series, the indicator pages on the Commitments of Traders net position and leveraged funds net position show what each one counts.

What can positioning tell you?

Positioning can help you notice:

• unusually large net-long or net-short exposure in a market;
• a split between participant groups, such as asset managers leaning one way and leveraged funds leaning the other;
• fast week-to-week shifts in exposure;
• crowding, meaning many participants hold similar exposure;
• price moves that happen against, or because of, existing exposure.

What can positioning not tell you?

Positioning does not tell you why a position exists. A short futures position might be an outright bearish bet, a hedge against a stock portfolio, one leg of a basis trade, or part of a spread. Reading every short as bearish is a common error.

Positioning also does not tell you when a crowded trade will unwind. Crowding can persist for a long time, and a position can look extreme for months. The most you can say is that a crowded market may be more sensitive if a common catalyst arrives.

How do you compare positioning across time?

Raw contract counts are hard to compare across years, because open interest (the total number of outstanding contracts) changes. More comparable measures include net position as a share of open interest, a rolling percentile against the market's own history, a z-score (how many standard deviations from average), and the weekly change. Looking at gross long and gross short positions separately also helps, because a flat net position can hide large offsetting exposures.

What is sentiment?

Sentiment is the broadest and least concrete of the four. It tries to capture optimism, pessimism, fear, confidence, or expectations. Because it is a feeling rather than a holding, it can only be measured indirectly, and different measurement routes produce very different indicators under the same label.

What are the main ways sentiment is measured?

• Surveys ask people what they expect. The AAII investor survey is a familiar example. See AAII and investor surveys as contrarian signals.
• Options activity, such as put/call ratios, uses what people trade as a proxy for what they feel.
• Search and media measures track attention and narrative intensity.
• Fund flows show where money is going in and out of funds.
• Composite gauges blend several inputs into one number.

Why do two "sentiment" indicators disagree?

Because they measure different behaviors. A survey records an opinion, which can be shaped by recent price moves, by question wording, and by who chooses to respond. It may not match what the same respondents actually own. An options ratio records transactions, but the transactions have many motives.

The put/call ratio is the clearest example. Puts are often bought as protection, and calls are often bought for upside, so a high put/call ratio is commonly read as bearish. But Cboe's analysis of early-exercise order flow shows that heavy volume in deep in-the-money options, driven by large participants managing assignment risk, can inflate equity put/call ratios without reflecting new bearish conviction. The article also describes alternative constructions, such as focusing on out-of-the-money and at-the-money contracts, that aim to reduce that distortion. The lesson is that a ratio is a measurement of activity, and the meaning of the activity needs a second step. See put/call ratio and options sentiment and the equity put/call ratio indicator page for how the series is built.

What can sentiment tell you, and what can it not?

Sentiment can show when attitudes have become unusually one-sided and when attention to a topic is intense. It cannot show actual exposure, and it cannot show how many stocks are participating. A very optimistic survey tells you about opinions, not about who is already invested or how widely the market is rising.

Every sentiment indicator should carry a short label stating its source, how it is built, and what it actually measures. Without that, "sentiment" can mean almost anything.

What is breadth?

Breadth asks whether a market move is shared widely or driven by a narrow group. It looks beneath the index level at the constituents.

This matters because many popular indexes are capitalization-weighted, meaning larger companies count for more. When a small number of very large companies rise, the index can climb even if most stocks are flat or falling. An equal-weight version of the same companies gives each one the same influence, so comparing the two is a rough way to see how much of a move depends on the biggest names. It is a useful participation lens, though not a full breadth measure.

What are common breadth measures?

• advancing versus declining securities, often cumulated into an advance-decline line;
• new highs versus new lows;
• the percentage of stocks trading above a moving average;
• sector participation;
• the median stock's return compared with the index return;
• diffusion indexes built from the trends of many constituents.

For a technical-analysis treatment, see the market breadth section.

What can breadth tell you?

Breadth can reveal concentration. An index can reach a high while fewer constituents join the move. Widening participation suggests a trend leans less on a few names. Narrowing participation suggests concentration, rotation, or both.

What can breadth not tell you?

Breadth does not tell you whether investors are optimistic, whether options are expensive, or whether futures traders are crowded. A narrow market can keep rising for a long time, and a broad market can fall broadly. Breadth is best read as a market-structure check on how a move is built, not as a forecast.

What is volatility?

Volatility describes the size of price movement. It can be measured two ways, and keeping them apart avoids a lot of confusion.

What is realized volatility?

Realized volatility is calculated from returns that have already happened, typically as the standard deviation of daily or weekly returns over a window. It describes the past.

What is implied volatility?

Implied volatility is inferred from option prices. It reflects what the options market is pricing in for the future, which includes both expectations of movement and the price participants are willing to pay to hold protection.

The best-known implied-volatility gauge is the Cboe Volatility Index, or VIX. Cboe describes the VIX as a leading measure of market expectations of near-term volatility conveyed by S&P 500 option prices. Cboe's explainer on VIX and VIX1D adds the key points: VIX is a constant 30-day expected-volatility reading built from S&P 500 options, it is non-directional, and a shorter-horizon sibling, VIX1D, looks at same-day expectations and can react much more sharply to short-term shocks.

"Non-directional" is the phrase to remember. VIX tells you how large the options market expects moves to be, not whether prices will rise or fall. For more, see the Cboe Volatility Index indicator page.

What is term structure?

Volatility expectations differ by maturity. Near-term expectations can spike around a specific event while longer-term expectations stay calm, or the reverse. The shape across maturities is called the term structure. See VIX term structure and, for the gap between implied and realized volatility, the volatility risk premium.

What can volatility not tell you?

High volatility does not mean prices will fall, and low volatility does not mean safety. A market can drift lower with low realized volatility, or rally sharply with high volatility. Volatility measures the size and price of movement, not its sign.

Where do the four families overlap?

The families are conceptually separate, but real indicators often blend them. Classifying each indicator by its primary measurable object, then noting secondary uses, keeps the picture clear.

Put/call ratios are built from actual options trades, so they reflect behavior and positioning. They are often used as a sentiment proxy, but hedging, product mix, and exercise activity can distort that reading.

The VIX measures expected volatility derived from option prices. People often call it a "fear gauge," which pulls it into sentiment discussions, but its measurable object is expected volatility. "Fear gauge" is an interpretation layered on a volatility number, not its definition.

COT data is positioning. Analysts sometimes describe extreme positioning as "bullish sentiment," which blurs two ideas. A more precise sentence is "positioning is consistent with crowded long exposure," which says what the data shows and nothing more.

Breadth is often used to confirm sentiment, but it measures participation, not belief.

Credit spreads and leverage sit near the edges of this framework. They are market prices and balances that reflect risk appetite and borrowing, and they often serve as an independent check on the four families above. See credit spreads and risk appetite and margin debt and leverage.

What is the anti-double-counting rule?

Suppose a composite gauge lists VIX, VIX term structure, a skew measure, a put/call ratio, and an implied-volatility percentile. That looks like five independent indicators. In reality all five come from the same options market and tend to move together. Giving each full weight makes the composite an options-market score wearing a "broad sentiment" label.

The fix is to weight by family rather than by indicator count. In a simple illustration with arbitrary, made-up weights, an options and volatility family might receive 20 percent of the total, split among its members, while breadth, positioning, credit, and leverage each receive their own 20 percent. The exact weights matter less than the discipline: correlated members of one family should not collectively outvote independent families simply because more data points exist.

The related question of how to build such a framework is covered in Sentiment Composite Framework, and a companion article on combining breadth, volatility, and sentiment without double-counting works through the overlap problem from the other direction.

How can the four axes be read together?

Instead of forcing one score, you can describe the market on four axes and keep the answers separate.

Labels should always be words, not just colors, so the reading is clear to every reader and to screen readers. The combination of four labels says more than any single number could.

What does the combination look like in practice?

The eight sketches below are descriptive patterns, not predictions. Each shows how the families can confirm or contradict one another.

• Optimistic, broad, calm, not crowded. Sentiment is constructive, participation is wide, volatility is subdued, and positioning is not extreme. The families confirm one another.
• Optimistic, narrow, calm, crowded. Headline mood looks strong, but concentration and crowding make the picture more fragile than the mood suggests.
• Pessimistic, broad decline, high volatility, short positioning. Strong agreement across families. The open question is whether short positioning is already so large that a rebound is more likely to be sharp.
• Pessimistic surveys, broad advance, low volatility. Stated opinion disagrees with market behavior. This can happen when investors stay cautious while prices improve.
• Bullish positioning, weak breadth, rising volatility. Existing long exposure rests on a narrowing base while uncertainty rises.
• Bearish positioning, improving breadth, falling volatility. The market may be healing while participants remain defensive. That does not guarantee a squeeze, but the divergence is informative.
• Neutral sentiment, narrow market, low volatility. A quiet but concentrated market. Calm should not be mistaken for broad health.
• High narrative fear, stable prices, expensive options. Attention and hedging demand are elevated around an event while realized price action stays contained.

The point of the list is that preserving contradictions is more informative than averaging them away.

Worked example: same price, different internal state

The numbers below are invented to illustrate the idea. They are not real market data.

Imagine two separate months in which a large-cap index finishes up 3 percent.

Month 1. The equal-weight version of the index rises 2.8 percent. About 72 percent of constituents sit above their 50-day moving average. The VIX falls. High-yield credit spreads narrow. Leveraged-fund positioning sits near its historical median. Survey sentiment is moderately positive. A careful description: broad participation, easing stress, and positioning that is not extreme.

Month 2. The equal-weight version falls 0.5 percent. Only about 38 percent of constituents sit above their 50-day moving average. The VIX rises. High-yield credit spreads widen. Leveraged-fund long exposure sits at a high historical percentile. Survey sentiment is very optimistic. A careful description: headline strength with narrow participation, rising risk pricing, and crowded exposure.

The headline return is identical. The internal state is not. A reader who looked only at the index, or only at the survey, would have called both months "bullish." A reader who checked all four families would see two quite different markets, and would have more reason to ask what could change the picture.

How do publication timing differences complicate things?

The families do not update on the same clock, and treating delayed data as if it were live creates false stories. A typical set of inputs might look like this:

• index price: real time;
• VIX: intraday;
• breadth: end of day or intraday;
• COT positioning: weekly, describing an earlier day;
• margin balances: monthly;
• surveys: typically weekly;
• search data: sampled and indexed relative to a baseline.

When you line these up, record both the observation date (the day the data describes) and the retrieval or publication date (when it became available). A positioning reading that reflects a Tuesday cannot have informed a decision made on the preceding Monday, and a backtest that ignores that gap will look better than any real process could have been. See Data Latency, Publication Lags, and Vintage Control for the detailed treatment, and the source ladder for how to rank evidence by how directly it measures what it claims.

How should each family be normalized?

Different data-generating processes call for different treatment, so a single universal method is a poor choice.

• Positioning: net exposure relative to open interest, plus a percentile within the market's own history.
• Sentiment: normalize within the same source. Compare a survey with its own past readings, not with a different survey. Search interest is indexed relative interest, not an absolute count of searches.
• Breadth: percentage participation, or the gap between a broad measure and the headline index.
• Volatility: levels, historical percentiles, recent changes, term structure, and the comparison between implied and realized volatility.

Thresholds should be economically interpretable and written down before outcomes are tested, whenever possible. Tuning a cutoff to fit historical returns tends to produce a rule that looks excellent in the past and fails afterward.

What are the most common mistakes?

1. Calling positioning "sentiment." Exposure is not belief.
2. Assuming all put buying is bearish speculation. Puts are widely used for hedging, spreads, and structured exposures, and exercise flow can inflate volume.
3. Using the VIX as a direction forecast. It measures expected size of movement, not sign.
4. Treating narrow breadth as an automatic top signal. Concentrated markets can persist for long stretches.
5. Mixing live and delayed data without timestamps. This creates false narratives and invalid backtests.
6. Counting several indicators from one market as independent. Family weighting reduces this.
7. Fitting thresholds to past returns. Interpretable, pre-stated thresholds hold up better.

What does this framework not tell you?

It does not tell you what to do with a portfolio. A description of the market's state is not a recommendation, and no combination of the four families produces a dependable signal. Positioning, sentiment, breadth, and volatility can all change quickly, and each can stay extreme longer than most people expect. The framework helps you ask better questions and spot contradictions. It does not remove uncertainty.

It also does not say that more indicators are better. A small set of transparent measures, one chosen from each relevant family, usually gives a clearer picture than a long list of overlapping ones. For the broader backdrop that these readings sit within, see macroeconomics and market regimes, and for cross-market confirmation see cross-asset risk appetite. For standard definitions of terms used here, the glossary is the place to start.

A reading sequence for a careful reader

A careful reader works through the families in an order that keeps the question clear:

1. Name the question. Are you studying crowding, participation, uncertainty, or attitude? Each points to a different family.
2. Pick one primary measure per relevant family. Starting with twenty indicators makes contradictions impossible to sort out.
3. Check source and freshness. Know when each observation described the market and when it became available.
4. Compare each series with its own history. Avoid arbitrary universal cutoffs.
5. Check the independent families against each other. Compare positioning with breadth, volatility, and credit.
6. List the contradictions. Write them down instead of averaging them away.
7. Write one plain sentence. For example: "Positioning is crowded long, breadth is narrowing, and implied volatility is rising despite positive survey sentiment."
8. State what would change the reading. A description that cannot be proven wrong is not much of a description.
9. Keep description separate from action. Reading a market state is not personalized investment advice.

Frequently Asked Questions

References

• CFTC: Commitments of Traders
• CFTC: Commitments of Traders Explanatory Notes
• Cboe: VIX Volatility Products
• Cboe: What the VIX and VIX1D Indices Attempt to Measure and How They Differ
• Cboe: How Early Exercise Order Flow Impacts Equity Option Put/Call Ratios