What sentiment actually measures

Ask most retail traders what market sentiment is and you get something about mood — whether people feel bullish or bearish, what the headlines say, how the news is skewed.

That is not what sentiment analysis measures, and the gap between the two explains why so much sentiment commentary is useless.

Opinion is free. Positioning costs money.#

Anyone can say they are bullish. Saying it costs nothing, commits nothing, and can be reversed silently.

Buying costs money. It commits capital, creates exposure, and — critically — creates a future obligation to act. Someone who has bought must eventually sell. Someone who has sold short must eventually buy back.

Sentiment analysis measures the second thing. Not what participants say, but what they have already done and therefore what they must do next.

Why positioning predicts anything at all#

If nearly everyone who wants to buy has already bought, who is left to push price higher?

That is the whole logic. Extreme positioning is self-limiting, because the buying that drove it has exhausted the pool of future buyers. And the crowded side has an exit problem: when everyone holds the same position, everyone reaches for the door at once.

This is why sentiment extremes are contrarian signals. Not because the crowd is stupid — the crowd is usually right during the middle of a move — but because at an extreme there is nobody left to keep being right.

What actually gets measured#

Four families of data, covered properly in later lessons in this track:

Derivatives positioning. Open interest, put-call ratio, where option writers are committed. The most direct measure available, because a derivatives position is an explicit obligation.

Institutional flows. FII and DII buying and selling. Large, slow-moving capital that is hard to disguise and hard to reverse quickly.

Volatility pricing. India VIX, implied volatility levels. This measures what the market is paying for protection, which is fear expressed in rupees rather than in words.

Breadth. How many stocks are participating. A rally led by five names is a narrower, more fragile thing than one where most stocks are rising.

Each measures commitment. None measures opinion.

Sentiment is a condition, not a signal#

The most common misuse is treating a sentiment reading as a trade trigger. "PCR is high, so buy."

That fails, because sentiment tells you about the environment, not about timing. Positioning can stay extreme for a long time and get more extreme. A market that is overbought can become far more overbought.

The right use is as a filter on other signals:

• A bullish technical setup in a market where positioning is already crowded long is a lower-quality setup.
• The same setup where positioning is washed out is a higher-quality one.

Sentiment changes your confidence and your size. It does not, by itself, tell you to act.

The reflexivity problem#

There is a genuine complication worth stating early.

Sentiment data is public. Everyone can see the same PCR, the same FII figures, the same VIX. So participants position based on the data, which changes the data, which changes how it should be read.

This is why fixed thresholds do not survive. "PCR above 1.3 is bullish" may have worked in some period and then stopped working, because enough people traded it that the crowding moved.

The practical response is to read sentiment relative to its own recent range rather than against absolute rules. Is this reading unusual for this instrument, in this regime? That question survives; "PCR above 1.3" does not.

Where to start#

Beginner-level sentiment work is genuinely simple:

1. Is volatility high or low relative to the last few weeks? This tells you whether options are expensive or cheap, and therefore whether buying or selling premium is structurally favoured.
2. Is positioning one-sided? Heavy concentration on one side of the option chain means a crowded trade.
3. Is breadth confirming? If the index is rising but most stocks are not, the move is narrower than it looks.

Three questions, all answerable in a couple of minutes, all telling you about the environment rather than the direction. That is what sentiment is for.

Check yourself

0 of 4 answered

1. 1.What is the core distinction that sentiment analysis rests on?

   Good news versus bad newsOpinion, which is free, versus positioning, which commits capital and creates future obligationsRetail traders versus institutional tradersShort-term versus long-term views

2. 2.Why are sentiment extremes contrarian signals?

   The crowd is usually wrong about directionIf nearly everyone who wants to buy has bought, there is nobody left to push price higherExtremes always mean-revert within a weekInstitutions deliberately trade against retail positioning

3. 3.PCR reaches an unusually high reading. What is the correct use of this information?

   Buy immediately — it is a bullish signalTreat it as an environmental condition that raises or lowers confidence in other signalsIgnore it entirely; PCR is meaninglessWait for it to reach exactly 1.3 before acting

4. 4.Why do fixed sentiment thresholds tend to stop working over time?

   Exchanges change how the data is calculatedThe data is public, so participants position on it, which changes the data and how it should be readThe thresholds were never based on real dataVolatility makes all thresholds meaningless

What to take away

• Sentiment measures positioning, not mood or headlines.
• Positioning matters because it creates future obligations to act.
• Extremes are contrarian because they exhaust the pool of future participants.
• Sentiment is a condition that filters other signals, never a trigger on its own.
• Early and wrong are indistinguishable in a leveraged, expiring position.
• Read sentiment relative to its own recent range — fixed thresholds decay through reflexivity.