10.2 Market Sentiment and Technical Analysis
Key Takeaways
- Sentiment is the prevailing mood, opinion, or positioning of market participants — how bullish or bearish the crowd is — inferred from what people say, how they position, and how they price risk.
- The crowd matters because price is a social outcome: informed hedgers and insiders, large speculators, and uninformed small traders jointly set the tape.
- The crowd is often right during an established trend (participation fuels the move) and often wrong at turning points, when opinion is one-sided and fuel in that direction is spent.
- Sentiment indicators are hard to time, may be contrary or confirming depending on context, are subject to lags and revisions, and can lose value if the indicator itself becomes crowded.
- A sentiment extreme is a hypothesis, not a clock: wait for price to stop making new extremes in the crowded direction before treating the crowd as spent.
Sentiment is the prevailing mood, opinion, or positioning of market participants: how bullish or bearish the crowd is. It is not a second price. It is not the CMT Association's grade of whether your chart is "correct." Independent OpenExamPrep teaching for this CMT Level I unit (still Theory and History, 38% of the 2026 Program Guide) treats sentiment as a measurable overlay on price: surveys, option prices, futures positioning, short interest, fund cash, and headlines that quantify whether the crowd is one-sided.
Three channels, same object:
- Opinion — what people say (advisories, surveys, news tone).
- Positioning — what people hold (futures open interest by trader class, short interest, fund cash, margin).
- Price of insurance — what people pay to hedge or to speculate on vol (VIX, put/call).
If you cannot say which channel a series belongs to, you are not ready to interpret it.
Why the crowd matters
Technical analysis is the study of market-generated data. That data is generated by a crowd. Charles Dow's averages, Wyckoff's composite operator, and candlestick language all assume that many participants, not one rational representative agent, leave a trace. Behavioral finance in the prior section named the biases. This section names the aggregation: when enough people share a bias, it shows up as a trend, a climax, or a dead-cat pause.
Split the crowd, at Level I resolution, into informed and uninformed participants.
| Category | Typical members | How technicians use them |
|---|---|---|
| Informed | Corporate insiders; commercial hedgers who know the physical market; sometimes large specialized specs | Positioning that is against a mature trend can be a warning that the people with a reason to hedge or to know the books are no longer sponsoring the move |
| Uninformed / less-informed | Small speculators, late advisory followers, headline-driven retail | One-sided opinion and positioning at extremes is the classic contrary input |
| Large speculators | Reportable noncommercials, many CTAs and funds | Often trend-following; their crowding can confirm a trend and then mark the uncomfortably popular end of it |
The labels are roles, not moral grades. A commercial in wheat who is short futures against inventory is not "smarter than the chart" in every market. A small spec who is long a healthy uptrend is not always the sucker. The exam skill is to know whose book you are reading and whether that book is usually with the trend or early at turns.
The crowd matters for a second, mechanical reason: someone has to take the other side. If everyone who wants to be long is already long, incremental buying power is gone even if the story is still bullish. Sentiment tools try to see that empty tank before the last up-close.
The crowd during trends versus at turning points
This is the Level I distinction that keeps candidates from treating every poll as a fade.
During an established trend, the crowd is often right. Participation is the trend. Rising price with rising speculative longs, expanding volume, and bullish (but not unanimous) opinion is how an advance is supposed to look. Fading every mildly bullish advisory in a months-old uptrend is not contrary thinking. It is arguing with sponsorship. Humphrey Neill's contrary-opinion tradition, taught in the external-data section, does not say the majority is always wrong. It says the majority is dangerous when it is nearly unanimous and when the easy money in that direction has been taken.
At turning points, the crowd is often wrong. Tops form when bullish opinion and long positioning are one-sided: advisers are fully invested in the story, put/call is depressed, VIX is asleep, fund cash is thin, commercials (where relevant) are hedging into strength. Bottoms form when bearish opinion is one-sided: advisers have thrown in the towel, put buying is frantic, VIX is spiked, cash has piled up. The crowd that was right for months is now the exit liquidity for whoever still has dry powder.
Worked sketch, no magic numbers: an equity index has rallied for 14 months. Advisory services are 70%+ bulls, equity put/call is scraping the lower end of its recent range, and mutual-fund cash is at a cycle low. That cluster is a turning-point hypothesis, not a sell-at-the-open order. The same readings six months earlier, when bulls were 52% and cash was mid-range, were trend-compatible. The difference is extremity plus duration, not the existence of bullish people.
A second sketch on the way down: after a 25% decline, magazine covers declare the death of equities, VIX has doubled, and small specs in the related futures are crowded short. The crowd is now bearish at a potential low. Contrary logic says the next durable move is often up — if price stops making lower lows. Sentiment without that price condition is a diary entry.
Remember the behavioral link. Herding and regret aversion keep people with the trend (I will not be the only one missing it). Loss aversion and availability of the last crash keep people from buying a low. Sentiment series are how those feelings get counted.
Challenges of using sentiment indicators
Sentiment is useful and easy to misuse. Four challenges belong on the exam.
| Challenge | What goes wrong | What to do instead |
|---|---|---|
| Timing | Extremes can persist. A one-sided crowd can get more one-sided for weeks or months. VIX can stay low in a grind-up. Advisory bulls can sit above 60% while price makes higher highs. | Treat an extreme as a condition, not a date. Require a price failure: loss of a swing high, a trendline break, a weekly close back through a moving average you already use. |
| Contrary versus confirming | The same series is a confirming tool in the middle of a trend (specs adding longs as price rises) and a contrary tool at an exhaustion extreme (specs max long as price stalls). Candidates who memorize "put/call is always contrary" fail both uses. | Ask: is this reading normal for this trend, or historically stretched? Confirming = sponsorship. Contrary = empty tank. |
| Data revisions and lags | Commitments of Traders figures are as-of Tuesday and released Friday. Short interest reports have historically been twice-monthly snapshots, not a live tape. Surveys revise. Seasonal adjustment and holiday weeks distort. You are reading last week's crowd, not this morning's. | Know the as-of date. Do not fight a Monday breakout with a COT print from last Tuesday as if it were a tick. Pair lagged positioning with live price, volume, and option prices. |
| Crowding of the indicator itself | Once a series is a celebrity (VIX, a famous put/call threshold, a magazine-cover rule), discretionary traders position for the rule. The old extreme may not fire, or it may overshoot as everyone tries to be the contrarian. Grossman–Stiglitz logic from the EMH unit applies: a fully exploited public signal cannot stay easy money. | Prefer clusters of sentiment (positioning + insurance + opinion) over one famous number. Re-estimate what "extreme" means in the current regime. |
Timing, worked
A candidate sees VIX at 12 and sells the S&P 500 because "complacency." Price then rallies another 8% over two months while VIX meanders 11–14. The sentiment label (complacency) can be fair the entire time. The trade was a timing error. Low VIX describes a regime. It does not timestamp the session of the break.
Contrary versus confirming, worked
Rising noncommercial longs in crude while price makes higher highs is confirming sentiment: large specs are sponsoring the uptrend. The same noncommercial book, months later, at a record net-long while price fails to make a new high on weaker internals, has flipped to a contrary warning. The series did not change names. The context changed.
Revisions, worked
You cannot treat Friday's COT as if commercials covered on Friday. They reported what they held on Tuesday. A mid-week squeeze can already have flipped the tape. Short interest published with a lag can miss a covering rally that already happened. Build that lag into the story or do not use the series for a same-day trigger.
Crowding of the indicator, worked
If every desk fades a VIX print above 30 the same way, the fade is in the price. March 2020-style spikes still marked fear, but the first VIX 30 print of a crisis is not the same trade as the fifth. Celebrity thresholds decay. That is a sentiment-specific version of the adaptive-markets warning from the EMH chapter.
How to study this unit
Write four sentences you can reproduce under Prometric time:
- Sentiment is the crowd's mood, opinion, and positioning.
- The crowd is often right in a trend and often wrong at one-sided turning points.
- Informed versus uninformed books are not interchangeable.
- Timing, contrary-versus-confirming context, revised/lagged data, and crowding of the tool itself are why a sentiment extreme is a hypothesis that still needs price.
The next two sections fill the toolbox: measures from market data (VIX, futures open interest, options volume and open interest, COT groups, insiders, short interest) and measures from external data (news, advisories, contrary opinion, fund cash).
Key Takeaways
- Sentiment: mood, opinion, and positioning of the crowd
- Crowd often right during trends; often wrong at one-sided turning points
- Informed (insiders, commercials) versus uninformed (small specs, late opinion)
- Four challenges: timing, contrary vs confirming, data revisions/lags, crowding of the indicator
What is "sentiment" as CMT Level I uses the term for financial markets?
How does the crowd typically behave during an established trend compared with a turning point?
Which statement best names the practical challenges of using sentiment indicators?