Sentiment indicators are tools that try to quantify the collective mood of market participants — fear vs. greed, complacency vs. panic, crowded vs. washed-out — so it can be used as an input to decisions. They exist because price alone does not reveal positioning: two markets at the same level can be supported by very different crowds, and the prevailing theory of this whole category is contrarian — that emotion and positioning become most extreme precisely when they are about to reverse, so the crowd is "wrong at the extremes." The core tension that runs through every member of this family is the gap between that intuitive premise and the measured evidence: sentiment reliably describes how stretched positioning has become, but it is a weak and unstable timing tool, and several of the most-quoted indicators have a popular reputation far stronger than their backtested edge. This section is the map of the major gauges; the per-indicator mechanics, thresholds, and honest base rates live in the child nodes.

What this section covers

This branch defines the recognized, broadly-used sentiment gauges and how each is read. It sits under Behavioral Finance & Market Sentiment: the parent branch covers why sentiment exists (herding, fear/greed, overreaction) and the academic anomalies; this node covers the instruments that measure it. It deliberately defers indicator-specific formulas and evidence to its five children rather than duplicating them.

A useful way to organize the field — and the way the children are split — is by what kind of data the gauge uses:

• Survey / "soft" sentiment — what people say. → AAII & Investor Surveys. Weekly polls of opinion (AAII individuals, Investors Intelligence newsletter writers, BofA's fund-manager survey). Cheap, long-history, but small self-selecting samples and noisy.
• Options / volatility "hard" sentiment — what people pay for protection. → Put/Call Ratio and VIX & Volatility Gauges. Real-time, hard data, but the public series mix informed hedging with uninformed speculation.
• Positioning — where real money is actually committed. → Commitment of Traders (COT) and Fund Flows & Positioning. Revealed-preference data (futures positions, fund inflows, cash allocations), but lagged and weekly.

A fourth, cross-cutting form is the composite — blends of several of the above into one number. The best-known is CNN's Fear & Greed Index, an equal-weighted 0–100 average of seven inputs: stock-price momentum, 52-week highs vs. lows, breadth volume, the put/call ratio, the VIX, junk-bond demand, and safe-haven demand (stocks vs. Treasuries) (CNN). SentimenTrader's "Smart Money / Dumb Money Confidence" spread is another composite expressing the same idea in the institutional-vs-retail framing. Composites smooth the noise of any single gauge but inherit the weaknesses of their components and hide which one is driving the reading.

The core interpretive frame

Two ideas recur across every child and are worth stating once:

1. Contrarian-at-extremes. Extreme bullishness/complacency (low put/call, low VIX, high bull %, crowded long positioning, fund-manager cash at a low) is read as a caution flag; extreme fear/capitulation (high put/call, VIX spike, record bearishness, washed-out positioning) as a bounce setup. The rule only applies at genuine extremes — in the middle of a range these gauges carry little reliable information, and "extreme" can persist for months in a strong trend.

2. Smart money vs. dumb money. The popular framing splits participants into a contrarian, hedging "smart money" (commercials in COT, institutional hedgers) and a trend-chasing "dumb money" (retail call buyers, large speculators, retail fund flows). It is a useful organizing heuristic — and the retail-flow contrarian effect has real long-horizon academic support (Frazzini & Lamont's "dumb money") — but the labels are imperfect (COT swap dealers sit in the "commercial" bucket despite being financial) and should not be taken as precise.

When sentiment matters — and when it doesn't

Sentiment is most valuable at emotional extremes and over slow horizons (weeks to months). It earns its keep as a context/regime overlay: it tells you the fuel state of a move (dry powder vs. exhaustion) and whether a tape is euphoric or capitulated. It is least valuable as a stand-alone, short-horizon timing trigger — most of these series are weekly or lagged (COT carries a three-day reporting lag; surveys are weekly), too coarse for entry timing, and their mechanical-threshold backtests are generally modest and regime-dependent. The disciplined use across the whole category is the same: define "extreme" objectively against the series' own history (percentile or z-score), smooth the noise, and require price/catalyst confirmation before acting.

Adoption & evidence (section-level)

These are among the most-watched numbers in finance, quoted daily across media and dashboards, and adopted by retail and institutions alike. But the section's honest through-line is that popularity outruns measured edge, and the children flag this carefully and consistently:

• The robust academic results often belong to different instruments than the popular gauge and must not be conflated — e.g. Pan–Poteshman's signed-volume option edge is not the CBOE put/call ratio CNBC quotes; Baker–Wurgler's market-based sentiment index is not the AAII poll.
• COT's "follow the commercials" folklore is contradicted by peer-reviewed work (Sanders, Irwin & Merrin) finding positions largely react to prices rather than forecast them.
• VIX is a genuine coincident stress gauge but a poor forecaster of future realized volatility; its most robust empirical fact is the volatility risk premium (a persistent overstatement), not a directional forecast.

The defensible summary: sentiment indicators are strong describers of crowd state and occasionally flag true capitulation/euphoria; they are weak predictors of timing. Read the children for each gauge's specific base rates and disputes.

System relevance

For the Augustus trade-setup agent, this entire section feeds the backdrop / conditioning layer, never the trigger layer. The consistent contract across all five children is identical and worth encoding once at the section level: sentiment extremes should adjust conviction and position sizing at the margin — raising the ceiling on a long near capitulation, tightening risk and shrinking size into euphoria — but a sentiment reading alone must never originate or override a price-and-structure-based thesis. Hard caveats inherited from the children: feed the correct series (equity-only put/call, not Total/Index); normalize to each series' own range; respect the weekly/lagged cadence; and remember that several of these gauges react to price as much as they lead it. The positioning children (COT, flows) link naturally to the Market Regime Engine as read-only macro context.

Sources

• Market sentiment — Wikipedia (definition; survey vs. trading-data measurement; contrarian framing): https://en.wikipedia.org/wiki/Market_sentiment
• CNN Business, "Fear & Greed Index" (seven equal-weighted components, 0–100 scale): https://www.cnn.com/markets/fear-and-greed
• SentimenTrader, "Smart Money / Dumb Money Confidence Spread" (institutional-vs-retail composite framing): https://sentimentrader.com/indicators/smart-money-dumb-money-confidence-spread
• AAII, "Investor Sentiment as a Contrarian Indicator" (contrarian-at-extremes logic, mixed record): https://www.aaii.com/journal/article/feature-investor-sentiment-as-a-contrarian-indicator
• Child nodes (mechanics, thresholds, base rates, and disputes): Put/Call Ratio; VIX & Volatility Gauges; AAII & Investor Surveys; Commitment of Traders (COT); Fund Flows & Positioning — all in this folder.

Disputes flagged: This is an overview; the contested, source-specific claims (Pan–Poteshman vs. public put/call; Baker–Wurgler vs. AAII; Sanders/Irwin on COT reverse causality; VIX forecasting power) are documented in the respective child nodes and should be retrieved there before relying on any single gauge's predictive value.