Sentiment Indicators for CMT Candidates: Options, Surveys, VIX, and COT

Sentiment is a measurement problem

Sentiment describes participants' expressed views or positioning. A bullish survey response, a put option trade, and a futures position are different observations. They should not be combined as if they measured one identical psychological state. The 2026 CMT Program Guide includes sentiment measured from market and external data at Level I, and applied stock and derivatives sentiment at Level II. These public objectives include VIX, options activity, futures open interest, COT, and surveys. They do not state that any single threshold times a reversal.

A useful first question is whether an indicator measures opinion, activity, positioning, or implied volatility. The second is who is in the sample and when the data become available. Only then ask whether the observation is unusual relative to its own history.

Put/call ratios

A simple options-volume put/call ratio is put contracts traded divided by call contracts traded over the same period and universe. If 900,000 puts and 600,000 calls trade, the ratio is 900,000/600,000 = 1.5. That means put volume is 50% larger than call volume. It does not prove that 900,000 bearish investors bought puts; some puts may be sold, hedged, or part of spreads. The ratio also differs when based on open interest rather than daily volume, and index options differ from single-stock options.

A high value may indicate concern or hedging, but “high” must be defined relative to the selected series and regime. One analyst might study the 90th percentile of the past five years. That is an illustrative research threshold, not an official CMT rule. If the ratio reaches an extreme while price is falling, a contrarian might watch for stabilization; a trend follower might read persistent demand for protection as risk. A price event is needed to distinguish those decisions.

Suppose an index closes at 95 with a prior support at 94, and the put/call ratio rises to 1.5. A conditional plan might wait for a close above 98 and improving breadth before considering a rebound. A close below 94 would weaken the rebound thesis. If the ratio remains high, it can coexist with continued downside. The statistic is context, not a timing device.

VIX measures expected variation, not direction

Cboe defines VIX as an options-derived, constant-maturity 30-day measure of expected S&P 500 volatility. See Cboe's VIX FAQ. A VIX reading of 24 conventionally represents about 24% annualized implied volatility, not a prediction that the index will fall 24% in the coming month. A rough one-standard-deviation 30-calendar-day magnitude under a square-root-of-time approximation is 24% × √(30/365) ≈ 6.88%. Applied to an index level of 4,000, that is about 275 points in either direction as a simplified range estimate. The model assumes scaling that real markets need not obey, and an implied range is not a guaranteed interval.

VIX often rises during equity stress, but the relation can vary. Its level also reflects option pricing and risk premiums, not a direct count of frightened investors. A spike may coincide with a market low or with the start of a larger decline. Do not call every high VIX reading a buy signal. For the distinction between volatility and price trend, see Technical Indicators.

Surveys and outside data

An investor survey asks a defined group for a view. Before using it, note the sampling population, question wording, response rate, and release schedule. If 120 of 300 respondents are bullish, the bullish share is 40%. That is a statistic about respondents, not about all investors or the dollars they control. A survey can be early, late, or persistently optimistic during a trend. Historical percentile comparisons should use a consistent methodology.

News and advisory sentiment measures classify published text or recommendations. Automated scoring can misread sarcasm, changing language, and repeated syndication. A backtest must use information at the publication timestamp rather than an after-the-fact revised data set.

COT: reported positioning with a delay

The CFTC Commitments of Traders page explains that reports summarize open interest by trader classification and that the weekly data generally represent Tuesday positions released Friday. The classification reflects predominant business purpose, not the purpose of each individual trade. A report is thus a delayed positioning snapshot, not a live order-flow monitor.

Suppose a reported group holds 80,000 longs and 55,000 shorts. Its net position is +25,000 contracts. If the previous net was +10,000, the change is +15,000. It would be misleading to call this an exact directional bet by every group member. Check whether you are using legacy, disaggregated, or financial-futures categories; they are not interchangeable. Also compare net position with total open interest and its own historical range.

A disciplined synthesis

Make a three-column note: indicator observation, alternative explanation, and price condition. Example: “High put/call volume; may reflect hedging rather than outright bearish bets; act only after a close above resistance and a risk level below support.” If COT positioning and surveys disagree, preserve the disagreement. Sentiment measures have different populations and timing.

A contrarian signal should be tested against a predeclared baseline. Count all extremes, including those followed by further trend movement. Avoid choosing the percentile, smoothing period, and exit date after seeing which combination worked. If the chosen indicator does not improve decisions after costs or loses usefulness in a later period, drop it. The purpose is to manage uncertainty, not to collect colorful indicators.

Common mistakes and self-check

Do not interpret put volume as net put buying, VIX as a directional forecast, COT as real-time data, or a survey percentage as a probability of market rise. Do not mix index and equity put/call series. A sentiment extreme can persist longer than a position's risk budget.

Self-check: A put/call volume ratio is 0.75 and call volume is 800,000. What is put volume? Answer: Puts = 0.75 × 800,000 = 600,000. The ratio alone says nothing about whether the calls were bought or sold.

Next, choose one sentiment series, write its exact numerator, denominator or survey population, and publication lag, then state a separate price condition that would change your view.