Manshu HuangQuant Research BlogsCode portfolio
← All research notes

Implied volatility

Implied volatility: from market price to research hypothesis

Examining low IV, directional signals, volume, skew, and term structure through the distinction between a pricing observation and a testable claim.

Low implied volatility makes an option less expensive only under a controlled comparison. It does not, on its own, make the option undervalued. The distinction is the starting point of this review.

The writing of 诸夏的复活 repeatedly combines a directional trend with low implied volatility. Other posts discuss the volatility surface, the effect of news, and a proposed relationship between implied volatility and trading volume. These ideas provide hypotheses to examine. Here I compare selected publicly visible descriptions with exchange education and academic work; I do not present a new backtest or verified trading results.

1. Implied volatility is inferred from a price

An option's implied volatility is the volatility parameter that makes a specified pricing model reproduce its observed market price. The inputs and conventions matter: the underlying, strike, expiry, rates, and exercise or settlement rules must be identified first. CME glossary

Market option price = Model price(underlying, strike, time, rates, σimp)

This is an inversion, not an independent measurement of future realized volatility. The recovered parameter depends on the model. A convention appropriate for a European option on futures is not automatically appropriate for an American-style or average-price contract.

The market price also reflects the willingness to bear risk and the conditions under which the option can be traded. Calling IV a forecast without qualification hides that distinction. It is an informative price-based quantity, but not a promise about what the underlying will subsequently do.

2. Three different meanings of “low”

Comparison What it measures What it does not establish
Low absolute IV The numerical level under a specified convention A universal bargain across different assets
Low historical percentile Its rank within a defined, comparable history That the current regime should resemble the past
Low relative to a valuation view A gap against a model or forecast after matching the exposure That the gap survives model error and trading costs

The author's distinction between absolute and relative low IV is useful, but an absolute threshold such as 20% cannot be transferred mechanically between equity indices, natural gas, and other markets. The distribution of risk and the way a volatility quote is defined differ.

Historical comparisons also need consistent moneyness and maturity. A front-month contract rolling toward expiry is not a constant-horizon series. Mixing those observations can create an apparent signal from changing contract composition.

For ordinary vanilla options with positive vega, holding the other inputs fixed, a lower volatility input reduces the model premium. CME makes that conditional relationship explicit. It is a pricing sensitivity, not a proof of positive expected return. CME discussion of implied volatility

3. IV, direction, and win probability are separate questions

The source descriptions contain a tension. One post says that a buyer's win rate is inversely related to IV; another says IV affects cost and payoff odds but does not determine win rate or direction. These are not interchangeable statements.

A controlled payoff comparison helps clarify the issue. If the terminal-price distribution and contract are held fixed, paying a smaller premium improves profit outcomes. In observed markets, however, lower-IV periods can also have smaller subsequent moves. The distribution is not held fixed. An empirical relationship between IV and win rate therefore needs a defined instrument, exit rule, sample, and treatment of costs.

A scalar volatility level also does not encode a signed price forecast. That does not prove that the entire option surface contains no directional information. Skew and other surface features can be tested for predictive content. CME's study of CVOL skew is an example of framing that question empirically rather than declaring either universal predictability or universal impossibility. CVOL skew research

Win rate is only one outcome. The magnitudes of gains and losses, tail exposure, holding time, and capital required also matter. Neither a high win rate nor a large option-return multiple establishes an attractive strategy by itself.

4. Read the surface, not a single number

The author's surface discussion correctly draws attention to variation across strikes and expiries. But different IVs are not independent prices that can be set arbitrarily. They are linked by the payoff structure and static no-arbitrage conditions.

Gatheral and Jacquier show how SVI volatility surfaces can be parameterized to avoid static arbitrage. This supplies a disciplined way to think about fitting and comparing a surface: smoothness and fit quality alone are not the only requirements. Arbitrage-free SVI volatility surfaces

Three dimensions deserve separate attention:

  • Level: How much volatility is priced around a chosen reference strike and horizon?
  • Skew: How does pricing differ between upside and downside strikes?
  • Term structure: How is risk distributed across maturities?

CME's CVOL materials distinguish the aggregate index, ATM volatility, and skew measures. Those are different summaries of an option market and should not be substituted for one another without checking the methodology. CVOL FAQ

Calendar and diagonal spreads consequently require more than identifying different IV levels. The legs have different exposure to time, price, and surface changes. An apparent spread in quoted volatility is not automatically an arbitrage.

5. News and volume are candidate explanatory variables

The author's point that a fundamental announcement need not move IV in the same direction as the underlying is useful. The relevant question concerns uncertainty over the remaining option horizon, not simply whether the latest price move was large. An event can move the underlying while resolving uncertainty that had previously been priced into options.

The stronger claim that IV is a periodic function of underlying trading volume requires much more evidence. IV is defined through option prices; that definition is not a deterministic mapping from volume. Volume could still be a useful predictor or conditioning variable. Establishing that would require a stable measurement convention, controls for intraday seasonality and price movements, and an out-of-sample comparison.

Likewise, combining a trend signal with low IV does not permit adding their individual success probabilities. The relevant object is a joint conditional distribution. A claim of interaction needs evidence that the combined rule improves on trend-only and IV-only baselines.

6. A favorable view can still produce a losing option

For a small move, a local sensitivity approximation can organize the sources of a change in option value:

ΔV ≈ Delta · ΔS + ½ Gamma · (ΔS)² + Vega · Δσ + Theta · Δt

Here Theta is defined with respect to elapsed calendar time. The expression omits rates, higher-order interactions, financing, and trading costs. It is an explanatory approximation, not an exact PnL attribution across large jumps. CME's Greeks material similarly separates the effects of the underlying, time, and volatility. Options premium and the Greeks

A correct directional forecast can be offset by a reduction in IV, elapsed time, or execution costs. Conversely, a position can gain from repricing uncertainty even without the anticipated directional move. This is why “the underlying went the right way” is an incomplete evaluation of an option thesis.

Comparing implied and subsequent realized volatility is another useful starting point, often discussed through a volatility risk premium. Cboe's research discusses that distinction. For a variance-based study, the quantities must be expressed in comparable variance units and over matching horizons; a single option's PnL is still not mechanically equal to that difference. Considerations in volatility trading

7. Turning the ideas into a testable research design

The review suggests four experiments, rather than a ready-made trading rule:

Hypothesis Baseline comparison Essential control
Low relative IV improves option outcomes Absolute IV versus historical rank versus a forecast-based measure Consistent horizon, moneyness, and executable prices
Trend and IV interact Trend-only, IV-only, and combined rules Freeze thresholds before the test period
Volume helps predict IV changes A model with and without volume features Seasonality, lag structure, and causal timestamps
Surface information improves decisions ATM-only versus skew/term-structure features No-arbitrage checks and matched risk exposure

These are proposed tests. No returns, Sharpe ratios, significance levels, or win rates are claimed for them here. The source's numerical assertions—including a fixed percentage of time spent in a low-IV regime—are not adopted as facts without a reproducible sample and definition.

The most useful contribution of the blog is a set of questions about what the option market has already priced. Answering those questions requires separating valuation, prediction, and decision-making. A low quote becomes a research opportunity only after specifying what it is low relative to and how that difference could be realized.

Source notes reviewed

The following selected posts supplied the hypotheses discussed above. Their public descriptions were reviewed; the links are attribution, not endorsement of every claim or a reproduction of all image text.

  1. 我的期权体系:既要绝对低隐波,更要相对低
  2. 隐含波动率就是标的成交量的周期函数
  3. 强趋势弱隐波,如何落实到指标上?
  4. 扎心真相:隐含波动率,永远判断不了标的方
  5. 期权醒醒!隐波只管成本和赔率,不管胜率和
  6. 同一标的下的不同期权合约:波动率曲面造就
  7. 期权买方的铁律:买方胜率与隐含波动率成反
  8. 标的基本面变动,为何隐含波动率未必同向响
  9. 期权卖方,吃透极端高隐含波动率,舍此无他
  10. 期权买方,吃透极端低隐含波动率,舍此无他
  11. 上证50ETF期权隐波真相:73%时间在低位,只
← All research notes