Direct answer
To assess a Volatility Ratio, you need input data that captures volatility in at least two comparable ways (for example, over different time windows, or using different volatility measures), plus the exact definition of the ratio you are computing. Because volatility calculations depend on data choices, you also need to document provenance (where the data comes from), timeliness (what period it covers and whether it is complete), and quality checks (how missing values, outliers, and corporate/event effects are handled).
If you want to independently explain and verify your results, the minimum set of “needed data” is: (1) the underlying price series you will use, (2) the observation interval(s) and sampling frequency, (3) the method for converting prices into a volatility estimate, (4) the ratio formula, and (5) the assumptions that connect those steps.
Mechanism and definition (what the ratio needs)
A “volatility ratio” generally compares volatility in one context to volatility in another context. The comparison is only meaningful when the inputs are defined consistently.
Typical data inputs you should be able to state precisely:
- Price series: at least one instrument’s time-ordered prices (e.g., mid, bid, ask, or last). Choose one and keep it consistent.
- Timeframe / lookback window(s): the number of bars or time span used to estimate each volatility component.
- Sampling frequency: whether you compute from hourly, 5-minute, daily, or other bars. Different frequencies change volatility estimates.
- Volatility estimator: how you compute volatility from prices. Common choices include volatility of returns (e.g., standard deviation of returns) or volatility based on ranges (e.g., high–low range-derived measures).
- Return or change definition (if applicable): arithmetic vs log returns, and whether returns are computed from close-to-close or another price field.
- Ratio formula: e.g., “volatility in window A divided by volatility in window B,” using the exact units your volatility estimator produces.
Because multiple “volatility ratio” variants exist, you should treat the ratio definition as part of the required data. Two people using different estimators or windows may compute different ratios from the same price series.
Evidence or example (what to document so the math is verifiable)
Here is a verification-focused example of what you would document, without assuming any specific live market values:
- Assumption 1 (price field): You use the same price field for all calculations (for example, a single type of quote or bar close).
- Assumption 2 (sampling): You compute returns from consecutive bars of a fixed interval (for instance, one bar per hour), with no gaps.
- Assumption 3 (volatility estimator): You compute volatility as the standard deviation of returns over each lookback window.
- Assumption 4 (ratio): You compute Volatility Ratio = (volatility over short window) / (volatility over long window).
To independently verify, someone else needs the same ingredients: the price series, the exact interval(s), the estimator steps, and the ratio formula. If any of these are unclear, replication becomes guesswork.
Limitations and risks (material failure modes)
Volatility Ratio assessments commonly fail for reasons that are not about “math correctness” but about data alignment and assumptions:
- Timeframe mismatch: If the short and long windows are not aligned to the same sampling frequency or if one includes missing periods, the ratio can reflect data artifacts rather than volatility differences.
- Outliers and jump behavior: Sudden price moves can dominate volatility estimators (especially standard deviation of returns). A small number of extreme observations can distort the ratio.
- Non-trading and data gaps: Overnight gaps, holidays, or missing bars change the effective information content. If gaps are filled or removed incorrectly, volatility estimates can be biased.
- Provider-specific microstructure: If you use bid/ask, mid, or last-traded prices, volatility can differ materially due to spread behavior and trading mechanics. The ratio is only interpretable relative to the exact price field used.
- Historical relationship limits: Even if the ratio shows consistent behavior in the past, historical volatility relationships do not establish future results.
These limitations mean you should avoid treating the ratio as a standalone “signal.” Instead, focus on how the computed ratio changes under your documented assumptions.