Volatility signature plot: realized volatility by sampling interval
A volatility signature plot places the sampling interval on the horizontal axis and the corresponding realized-volatility estimate on the vertical axis. If very fine sampling is contaminated by bid-ask bounce or other microstructure noise, the high-frequency end can rise away from estimates at coarser intervals.
- Horizontal axis
- Sampling interval or frequency
- Vertical axis
- Realized variance or volatility estimate
- Main limit
- The curve diagnoses noise; it does not prove one cause
Key Takeaways
- Calculate the same day or window at several sampling intervals.
- Use consistent sessions, return definitions, and annualization across the grid.
- A fine-scale rise can indicate microstructure noise such as bid-ask bounce.
- The plot does not choose a universal optimal interval for every asset and period.
Build the curve
Clean timestamps, select a session, choose a price series, and create returns at a grid of intervals. Sum squared returns within each evaluation window. Repeat for every interval and plot the resulting variance or its square root.
Realized variance grid
for interval Δ in samplingGrid:
returns = logPriceChanges(sample(price, Δ))
realizedVariance[Δ] = Σ returns²Read the high-frequency end carefully
When prices include microstructure noise, using more observations can add noise faster than information. Bid-ask bounce is one possible source. Timestamp errors, stale quotes, discreteness, and asynchronous trading can also affect the curve.
Keep comparisons on one contract
Use the same session boundary, timezone, overnight policy, return formula, missing-data rule, and annualization for all points. Record how irregular events are sampled. Otherwise the curve can reflect preprocessing differences instead of frequency effects.
Platform limits shape the plot
A chart platform may not expose the same historical depth at every fine interval. Pine Script also runs on the data loaded for the chart and requested lower timeframes. Pineify can generate a study for the available grid, but it cannot create tick history or remove vendor limits.
Generate a sampling-grid study
Specify the session, price field, interval grid, return formula, annualization, and missing-data behavior. Pineify can create editable Pine Script for the data TradingView exposes. Inspect history coverage before interpreting the curve.
This page is educational and does not provide investment advice. The methods described are diagnostics or estimates, not trading signals. Results depend on data quality, sampling, market structure, and model assumptions.