Volatility

YAML type: volatility | Category: Adaptive

Detects regime shifts in variability (noise), not just mean level.

Description

Compares short-window scale (mad or std) to a longer baseline scale using ratioThreshold.

How it works

Volatility illustration

Volatility compares a short-window scale (mad or std over windowDays) to a longer baseline scale (baselineDays). When the ratio crosses ratioThreshold in the configured mode, it flags a variability regime shift.

The mean level can stay stable while noise rises or falls -- that is exactly what this detector is for.

Setup

Use method: mad for robustness to outliers; tune ratio for how large a volatility jump matters.

Configure detectors on the data view Analysis tab or in YAML. Validate with Data Preview before enabling production schedules.

Settings

Parameter	Default	Description
`windowDays`	`14`	Short window for current scale
`baselineDays`	`60`	Longer baseline scale
`ratioThreshold`	`2.0`	Current/baseline ratio to flag
`method`	`mad`	Scale method -- see Allowed values
`mode`	`above`	Direction -- see Allowed values

Allowed values

`method`

How short-window and baseline scale (variability) are measured.

Value	Meaning
`mad` (default)	Median absolute deviation from the median (robust to outliers)
`std`	Population standard deviation
`stdev`	Alias of `std`

`mode`

Value	Meaning
`above`	Flag increases / values above the comparison (or high side of a band)
`below`	Flag decreases / values below the comparison (or low side of a band)
`both`	Flag either direction

above: current/baseline scale ratio >= ratioThreshold (noise up)
below: ratio <= 1/ratioThreshold (noise down)

Shared series filters

Most detectors also accept optional series filters in params:

Parameter	Purpose
`maxGapAmount`	Skip series when gap amount is too high (`null` = no limit). Catalog create-default is often `5`.
`minValue`	Skip series whose average value is below this (`null` = no limit). Catalog create-default is often `1000`.
`maxValue`	Skip series whose average value is above this (optional; omit/`null` = no limit)
`useSpecialDates`	When `true`, suppress anomalies that overlap special dates

When a filter skips a series, Data Preview shows the reason.

Configuration example

analysis:
  detectors:
    - type: volatility
      params:
        windowDays: 14
        baselineDays: 60
        ratioThreshold: 2.0
        method: mad
        mode: above

Using the detector

Alerts when the process becomes noisier (or quieter) even if the mean is stable.

Use cases

Trading / pricing instability
Ops metrics that become erratic after a release
Detecting calmed regimes after remediation (mode: below)