What data is needed to assess ATR?

Mechanism and definition of ATR

Average True Range (ATR) is a volatility measure built from historical price movements. To assess ATR, you need to know exactly which price series and which “true range” definition a method uses.

In the common approach, True Range for each period is based on the current high and low and their relationship to the previous period’s close. ATR is then computed as an average of True Range over a chosen number of periods (for example, N periods). The assessment therefore depends on:

• The raw OHLC inputs (open, high, low, close) for each period.
• The aggregation choice for ATR (length N and the averaging method).
• The handling of “previous close” across gaps or missing candles.

Because ATR is derived from past ranges, it is a descriptive statistic. It does not automatically represent future movement, and it is sensitive to how the input data are formed.

What data you need to assess ATR

To independently assess ATR, collect the following inputs and metadata.

1) Price data (OHLC) with clear construction rules

You need a time-ordered series of candles or bars that includes at minimum:

• High, low, and close for each period.
• The previous close for the True Range calculation.

State the data construction method: whether candles are based on exchange time or provider time, and whether any adjustments (such as corporate-action style adjustments in other markets) are applied. In FX, the main issue is usually consistent candle formation and timestamping rather than corporate actions.

2) Timeframe and session boundaries

ATR depends on the timeframe of the bars used to compute it (e.g., 1-hour vs 1-day). You must record:

• The exact timeframe.
• Whether the dataset includes full or partial periods.
• How weekends and holidays are handled (missing bars vs carry-forward behavior).

If a provider uses different market hours or different break handling, the ATR values can differ.

3) ATR parameters: period length and averaging method

To reproduce ATR, record:

• The lookback length N (the number of periods used).
• The averaging approach (simple moving average or a smoothing method, if applicable).
• Whether the computation starts from a fixed initial window or uses a recursive initialization.

Without these parameters, two ATR “values” may reflect different computations rather than different market conditions.

4) Data provenance (who supplied it and how it’s audited)

Document where the OHLC data came from:

• Provider or data feed name.
• Whether it matches a known venue’s pricing feed or an aggregated feed.
• Any published methodology for data quality checks.

Even if the formula is identical, different feeds can yield different highs/lows and therefore different True Range values.

5) Timeliness and update policy

ATR is time-sensitive because it updates as new bars form. Record:

• The timestamp of the most recent bar included.
• The update frequency and whether the dataset is “final” or still forming (for example, an incomplete current candle).
• Whether ATR is recomputed after corrections to historical bars.

6) Quality checks on input integrity

Before computing or comparing ATR, perform basic checks:

• Verify time order and consistent spacing between periods.
• Flag missing bars and confirm how they are treated.
• Check for duplicated timestamps.
• Confirm that high ≥ low for every bar.

These steps reduce the risk that discrepancies come from data errors rather than true volatility differences.

Evidence or example of what to verify

A practical way to assess ATR is to verify it through controlled recomputation:

1. Fix the parameters: choose one timeframe, choose N, and specify the averaging method.
2. Use the same OHLC series for the entire history required for the lookback.
3. Compute True Range period-by-period using the stated formula.
4. Compute ATR as the average/smoothed series over N.
5. Compare your result with a second computation only after confirming that both datasets use the same candle definitions, timestamps, and initialization.

If the outputs differ, inspect the inputs first: mismatched timeframe, a different “previous close” caused by missing bars, or different bar completion rules will usually explain most discrepancies.

Limitations and risks (material failure modes)

Several limitations affect how ATR should be interpreted and assessed.

• Regime change: ATR summarises past ranges; volatility can shift due to new information, liquidity changes, or structural market changes.