Local spread and ATR data atlas
Explore your own spread and ATR observations without uploading them. Filter a pair, source, ATR definition and fixed UTC session window, then export the selected data and a reproducible descriptive report. No real market dataset is bundled with this tool.
The demonstration is synthetic and must not be cited as observed market data. Imported files are user supplied and are not independently verified. A public benchmark requires verified provenance and a consistent sampling protocol.
Paste CSV or inspect the import format
Required columns: time,pair,spread_pips,atr_pips,source,account. Optional: atr_timeframe,atr_period. The aliases spread and atr also mean pips. Use comma-separated CSV and a decimal dot. Time must include Z or an explicit UTC offset. Quoted fields, embedded commas and escaped double quotes are supported.
No dataset loaded.
The tool is loading. If it does not become ready, enable JavaScript and reload this page.
Processing stays in your browser. No account or upload is required. Results depend on your inputs and are educational calculations rather than investment advice.
What can be cited
This tool provides the calculation method and a way to analyze a local sample. It does not provide independently collected spread history, verify brokers or rank execution quality. A synthetic demonstration can illustrate the method; it cannot support a claim about real spreads, sessions or a currency pair.
To publish a useful benchmark, document the observation source and collection dates, broker and account type, instrument specification and pip convention, bid/ask capture method, timestamp resolution, sampling frequency, missing periods and exclusions. For ATR, also record the timeframe, lookback, smoothing method and completed-bar convention. Retain the raw file and explain any data cleaning. User-entered source labels are provenance claims rather than verification.
Metrics and grouping
Spread-to-ATR ratio per observation = spread in pips ÷ ATR in pips.
Median = the 50th percentile.
P90 = the 90th percentile, using linear interpolation at sorted index (n − 1) × 0.9.
Spread and ATR must use the same pip definition within an observation. The ratio is calculated for each row before its median and P90 are calculated; it is not a ratio of aggregate medians. Zero ATR observations remain in the spread and ATR sample but have no ratio. The report shows the ratio sample count separately. Repeated timestamp, pair, source, account and ATR identifiers are flagged and retained. Closely spaced observations can be dependent, so n is a row count rather than a count of independent events.
Summary tables keep pair, account context and ATR definition separate. Account currency is retained as context; no monetary conversion is performed. Spread is a quoted observation rather than a guarantee of executable price, and ATR is a historical range measure rather than a maximum future move. Comparisons require consistent sources and sampling.
Session convention and dates
The reporting windows are fixed in UTC: Tokyo 00:00–09:00, London 08:00–17:00, New York 13:00–22:00 and Sydney 22:00–07:00. The start is included and the end is excluded. Windows overlap, so an observation can belong to several windows. These are reporting conventions without daylight-saving adjustment, not live exchange opening hours. The selected sample’s first and last timestamps, declared sources and filters are included in the exported report.
Related tools
Trading cost sensitivityPortfolio exposureTrading sessions and time zonesStrategy testing protocol
Method and tool reviewed 26 October 2026. Editorial methodology · About the author
