A spread is an observation at a moment
The bid and ask quoted by a platform change over time. The difference between them is the spread at that moment. A single screenshot is useful evidence for that observation, but it is not an all-day average or a guarantee about the next quote.
Use a published sampling schedule
Our spread protocol samples at 08:00, 14:00 and 20:00 UTC over five trading days. A complete cycle contains fifteen observations per pair. Publishing the schedule makes the method inspectable and reduces the temptation to select only the most attractive readings.
Compare the same thing
Our summary uses EURUSD on the account named in the file. Other pairs can use different units and should be read separately. Commissions, swaps and slippage are not part of the spread measurement; the headline figure is not an estimate of every trading cost.
Keep the sample count next to the mean
An arithmetic mean compresses a set of observations into one number. It can conceal variation within the window. Read the count, dates and underlying log as well as the mean. A partial cycle needs to be understood as partial evidence.
Respect freshness and corrections
A spread observed last month may not describe conditions today. The file shows when the latest cycle ran and whether a re-test is due. If a sample is corrected, the original remains in the log with a link to its replacement.
Know what this method leaves out
Fixed-time samples do not cover every market event or capture every quote. These tests describe the observations collected under the published method. They do not predict execution, profitability or the result of using a different account type.