A 24-hour price change compares two observations within a moving window. It can differ from a calendar-day return, and the percentage can change because the earlier reference moves. Read the provider’s definition, pair and timestamp before interpreting a daily gain or loss.
An unchanged current price can still have a changing 24-hour percentage when the earlier reference changes.
BTKWIN.COM · EDUCATIONAL EXAMPLERolling window versus calendar day
A rolling comparison generally relates an observation to one about 24 hours earlier under the provider’s method. A calendar-day comparison uses a defined daily boundary. They can produce different numbers at the same clock time without either calculation being inherently wrong.
Record the timezone for calendar boundaries and the timestamp for the observation. A daily candle and a 24-hour ticker field need not share the same start. Do not label every daily figure as the same return.
Read the source’s field definition. Coinbase’s price-page explanation illustrates that a page can contain several separate statistics, including price, market capitalisation and 24-hour volume. Their units and meanings still need to be kept distinct.
Calculate the change from its actual start
For a positive starting price, percentage change is the ending price minus the start, divided by the start, times 100. A hypothetical move from 100 to 110 is 10%. A move from 110 back to 100 is about negative 9.09%, because the denominator changed.
Keep full precision in the underlying observations and round the displayed result. If a source reports a slightly different percentage, inspect its timestamp, field and rounding before assuming an error. A last trade sampled a few seconds later can differ.
The percentage-change guide explains denominators and successive returns. A daily percentage is a market statistic, not automatically the result on your cash.
Understand why the percentage changes
| Fictional snapshot | Current price | Price 24 hours earlier | Calculated change |
|---|---|---|---|
| First | $110 | $100 | +10% |
| Later | $110 | $105 | +4.76% |
In the invented example, the current price stays at 110 while the earlier comparison becomes 105. The displayed percentage falls even though the current quote did not fall. The window’s moving start explains the change.
This is why a screenshot showing a lower daily percentage does not by itself prove a price decline since the previous screenshot. Compare the actual current values and timestamps. State which question you are answering: change since your observation or change over the provider’s rolling interval.
High and low are not your achieved trade
A 24-hour high and low describe observed extremes within the stated window. They do not establish that you could buy the whole intended quantity at the low and sell it at the high. The exact order, timing and available liquidity remain unresolved.
The order of extremes also matters. A summary table can show a range without describing the path. A strategy that assumes the low occurred before the high may be impossible for a particular interval.
Use the candle guide to understand interval summaries. A range is not a realised return, and selecting perfect endpoints after seeing the data creates a different claim from a predefined trading method.
Keep pair and source consistent
BTC/USD and BTC/USDT use different quote assets. A converted local-currency display adds another rate. A derivative, spot venue and aggregate index can also have different series. Match the instrument before comparing percentages.
Different venues can record different trades, and a feed can be delayed or interrupted. Preserve source, field and event time where available. The source-check guide provides a reproducible observation record.
If one provider’s daily figure is calendar-based and another’s rolling, label the difference. Averaging the two percentages does not resolve the incompatible definitions. Choose a consistent measure for the question instead.
Volume describes activity, not your available exit
A 24-hour volume field measures activity over the provider’s period and scope. Check whether it is expressed in base units, quote currency or a converted value. One venue’s figure and an aggregate are not interchangeable.
High volume over a past day does not guarantee current depth at the price you need. Orders can disappear, spreads can widen and service routes can become unavailable. Execution depends on the actual quantity and conditions.
The liquidity guide separates past activity from available orders and completed fills. Keep that distinction when a ranking sorts assets by daily trading volume.
Your investment result needs your transactions
If you bought at another time or made several purchases, the provider’s 24-hour return does not describe your acquisition cost. Use actual units, cash paid and fees. A later sale adds execution and exit costs.
A hypothetical chart price rise from 100 to 120 is 20%, but a 1,000-dollar fee-inclusive purchase with 1% fees on both sides produces about 17.62% net return under the calculator’s convention. The price movement and cash result are different calculations.
Label an unsold position’s marked result unrealised. A current valuation does not establish that the entire quantity can be sold and withdrawn at the selected number. Keep the hypothetical example separate from the current external feed.
Save a usable daily observation
Record the pair, venue or methodology, observation time, timezone, field definitions and comparison start. Preserve the original value rather than overwriting it when the window moves. Note missing or stale data explicitly.
Use the snapshot worksheet to compare assets or dates consistently. A useful 24-hour statement identifies its reference and limits. It does not turn a moving statistic into a forecast or a guarantee of the next day’s outcome.
Preserve two questions in a daily log
A useful log can show both the provider’s 24-hour change and the change since your own earlier observation, with separate baselines. Label each one explicitly. Otherwise, a moving statistic can be mistaken for what happened during the time you were watching.
If the earlier reference is unavailable, do not reconstruct it from a rounded percentage without acknowledging the estimate. The provider may use more precise values or another sampling convention. Keep the original published field and its definition alongside your own calculation, and explain any unresolved difference. That makes the record useful without claiming a precision the source does not support.
Review the log after the interval closes and preserve the original observations.
Continue reading
English content and examples reviewed: 27 October 2026. Read the Arabic edition.
