Raw versus derived
Raw market data is what a market emits directly: a trade at a given price and size, or a quote with a bid and an ask. Derived data is anything calculated from it. The distinction matters because raw data is a matter of record while derived data is a matter of method. For a derived value, the relevant question is not only its result but how it was calculated and over what window.
Common derived metrics
Most widely used figures are derived. A 24-hour change compares the current price with the price exactly one day earlier and expresses the difference as a percentage. VWAP, the volume-weighted average price, weights each trade by its size to summarize the average price paid over a window. Moving averages smooth price over a chosen number of periods. An OHLCV bar is itself derived, aggregating raw trades into open, high, low, close, and volume.
- 24-hour change: the current price versus the price 24 hours ago, as a percentage.
- VWAP: the average trade price weighted by volume over a window.
- Moving average: price smoothed over a fixed number of periods.
- OHLCV: raw trades rolled into open, high, low, close, and volume.
Why derived values differ between providers
Because derived metrics depend on method, two providers can report different values for the same metric while both represent the raw data correctly. A 24-hour change depends on which timestamp marks the start of the window. A VWAP depends on which trades are included and over what period. A moving average depends on the interval and the number of periods. A disagreement on a derived figure almost always reflects a difference in inputs or window rather than in the underlying market.
Computed values versus precomputed values
Derived data can be obtained in two ways: by receiving the raw inputs and computing the metrics directly, or by consuming precomputed values from the provider. Computing directly gives full control over the window and formula, which matters where the method must be exact or auditable. Precomputed values are faster to adopt and consistent across clients. Many systems use both, computing the metrics that require precision and reading provided values for the rest.
Derived data on SiftingIO
SiftingIO publishes a canonical price tick with a small, fixed set of fields and leaves richer derived metrics to the consumer, which keeps control of the window and formula behind values such as VWAP, 24-hour change, and OHLCV bars on the client side. Where SiftingIO does compute a value, the method is documented so results match. The aggregated fair price is itself a derived value, combining many sources into one reference price.