Building a Composite Price: How a Quote Is Assembled From Several Feeds
Posted: Tue Oct 27, 2026 9:26 pm
A broker's displayed price is rarely a copy of one source. A pricing engine usually combines several feeds, each with a different latency, depth and reliability, into a single composite quote.
The first step is cleaning. Stale entries, crossed quotes and obvious outliers are discarded. The second is weighting: feeds that have recently been accurate, fast, or deep may carry more influence than slower or shallower ones. Some engines select the best bid and best ask across providers, while others compute a blended mid price and then apply a spread around it.
Further layers follow. A volatility-based margin widens the spread when markets turn unstable, an inventory adjustment skews the midpoint, and a markup covers operating costs. The finished quote is therefore the result of a chain of decisions, each reasonable on its own, rather than a pure reflection of the interbank market.
Because each engine makes different choices, two brokers can quote differently even when drawing on similar sources. Differences are most visible during fast moves, when feed weighting and outlier rejection determine which updates are trusted and which are ignored. For anyone comparing platforms, the interesting question is not who has the lowest quiet-hour spread, but how the composite behaves when feeds disagree.
The first step is cleaning. Stale entries, crossed quotes and obvious outliers are discarded. The second is weighting: feeds that have recently been accurate, fast, or deep may carry more influence than slower or shallower ones. Some engines select the best bid and best ask across providers, while others compute a blended mid price and then apply a spread around it.
Further layers follow. A volatility-based margin widens the spread when markets turn unstable, an inventory adjustment skews the midpoint, and a markup covers operating costs. The finished quote is therefore the result of a chain of decisions, each reasonable on its own, rather than a pure reflection of the interbank market.
Because each engine makes different choices, two brokers can quote differently even when drawing on similar sources. Differences are most visible during fast moves, when feed weighting and outlier rejection determine which updates are trusted and which are ignored. For anyone comparing platforms, the interesting question is not who has the lowest quiet-hour spread, but how the composite behaves when feeds disagree.