Direct answer
To assess Raw Spread, you need data that lets you (1) define what “raw” means for the specific context, (2) collect the underlying bid/ask inputs used in the calculation, (3) capture the provenance and timeliness of those inputs, and (4) run quality checks that expose mismatches, hidden costs, and calculation errors.
Because Raw Spread is often discussed alongside other spread concepts, you should separate stable mechanics (how the metric is calculated) from variable conditions (market liquidity, quote timing, and provider execution behavior). For any example calculation, state assumptions explicitly (e.g., the timestamp used, the quoting convention, and what is included or excluded).
Mechanics and definition: what inputs are required
Raw Spread is typically assessed using bid and ask quotes for the same instrument at the same time. The minimum data needed is:
- Instrument and quote context
- The exact currency pair (or instrument) and contract specification used in the quote.
- The price format convention (e.g., decimal places) so you compute the spread in the intended unit.
- The timestamp or time window that anchors “the moment” you are measuring.
- The bid/ask inputs
- Bid price used for the calculation.
- Ask price used for the calculation.
- Whether these are live quotes, indicative quotes, or snapshots, and from what layer (e.g., market data feed vs provider-provided pricing).
- What “raw” excludes or includes Raw Spread assessments are most useful when you record exclusions and inclusions, such as:
- Whether commissions/fees are excluded from the spread number (many “raw” discussions keep them separate).
- Whether any markups, adjustments, or dealing conditions are excluded.
- Whether the metric is intended to reflect only quoted spread, or also any execution-related effects.
Evidence or example: how to structure the check
A practical way to assess whether Raw Spread data is meaningful is to run the same calculation with traceable inputs and then compare the result to the claimed interpretation.
Example structure (no live prices assumed):
- Assumption A: You have bid and ask prices for the same instrument, same timestamp.
- Assumption B: Spread is computed as ask minus bid in price terms.
- Assumption C: You convert to your desired unit (e.g., pips) using the appropriate conversion for that instrument.
Data you record for verification:
- Timestamp used for bid/ask.
- Bid and ask values (with their precision).
- The formula used (ask − bid).
- The conversion method to pips or another unit.
Then perform quality checks:
- Consistency check: bid and ask should belong to the same timestamp or narrowly defined window.
- Precision check: rounding can materially change the result in low-spread regimes.
- Attribution check: confirm the provenance of the quotes (where they came from) and whether they are representative of what a client could observe and execute.
If Raw Spread is reported by a provider, also collect any documentation that defines the metric and its calculation conventions, because “raw” can be presented differently depending on provider terminology.
You may find it helpful to compare with related concepts (for example, total transaction cost vs quoted spread) using the same underlying quote context, so you can see which parts are stable mechanics and which parts are variable conditions.
Limitations and risks: what can go wrong
At least one material limitation to plan for is quote–execution mismatch. Even if bid/ask quotes are correct at a timestamp, actual execution can differ due to latency, order book changes, or execution rules.
Other common failure modes:
- Timeliness drift: quotes taken at different times can create a misleading spread comparison.
- Hidden cost mixing: some datasets combine spread-like effects with commissions or other adjustments, making “Raw Spread” look better or worse than it is.
- Market regime change: historical relationships between spread measures and outcomes do not guarantee future results.
- Variable conditions: liquidity, volatility, and time-of-day effects can change bid/ask behavior quickly, so a single observation may not represent typical conditions.
Because outcomes vary with market conditions, costs, execution quality, and jurisdiction, Raw Spread should be assessed using consistent, documented inputs rather than treated as a single number that predicts trading results.