How Does Spread By Session Work in Forex?

Direct answer

“Spread by session” in forex means that the bid–ask spread you observe (the difference between the buy price and sell price) may vary depending on the time window you call a “session.” The core idea is simple: during hours with more participants and tighter order books, spreads often tend to be narrower; during hours with fewer participants or higher volatility, spreads often tend to be wider. The exact way a provider defines sessions, and the way it measures spreads, can differ.

Mechanism and definition

What a spread is

In forex, a provider typically quotes two prices for the same instrument: a bid (the price at which you can sell) and an ask (the price at which you can buy). The spread is the ask minus the bid. In practice, the spread you pay depends on the quoted liquidity and execution environment at the moment your order is matched or priced.

What “by session” adds

A “session” is a time window (for example, the part of the day associated with particular geographic trading activity). “Spread by session” means you break time into these windows and compare spreads within each window.

A plain way to think about it:

1. Choose a session definition (start/end times, time zone, and whether weekends or specific holidays are treated differently).
2. Collect bid and ask quotes (or recorded spreads) over time.
3. For each session window, compute a measure of spread, such as an average or median spread.
4. Report how that measure differs between windows.

The “output” is therefore a set of spread statistics indexed by session window. It is not a guaranteed relationship; it is a descriptive pattern based on the data used.

Inputs, outputs, and sequence

Inputs you need to understand

1. Session boundaries: exact start and end times, plus the time zone used.
2. Spread data source: whether the provider uses real-time quotes, historical records, or modeled values.
3. Measurement rule: average vs median; using all observations vs filtering out extremes; how missing data is handled.
4. Cost components: some providers combine spread with other trading costs (commissions or fees). “Spread by session” may isolate only the bid–ask component.
5. Execution context: spreads can reflect order-book depth at the time of quoting or can change between quote and execution.

Output you should expect

A “spread by session” description usually results in one or more of the following:

• A mapping from session window → typical spread statistic.
• A statement that spreads change “during” certain sessions.
• Sometimes a comparison (for example, relative tightness in one window vs another), but the exact numbers depend on the measurement approach.

Sequence for independently explaining it

To explain the concept without implying results:

1. Define the instrument and what “spread” means (ask minus bid).
2. Define the session windows and time zone.
3. Explain the data collection period (for example, a historical range) and the measurement statistic used.
4. State the descriptive output (typical spread varies by session window).
5. Clarify the limitations: outcomes depend on changing market conditions and the provider’s specific definition.

An important assumption in any example is that you use consistent rules across sessions; otherwise, the comparison may reflect your methodology rather than market behavior.

Evidence and a worked, non-promissory example

Consider a hypothetical dataset of quoted bid and ask spreads for one forex instrument. Assume you choose three session windows and use the median spread in each window to reduce sensitivity to occasional spikes.

• Session A (morning hours in your chosen time zone)
• Session B (overlap hours with higher cross-market participation)
• Session C (evening hours)

If, in your historical sample, the median spread in Session B is lower than in Session C, you can say: “In this sample, the typical (median) spread was narrower during Session B than during Session C.”

Two material limitations apply even in this clean example:

1. Volatility coupling: session windows may overlap with scheduled news releases or high-impact events. If volatility drives wider spreads, the observed difference may be an event effect rather than a time-of-day effect.
2. Provider methodology: a provider may compute “spread by session” from different quote sources, time zones, or filtering rules. Your computed stats may not match the provider’s if definitions differ.

Limitations and risks (what can go wrong)

1) Liquidity is not only about the clock

Even if a session window usually has higher participation, liquidity can change within the session due to news, sudden risk-off or risk-on moves, technical issues, or changes in participant behavior.

2) Quotes vs executions

A spread you see in a quote can differ from the spread you experience when executing an order, because execution depends on size, order type, and whether your order can be filled at the quoted prices.

3) Time zone and session definition mismatch

If one party uses a different time zone or different session boundaries, “spread by session” can look different even for the same underlying market.

4) Measurement choices can mislead

Average spreads can be distorted by occasional widening. Median spreads reduce the effect of outliers, but you may also lose information about how frequently spreads widen sharply.

5) Historical relationships may not hold

A common failure mode is assuming that because Session B was tighter in the past, it will be tighter in the future. That assumption can fail when volatility regimes or market structure changes.

Verification and next questions

You can verify the idea by checking spread behavior in your own dataset:

• Compare bid–ask spreads across the exact session windows you define.
• Use consistent measurement rules (same statistic, same filtering approach).
• If you are comparing across providers, separate bid–ask spread from other trading costs so your comparison is like-for-like.

Next, clarify what matters for your goal of understanding costs: do you want the typical quoted spread by session, or do you want the effective cost of execution (which can depend on commissions, slippage, and fill quality)? These are related, but they are not always identical.

If you can state your session definitions, the spread statistic used, and the data window, you can explain “spread by session” accurately and independently check whether the described pattern holds under your assumptions.