How can Session Liquidity be measured?

Session liquidity: what you measure

Session liquidity is a market’s ease of trading during a defined time window for a geographic/market “session” (for example, periods when major regional activity is concentrated). To measure it, you need measurable fields and a repeatable procedure.

A practical way to define what you measure is to separate:

• Liquidity state indicators (what you can observe): trading activity, available price levels, and transaction costs.
• Time definition (when you measure): the exact session start/end, and the timezone.
• Assumptions (what you assume): how you estimate liquidity when you only have trade data (no full order book).

Common measurable fields include:

1. Traded volume or number of trades within the session window.
2. Bid-ask spread (and spread distribution) around specific timestamps inside the session.
3. Order-book depth at selected price distances (if you have access).
4. Price impact proxy: how much the mid-price or execution price moves for a fixed notional size over a defined interval.

Mechanism: a repeatable measurement approach

You can measure session liquidity by calculating one or more metrics inside the same time window, then comparing them across sessions.

Step 1: define the session window

Choose a start and end time (e.g., based on a timezone) and apply it consistently across days. If the “session” definition changes (for example, daylight savings shifts), your results can change even if market behavior is similar.

Step 2: collect observable inputs with timestamps

For each session window, record metrics at consistent timestamps, such as:

• Average and percentile spread during the window.
• Volume totals and/or volume per minute.
• If available, depth at level(s) (for example, within a fixed number of basis points from mid).

Step 3: normalize where possible

Two sessions may have different overall activity because of broader market conditions. To reduce misleading comparisons, normalize where you can, such as:

• Spread measures scaled by recent volatility.
• Volume measures expressed per minute.

Step 4: compare with consistent execution assumptions

If you use price-impact proxies, you must state how the proxy is computed (fixed size? fixed interval? mid-price reference?). Without consistent assumptions, “better liquidity” can actually reflect a different calculation method.

Evidence or example: what “measurement” looks like

Here is one self-contained example using only concepts that can be computed from recorded market data.

Example: spread-based session liquidity

1. Pick a session window (start/end with a timezone).
2. For each minute (or chosen interval) inside the window, compute the bid-ask spread at that timestamp.
3. Summarize the session with:
   • Average spread
   • Median spread
   • 90th percentile spread (to capture worse-cost moments)
4. Repeat for multiple days and compare distributions.

If you also have depth data, you can complement spread with a depth-based metric:

• Define a fixed price distance from mid (e.g., “within X basis points”).
• Sum available liquidity in that range at selected timestamps.
• Compare the distribution of that depth across sessions.

These methods make session liquidity measurable because each step specifies the window, the timestamp cadence, and the metric definition.

Limitations and risks: where measurements break

Session liquidity measurements can fail for material reasons:

Measurement can depend on data availability

• If you only observe trades and not the full order book, you cannot directly measure “true depth.” You might use proxies like traded volume or executed spread, but they describe realized liquidity, not available liquidity.

Market regime and costs distort comparisons

• Liquidity can vary with volatility, scheduled macro events, and risk sentiment. A session may look “liquid” in a low-volatility regime and “thin” during stress.
• Transaction costs are not only spreads: commissions, financing, and execution latency can change realized outcomes even when published spreads appear similar.

Timestamps and microstructure matter

• A session window defined too broadly can mix high-liquidity and low-liquidity sub-periods.
• Liquidity is not constant inside a session; a single snapshot can be misleading. Using percentiles and time series summaries reduces that risk.

Historical relationships are not predictive

Even if a session typically shows tighter spreads on some days, historical patterns do not establish future liquidity conditions.

Verification and next question

To independently verify session liquidity measurements, check whether your results remain stable when you:

• Change the timestamp cadence (e. g. , minute vs. 5-minute samples).