A trading schedule should fit both personal availability and the instrument’s actual trading hours. Ending observation is not the same as ending account exposure: a pending order can remain active after the screen is closed. This article connects informal session labels to dated clocks and a final position-and-order check.

Three regions, four centres and separate instrument hours

OTC FX is a network, not one exchange. Three regions—Asia-Pacific, Europe and North America—can include four centres. Define Sydney 08:00–17:00, Tokyo 09:00–18:00, London 08:00–17:00 and New York 08:00–17:00 in each named local zone. These analytical windows are not every bank’s or instrument’s hours. OANDA’s US schedule, for example, separates most FX products, TRY pairs and an extra NZD break. Verify entity, symbol, holidays and permitted order actions; “24/5” does not promise uninterrupted execution.

The BIS April 2025 preliminary release reported that sales desks in the United Kingdom, United States, Singapore and Hong Kong SAR together accounted for about 75% of OTC FX turnover on a net-gross basis. This is a dated geographical statistic, not an intraday session share or a measure of retail trading suitability.

Session labels group hours associated with regional desk activity. To compare a pair across sessions, specify dated windows and measure them directly. ATR is an indicator over its defined bars; it does not by itself allocate a daily move to an overlap window.

The Asian session — Sydney and Tokyo

Preserve the source centre’s date when converting. Sydney 08:00–17:00 on 15 January 2026 becomes Warsaw 14 January 22:00–15 January 07:00; on 15 July it becomes 00:00–09:00 that date. Tokyo’s 09:00–18:00 becomes Warsaw 01:00–10:00 in January or 02:00–11:00 in July. Sydney’s own clock changes fall on 5 April and 4 October: its 12 October window becomes Warsaw 11 October 23:00–12 October 08:00. Polish winter/summer alone misses this third conversion.

Regional releases can affect AUD, NZD and JPY pairs, but activity and costs also depend on global news, the provider and account. Compare spreads, commissions and observed execution rather than assigning a standard pip range to ECN or market-maker labels.

Session boundaries are conventions, not global exchange hours. Record the date, named time zones and provider schedule before converting a window to UTC or your own local zone.

Use the dated official announcement calendars for the Bank of Japan and RBA. The RBA decision is released at 14:30 Sydney local time; its UTC conversion varies seasonally. Do not assign a universal 03:00 UTC release time to a BoJ decision or assume other Asian releases cannot move major pairs.

Convert London hours without using country share as session share

London uses GMT or BST. Its defined 08:00–17:00 window becomes 09:00–18:00 Warsaw on both 15 January and 15 July 2026. New York becomes 14:00–23:00 on those dates, changing relative to Warsaw during the DST mismatches. A dated UK sales-desk share is neither this nine-hour window’s turnover nor proof of its lowest cost.

London links participants trading across regions, while execution and settlement also involve networks elsewhere. The CLS settlement service should not be described as concentrating all interbank settlement in London. A sales-desk location, execution venue and settlement arrangement are different attributes.

Model windows on 15 July 2026, all converted to Europe/Warsaw
Sydney 08:00–17:00 Australia/Sydney00:00–09:00 Warsaw
Tokyo 09:00–18:00 Asia/Tokyo02:00–11:00 Warsaw
London 08:00–17:00 Europe/London09:00–18:00 Warsaw
New York 08:00–17:00 America/New_York14:00–23:00 Warsaw
Sum of four lengths versus time covered36 hours versus a 23-hour union

Exclude each interval’s end. Four nine-hour lengths sum to 36, but the union covers 00:00–23:00. Pairwise overlaps are seven hours Sydney–Tokyo, two Tokyo–London and four London–New York, with no triple overlap: 36−7−2−4 = 23. Merge intervals for availability and avoid counting one observation as independent samples. The union is a clock calculation; its gap proves neither provider closure nor a market-volume or liquidity ranking.

A morning method needs explicit rules and dated data, including costs and the time each bar becomes complete. Session timing does not guarantee technically clean moves or determine most daily trends. Test the method using independently recorded outcomes.

Calculate the dated London–New York intersection

Take the later start and earlier end in a common zone. London is 09:00–18:00 Warsaw. On 16 March and 26 October 2026, New York is 13:00–22:00, making the overlap 13:00–18:00: five hours. On 30 March and 2 November it is 14:00–23:00, giving 14:00–18:00: four hours. New York changes on 8 March and 1 November; Warsaw and London on 29 March and 25 October. These dated windows do not measure turnover, spread or strategy profit.

Convert announcements separately. A hypothetical event at 08:30 America/New_York is 13:30 Warsaw on 16 March 2026 but 14:30 on 30 March; a permanent six-hour addition fails. These are clock exercises, not actual release dates. BLS labels Employment Situation times Eastern Time. The Fed’s 16 September 2026 statement specified 14:00 EDT. Check the document, date and publisher; one source-zone hour does not imply a constant UTC time or universal dealer activity.

Compare a strategy’s complete costs and execution in each candidate window. Spread, commission and slippage can matter even for a small intended move; an advertised spread is not negligible by definition. No universal pip target or ECN spread range establishes the best session.

A dated evening window may fit your schedule. Measure the instrument’s conditions rather than inferring a fixed liquidity drop or spread from the London closing label.

When to avoid trading — price gaps, holidays, low liquidity

Check the instrument’s weekly trading break and the executable prices available when it resumes. News can lead to a gap, and an ordinary stop may fill worse than its trigger. Weekend access and guaranteed-stop terms vary by product and provider; a fixed global closure schedule does not cover every instrument.

Check provider and release calendars for the jurisdiction and year. A stock-exchange holiday or early close neither closes every spot-FX instrument nor establishes a fixed liquidity reduction.

Before the provider’s weekly break, decide whether exposure is permitted by the plan and check active orders. Use an explicit cash-risk budget for any retained position. A wider ordinary stop does not cap gap loss, and a fixed Friday UTC hour does not define risk for every account.

Zero positions is only half the check

In this hypothetical model, observation ends at 18:00. There is no open position, but the server still holds an active, unfilled order to buy EUR 10,000. It has not expired and cancellation has not been confirmed. The model’s end-of-work condition is zero positions and zero active orders. The first count is zero and the second is one, so the condition remains unmet.

Scroll sideways to see the full diagram

No position does not remove an active orderAt the observation window’s end, there are no open positions but one unfilled order to buy ten thousand euros. The model condition requiring both counts to be zero is unmet. A full fill creates one position. Accepted cancellation without a fill leaves both counts zero. Read the state again after every event.End of the observation windowOpen positions0Active orders1 · buy EUR 10,000Model condition: both counts are zeroNow: unmetFull fillPositions: 1Orders: 0Cancellation acceptedPositions: 0Orders: 0Both model paths requirea fresh account-state check.
The lower boxes show alternative events in the model, not a forecast of the order’s fate. A cancellation request alone does not confirm acceptance.

With a full fill and no previous position, the model changes to one EUR 10,000 position and zero pending orders. With accepted cancellation before any fill, both counts become zero. In practice, check the server response and current state: the order may have filled between reading and acting. A session-end label or closing the terminal view does not confirm cancellation. Record expiry and the method for checking positions and orders at the end of your availability in the plan.

Assumptions: This is a fictional state for one symbol and account, with no other positions, orders or automation. The left path assumes a full EUR 10,000 fill; the right assumes cancellation accepted before any fill. No price, activation probability or provider-specific expiry rule is specified. Official source: MT5: order, deal and position.

Picking the session that fits your schedule

Choose observation hours you can attend, then measure dated instrument conditions. Session boundaries do not guarantee a price range or an acceptable cost.

Work, sleep and other commitments determine actual availability. Select a feasible period for the plan, without assuming a universal time budget, popularity ranking or health advantage.

Higher-timeframe rules still require execution and risk controls. Test dated outcomes and record failures; the bar period alone establishes neither signal reliability nor a monitoring deadline.

What to do tomorrow

  1. Convert the observation date with named zones, preserving source timestamps. Review clock changes; practice does not guarantee automatic conversion accuracy.
  2. Map commitments against dated instrument hours and the plan’s monitoring needs. Mark actual availability; employment does not imply a standard daily trading budget.
  3. Specify permitted positions and orders after observation ends. Test the feasible window under those rules; shortening it does not itself improve a strategy.
  4. Use official holiday calendars and verify provider hours. Apply the plan’s risk limits; smaller size or a wider ordinary stop does not automatically justify retained exposure.
  5. Compare the same defined setup across dated windows, recording costs, fills and outcomes. Thirty trades per window describe that sample; uncertainty and independent validation remain necessary.