In this article
After this article, you can
- Check whether a whole bar fits a study window, beyond its opening timestamp.
- Calculate an earlier window’s range from completed shorter bars.
Terms at hand: Backtesting · Trading Plan · Liquidity · spread (bid-ask spread) · Major Pairs
A morning range comparison depends on the exact time window and when each bar becomes complete. A bar opened before London activity starts can still contain later data. Define those intervals before deciding whether a price moved beyond the overnight range.
What actually happens in the market each morning
Tokyo and London are informal desk-activity labels, not centrally enforced exchange sessions. Their overlap depends on the working-hour conventions and date selected. Overlap alone does not establish the deepest liquidity or the best trading conditions.
A BIS geographical turnover share refers to the location of reporting sales desks, not to the volume of a named intraday session. It cannot show how much liquidity arrives at a particular London opening minute or prove that one session takes over another.
An opening timestamp does not place the whole bar inside a window
The export timestamp 06:30 marks the H1 opening. This bar covers [06:30, 07:30), so its intersection with a study window ending at 07:00 requires finer data.
Scroll sideways to see the full diagram
All OHLC data and observation times are fictional. The study window is [00:00, 07:00) UTC; earlier completed bars have High no greater than 101. M15 O/H/L/C: 99/100/98/99.5; 99.5/101/99/100; 100/108/99/107; 107/109/106/108. H1 opens at 06:30: this explicitly assumed boundary offset is nonstandard for some data sources. TradingView — Time confirms that bar opening and closing times are distinct, and an opening timestamp does not describe the entire interval.
A filter accepting “H1 opening time before 07:00” would include this bar and its later High of 109. In this sample, the range through 07:00 can be calculated from the two completed M15 bars: its High is 101. The full H1 has O 99, H 109, L 98 and C 108, but those values are complete only at 07:30. For an export, separately check timestamp meaning, closing time and aggregation boundaries. Setting the time zone correctly is not enough.
Why the overlap is so short
Define Tokyo as 09:00–18:00 JST and London as 08:00–17:00 local time for this illustration. Japan uses UTC+9 without seasonal clock changes. The UK and Poland change clocks on 29 March and 25 October in 2026. On 27 March and 26 October, Tokyo converts to 01:00–10:00 Warsaw and London to 09:00–18:00: the intersection is 09:00–10:00, one hour. On 30 March and 23 October, Tokyo is 02:00–11:00 Warsaw, giving 09:00–11:00, two hours. These dated conventions do not certify holiday access or a fully staffed desk. Defining Tokyo’s end as 17:00 JST instead shortens the shared interval by an hour.
The London–New York overlap also needs dated local-hour assumptions because seasonal offsets can differ. Compare observed spreads, available depth and execution for the chosen pair rather than assuming a fixed three-hour CET interval always gives the cheapest or most decisive trades.
Define the earlier range before evaluating a break
European desk activity and scheduled releases can change trading conditions. Determine the effect in a defined sample; a session label alone does not establish that volatility must increase.
At the assumed London start, Tokyo’s full-session high and low are still unknown: it ends one or two hours later under our convention. Instead choose an earlier reference interval, here [00:00, 07:00) UTC, excluding quotes from 07:00 onward. Its seven hours convert to 01:00–08:00 Warsaw in winter or 02:00–09:00 in summer; it is not the whole Tokyo session. Require the eligible data to be complete and available before the decision. Observe a separate first London hour, 08:00–09:00 local or 09:00–10:00 Warsaw, to avoid comparing a one-hour winter sample with a two-hour summer sample.
Use a hypothetical EUR/USD sample with pip = 0.0001. Reference Bid ranges from 1.0990 to 1.1020, a width of 30 pips. A later first-15-minute close is Bid 1.1023, three pips above that high; assume a purchase of EUR 10,000 executes at Ask 1.1025, with a two-pip spread. Assume the hour-end sale fills at Bid 1.1011: (1.1011 − 1.1025) × 10,000 = USD −14. An assumed USD 0.80 round-turn commission leaves USD −14.80, with no financing. Spread is already included in those execution prices and is not subtracted again. This is one invented path, not a typical morning or complete live-risk plan; a break does not guarantee continuation or the next fill.
Boundary inclusion can also change the condition. Assume all earlier Bid quotes are no higher than 1.1020, but the quote exactly at 07:00:00 UTC is 1.1050; evaluate a later Bid of 1.1030. The chosen t < 07:00 interval has high 1.1020, so the later quote is ten pips above it. A different t ≤ 07:00 interval includes 1.1050, leaving the same later quote twenty pips below the high. Label the rule before testing; time precision and bar-open/bar-close labels are separate issues. Neither comparison proves a profitable trade.
Compare instruments with the same measurement rules
EUR/USD and GBP/USD can respond to European news, but neither must be quiet overnight or become most active at the London open. Compare the instrument’s dated observations and release calendar before choosing it for a method.
EUR/JPY and GBP/JPY combine currencies associated with Asian and European activity. Their costs and price behaviour must still be measured for the account and date. The Tokyo and London guides explain the labels; they do not establish a universally best beginner pair.
Your next step
- Compare a defined overnight interval with a dated first-hour interval using completed bars. Record price range and costs without assuming that the arrival of European desks explains every difference.
- Convert both morning and afternoon windows for the observation date. Measure the pair’s spread and execution conditions in each; the afternoon overlap does not guarantee the lowest cost.
- Record the overnight high and low at the chosen boundary using only completed eligible data. Track subsequent breaks under a stated rule, including false breaks, rather than assuming a successful signal.
- Record matching Bid and Ask observations in several dated windows, together with commission and execution costs. Compare the resulting sample; it does not establish that the afternoon is always cheapest.
Check your understanding
One short exercise
About 2 min · you can try again
An H1 bar opened at 06:30 UTC ends at 07:30. M15 bars completed by 07:00 have High 100 and 101; later M15 bars have High 108 and 109. What is the earlier window’s High in this model?
- 109, because H1 opened before 07:00.
- 108, because only the last M15 is excluded.
- 101, from the two completed M15 bars before the boundary.
Check the answer and explanation
101, from the two completed M15 bars before the boundary.
Earlier bars have High no greater than 101, and both pre-boundary M15 bars are complete. Their maximum of 101 determines the window’s High.
Frequently asked
Why is the Tokyo–London overlap so short?
The overlap depends on which desk-hour windows you define and the date. Tokyo uses Japan Standard Time throughout the year, while London changes between GMT and British Summer Time. Convert the windows to a common time zone before comparing them. A permanent 08:00–09:00 CET interval is not a global schedule.
Is the Tokyo–London overlap a good time to trade?
That requires evidence from the actual instrument, account and method. There is no established universal 30–40-pip overnight range, guaranteed morning breakout or afternoon spread advantage. Define a reference window using only then-available data, keep the later observation length consistent and include failed breaks, costs and fills. In our hypothetical EUR 10,000 purchase/sale, a three-pip Bid break is followed by a USD −14.80 net result under the stated assumptions; it illustrates possibility, not frequency.
How should I compare pairs during the morning?
Use the same dates, sampling method and account conditions to compare price range, Bid/Ask spread, charges and actual fills. A euro, sterling or yen label does not prove the strongest reaction, cleanest trend or lowest cost. Include announcements and holidays separately. Choose the time boundary before seeing the result; a chart’s opening timestamp does not make its later high available at the start.
Tokyo–London or London–New York overlap?
Define both overlaps from dated local-market windows. Tokyo–London involves the end of Asian desk activity and the start of London activity; London–New York involves concurrent European and North American activity. Neither has one universal CET duration or a guaranteed spread advantage. Compare the actual instrument, costs and execution in consistent observation windows.
Sources & bibliography
-
Bank for International Settlements Triennial Central Bank Survey 2022 — geographical distribution of FX turnover · BIS geographical FX turnover data for 2022. www.bis.org ↗
-
Kathy Lien (Wiley) Day Trading and Swing Trading the Currency Market, 3rd ed. · A discussion of forex sessions and the opening of European trading activity. www.wiley.com ↗
-
Bank for International Settlements Triennial Central Bank Survey 2022 — OTC foreign exchange turnover · BIS information on OTC FX turnover for the 2022 survey. www.bis.org ↗
-
GOV.UK When do the clocks change? · UK government information on British clock changes. www.gov.uk ↗
-
NICT Japan Standard Time · The Japanese time authority defines JST as UTC(NICT)+9. Used only for the model’s time-zone conversion. www.nict.go.jp ↗
-
Japan National Tourism Organization Plan Your Trip: Time Difference · Official travel guidance states Japan uses UTC+9 without daylight-saving changes. Session boundaries remain assumptions, separate from holiday access. www.japan.travel ↗
-
Dziennik Ustaw Poland: regulation on summer-time transitions in 2022–2026 · The official 2026 Polish dates support the stated Warsaw/Tokyo conversions; no permanent future DST rule is inferred. www.dziennikustaw.gov.pl ↗
