Forex Sessions Explained: Hours & Overlaps

Forex sessions are the trading windows tied to the business hours of the world’s major financial centers — commonly grouped as Sydney, Tokyo, London, and New York — whose overlapping hours are what let the foreign exchange market trade nearly continuously from Sunday evening to Friday evening UTC. There is no single forex exchange that opens and closes; “session” is shorthand for when a given region’s banks and institutional desks are active, which is also when liquidity and typical volatility in that window tend to be highest. This article explains the session structure itself and what it changes about a trader’s plan and review — it does not recommend a session to trade or a strategy for trading it.

The four sessions and their approximate hours

These are conventional FX-session labels built around each center’s typical local business hours — not exchange-mandated opening and closing bells, since spot forex has no single exchange. The UTC boundaries below are approximate and move with daylight saving: Sydney, London, and New York all shift their clocks over the year, while Tokyo does not observe daylight saving and stays on the same UTC window year-round.

Session	Approximate UTC hours	Notes
Sydney	~21:00/22:00–06:00/07:00, depending on Australian daylight saving	Opens the trading week; lower overall liquidity than the three sessions below
Tokyo	00:00–09:00, stable year-round	Anchors the “Asian session”; Japan does not observe daylight saving
London	07:00–16:00 during British Summer Time; 08:00–17:00 during GMT	The world’s largest FX trading location by global turnover share; overlaps the end of Tokyo and the start of New York
New York	12:00–21:00 during Eastern Daylight Time; 13:00–22:00 during Eastern Standard Time	Overlaps London for several hours; the session close on Friday marks the market’s weekly pause

Because the UK and US daylight-saving transitions don’t fall on the same calendar date, there are a few weeks each spring and autumn when one side has already changed its clocks and the other hasn’t — during those weeks, treat the hours above as directional rather than exact, and check a current market-hours source for the specific week.

Three commonly discussed session overlaps are Sydney–Tokyo, Tokyo–London, and London–New York, and their extent depends on which of the ranges above is currently in effect. None of these are exchange-mandated boundaries — spot FX has no central exchange to mandate one — they’re conventions built on the session-hour definitions above.

Sydney–Tokyo. Tokyo opens at 00:00 UTC while Sydney doesn’t close until 06:00 or 07:00 UTC. Under these conventional session-hour definitions, that puts Sydney and Tokyo overlapping for roughly six to seven hours, not just the first couple of hours after Tokyo opens. Some trading publications describe a narrower stretch, informally around 00:00–02:00 UTC, as when activity is more concentrated; that narrower window is a descriptive convention some sources use, not a separate boundary, and this article doesn’t rely on it for anything beyond noting it exists.

Tokyo–London. Tokyo normally runs to approximately 09:00 UTC, and London opens around 07:00 UTC during BST or 08:00 UTC during GMT. That leaves roughly one to two hours of conventional overlap depending on the season — the point in the cycle where the Asian and European trading days meet. As with the other overlaps, this doesn’t establish that the window carries any particular volatility or presents a trading opportunity; it’s a structural handoff between two regions’ business hours, not a signal.

London–New York. This overlap is where the length actually shifts with the calendar: roughly 12:00–16:00 UTC when both the UK and US are on their respective summer clocks (BST and EDT), and roughly 13:00–17:00 UTC when both are on standard time (GMT and EST). Because the US shifts to and from daylight saving earlier in spring and later in autumn than the UK does, there are short windows each year — a couple of weeks around each transition — where New York is already on Eastern Daylight Time while London is still on GMT, or London has already returned to GMT while New York is still on Eastern Daylight Time. In both cases, London is at its later, GMT-based hours (08:00–17:00 UTC) while New York is at its earlier, EDT-based hours (12:00–21:00 UTC), so the conventional overlap temporarily expands to roughly 12:00–17:00 UTC — about five hours — before settling back into the usual four-hour seasonal window once both regions have changed their clocks.

Why the sessions actually differ

This isn’t a stylistic grouping — it reflects where trading activity concentrates. In the 2025 Triennial Central Bank Survey, global OTC foreign exchange turnover reached approximately $9.6 trillion per day, and the United Kingdom alone accounted for approximately 38% of that global turnover — the single largest share of any location — with the United States at approximately 19%, Singapore at approximately 11.8%, and Hong Kong SAR at approximately 7%; together, those four jurisdictions accounted for approximately 75% of total FX trading.¹ Academic research on intraday volatility in currency markets has documented a related, longer-standing pattern: Andersen and Bollerslev’s analysis of high-frequency Deutsche mark–dollar data found systematic, time-of-day volatility increases tied to the opening of the Tokyo, London, and New York business days, associated with the arrival of new participants and information rather than a random pattern.² That finding is about one historical currency pair and does not establish that any particular modern session or overlap carries the highest volatility across all currency pairs today; it supports the narrower, structural point that intraday FX activity follows systematic time-of-day patterns tied to major market openings, not a specific ranking of current sessions. Neither fact says a given session will move in a particular direction, or that overlap volatility is trading opportunity rather than trading risk — both are true of the same conditions.

That distinction matters because it is easy to convert a structural fact into an implicit strategy claim. “The London–New York overlap has historically carried the highest volume” is a market-structure observation. “Trade the overlap because it moves more” is a strategy claim this article does not make — whether a given trader’s method performs better or worse under higher volatility and tighter or wider spreads depends on that method, not on the session label itself.

Forex sessions, multi-session trading, and session shutdown are three different things

These three terms get used loosely enough that they’re worth separating explicitly, because each answers a different question.

Term	What it actually describes	Where it’s covered
Forex sessions	The market-wide structure of overlapping regional trading hours — a fact about the market, true for every participant regardless of their own schedule	This article
Multi-session trading	One trader’s deliberate choice to run more than one separately planned window in the same day, and how to review those windows without merging them	Multi-session trading
Session shutdown	The predefined act of closing a single trading window: locking the entry gate, reconciling exposure, and logging the evidence	Session shutdown

A trader who only ever trades one planned window a day, and that window happens to sit inside the London session, is using forex session structure to schedule when they trade — not running multi-session trading. Multi-session trading only enters the picture if that trader deliberately runs a second, separately planned window later the same day — for example, a London-session block and a separate New York-session block with their own predefined conditions. In that case, the two frameworks stack: this article helps decide which session windows fit the trader’s plan, and multi-session trading’s review-separation test applies to keeping those windows distinct afterward. Either way, whichever window a session ends up being, it still needs the same shutdown sequence at its close.

What session structure changes about a trader’s decisions

None of the following is a reason to prefer one session over another in the abstract — it’s what should be settled in a plan before the session starts, not discovered mid-session.

Spread and available liquidity can differ by session. Typical spread conditions and available liquidity in a thinner session (such as the late Sydney or early Tokyo hours) are not the same as in the London–New York overlap, so the same nominal position size can face different transaction-cost and execution conditions — realized slippage can differ even though the plan’s own stop distance and sizing rules haven’t changed. A plan written around one session’s typical conditions may need explicit adjustment, not silent reinterpretation, before it’s applied in a different one.
Attention and alertness are not constant across time zones. A trader based in a time zone where a given session falls overnight or very early is making decisions on a different sleep and attention baseline than one trading during their own daytime hours. That is a personal-schedule constraint to plan around, not a property of the session itself.
Session-hopping can substitute for a stopping rule. Chasing a session change — moving into London because the current session went quiet, or into New York because London didn’t produce a trade — is a way an undefined stopping condition gets rationalized as “waiting for a better window” instead of being named as continued activity past a boundary the plan should have set.
Daylight-saving shifts move the boundary itself. Because the regions don’t change clocks on the same dates, a session’s UTC start and end time can shift for a week or two out of sync with the others. A plan that references a session by clock time, not by which region’s session it is, can silently drift out of alignment with the actual market condition it was built around.

Reviewing performance by session, not by the clock alone

The same aggregation problem multi-session trading describes for a single trader’s multiple daily windows applies, in a different form, to a trader who regularly trades across genuinely different sessions. A trade log that only records a timestamp treats a London-overlap trade and a thin Asian-session trade as the same kind of observation, when the liquidity and typical volatility underneath them are documented to differ. Tagging trades by session — not just by time of day — lets a review separate “this setup underperforms” from “this setup was evaluated mostly in a session where the underlying conditions were unusually thin,” which are different conclusions requiring different fixes. This doesn’t mean every trader needs a session-by-session breakdown; it matters most for a trader who regularly trades more than one named session and wants to know whether a result is about the setup or about the conditions it was tested under. A forex trading journal can carry that session label next to the platform’s server-time timestamp, so the tag survives a time-zone mismatch.

A worked example

The following is a hypothetical illustration, not a report of an actual trader, holiday, or market event.

A discretionary trader based in a UTC+2 time zone builds a plan around the London–New York overlap, roughly 13:00–17:00 UTC during standard time, because it fits their own daytime hours and their method was developed watching conditions typical of that window. The plan specifies position sizing appropriate to that window’s typical spread and volatility.

One week, a national holiday in a major London-session country thins liquidity during what the trader’s log still records as “the London session.” The trader’s usual size, unadjusted, produces wider slippage than normal on two trades. In review, because the log notes which session and which unusual condition applied, the trader can attribute the wider slippage to an atypical liquidity day within the usual session — a condition worth flagging for next time a similar holiday falls in that window — rather than concluding the strategy itself degraded.

Failure modes

Failure mode	What it looks like	Why it matters
Treating overlap volatility as automatic edge	Sizing up specifically because it’s “the overlap,” without the plan otherwise calling for it	Converts a liquidity fact into an implicit strategy decision that was never actually tested
Session-hopping	Switching to a different session’s market whenever the current one isn’t producing a signal	Substitutes for a stopping rule and can extend activity well past a planned session length; low-opportunity session overtrading covers the narrower quality-of-opportunity pattern
Untagged review data	Trade logs record time-of-day but not which regional session or notable condition (holiday, data release) applied	Conditions that differ structurally get averaged together, hiding which one actually explains a result
Clock-time plans that ignore daylight saving	A plan keyed to a fixed clock hour drifts out of alignment with the actual session it was built around	The plan keeps running at the labeled time while the market condition it assumed has already shifted by an hour

What this doesn’t resolve

Two related questions sit just past where this article’s scope ends. First, a scheduled high-impact news release can land near a session boundary — for instance, a data print during the last minutes of London before New York opens — and produce volatility that isn’t cleanly attributable to either session’s usual pattern; a plan should decide in advance how it treats a boundary-adjacent release, since after the fact it’s easy to relabel the outcome to fit whichever explanation is more comfortable. Second, the market’s Friday-close-to-Sunday-open weekend gap is a distinct liquidity event from any session’s normal open, with its own gap-risk considerations for exposure carried across it — a question this article does not cover in depth.

Where Costante fits

Costante supports the behavioral side of trading around session structure: a trader can define a predefined trading window and its guardrails in advance — which hours the plan applies to, the rules that govern it, and the cutoff or shutdown point where the window ends — and use low-friction logging to record which session and any unusual condition applied to a given trade, so a later review can separate a setup problem from a conditions problem.

Costante does not identify session overlaps, automatically detect or tag which forex session a trade occurred in, configure or recommend per-session position sizes, adjust sizing automatically for volatility, recommend a session to trade, or provide any signal about when to enter or exit. Which sessions to trade, how to size within them, and how a session is logged remains the trader’s own decision and own data entry.

Frequently asked questions

What are the forex trading sessions?

They are the trading windows tied to the business hours of major financial centers, most commonly grouped as Sydney, Tokyo, London, and New York. Their overlapping hours are what allow the foreign exchange market to trade nearly continuously from Sunday evening to Friday evening UTC, with liquidity and typical volatility varying by which session (or combination of sessions) is currently active.

What is the best forex session to trade?

There isn’t a universal answer. The London–New York overlap is widely regarded as the most active and liquid overlap window in the 24-hour FX cycle, but higher liquidity and volatility change execution conditions rather than guaranteeing a better outcome for any specific method or superior strategy performance. Which session or overlap fits depends on a trader’s own strategy, risk tolerance, and schedule — not on volume alone.

Do trading sessions apply to markets other than forex?

The same underlying idea — activity concentrating around when major regional markets are open — is present elsewhere, but the specifics differ. Futures and equity markets have their own defined exchange hours and after-hours or extended-hours windows rather than a continuous 24-hour cycle built from overlapping regional sessions the way spot forex is.

Costante provides educational workflow tools, not financial advice. Trading involves risk.