Forex Trading Sessions Explained: London, New York, Tokyo, Sydney

A forex chart does not print a session label. The tape just runs, one continuous series from Sunday evening to Friday evening, and the thing changing underneath it is who is awake. Four names get used for those shifts, all of them borrowed from cities: Sydney, Tokyo, London, New York. None of them is a schedule anybody publishes. They are a rough map of when each region's banks, funds and brokers have people at their desks, and the map earns its keep because the same setup on the same pair behaves differently depending on which rooms are staffed.

Quick Answer

Everything below is the detail that paragraph compresses: why a market with no exchange has sessions at all, the full times table across UTC, GMT and ET with the caveats that make it honest, the two weeks a year when the London and New York clocks come apart and the overlap stretches to five hours, why the longest overlap of the day is also the quietest one, and the handoff stretches where a breakout has almost nobody behind it.

Why Does Forex Have Sessions If There Is No Exchange?

Because the sessions are describing people, not venues. Equities have an exchange, so they get an opening bell, a closing auction and a single order book with a published timetable. Currencies have none of that. Spot forex trades over the counter through a network of banks and dealers quoting each other and quoting their clients, which is why the US regulator's own rulemaking on the retail side of it is titled regulation of off-exchange retail foreign exchange transactions. Off-exchange is the operative word. There is no central book to open.

So when someone says the London session opens at 8am, the claim underneath is narrower than it sounds: London's banks staff their desks around then, and the volume they quote starts showing up in the tape. The session is a proxy for headcount. That is the whole idea, and holding onto it explains most of the behaviour further down this page, including why two overlaps of similar length can feel like completely different markets.

The week itself runs continuously. It opens when Sydney comes in on Monday morning local time, which lands on Sunday evening in New York, and it closes at New York's Friday afternoon, around 5pm ET. Between those two points there is no close and no reopen, just the rolling handoff. Then it genuinely stops for about 48 hours, which is the part people forget until a weekend headline reprices something and Monday opens away from Friday's last print.

One consequence worth carrying into the table below. Since no authority publishes the boundaries, platforms draw them differently. Your broker's session shading may start an hour before or after the numbers here, and neither is wrong, because there is nothing to be wrong against. Treat published session hours as a description of a tendency and treat the shading on your own chart as one vendor's opinion about that tendency.

Forex Trading Session Times in UTC, GMT and ET

Before the table, the vocabulary, because two of these words get used interchangeably and one of them is a trap. UTC and GMT are the same clock for anything a trader does with them, so a column of UTC times is also a column of GMT times. Forex trading session times in GMT and in UTC are the same numbers. The trap is your platform's server clock, which is a third thing: most MT4 and MT5 servers run on a fixed offset like GMT+2 or GMT+3 so that their trading day rolls over at New York's 5pm rather than at midnight, which means your candle timestamps are already shifted before any session logic touches them.

The convention below takes each centre's roughly 08:00 to 17:00 local business day and converts it. It matches the hours published on Dukascopy's forex market hours tool and the other widely used session maps to within about an hour per boundary.

The row that does not move is Tokyo. Japan has no daylight saving, so 00:00 to 09:00 UTC holds all year and every other cell in that table shifts around it. If you want one fixed anchor to think in, that is the one. The ET column is written for the northern winter because that is when the conversions are cleanest, and the Sydney and Tokyo rows start the previous calendar evening in ET, which is why a times table that prints clock values without saying which day looks wrong on first read.

What Happens to Session Times During Daylight Saving?

The sessions move, and the annoying part is that they do not all move on the same weekend. The United States switches on the second Sunday in March and back on the first Sunday in November, per the NIST daylight saving time reference. The United Kingdom switches on the last Sunday in March and back on the last Sunday in October, per the GOV.UK clock change dates. Japan does not switch at all. Australia switches in the opposite direction from the northern hemisphere, so Sydney springs forward around the time London falls back.

Put the US and UK dates side by side and two gaps fall out of them. From the second Sunday in March to the last Sunday in March, two to three weeks depending on the year, New York has moved forward and London has not. From the last Sunday in October to the first Sunday in November, about one week, London has moved back and New York has not. In both windows the usual five-hour gap between the two cities narrows to four.

That does something specific and checkable to the overlap. Normally London and New York share four hours: in the northern winter, London runs 08:00 to 17:00 UTC and New York runs 13:00 to 22:00 UTC, which leaves 13:00 to 17:00 in common. In the mismatch weeks New York's day slides an hour earlier in UTC terms while London's stays put, so they share five hours instead of four. The London and New York session overlap is temporarily longer, in the direction most people would not guess.

The Tokyo and London overlap moves harder in relative terms. In the northern winter they share a single hour, 08:00 to 09:00 UTC. In the northern summer London opens an hour earlier in UTC while Tokyo does not move, so the shared window doubles to two hours. Nothing about market behaviour changed. The clocks did.

Which Forex Session Overlap Has the Most Liquidity?

London and New York, and the reason is unglamorous. Two of the largest dealing centres on earth have people at their desks at the same time, so there are more quotes competing on both sides of the price. More competing quotes means a tighter bid-ask spread and more size available at each level. That is the whole mechanism, and it is worth stating plainly because a lot of writing about the overlap makes it sound like the clock itself does something.

The spread is the most direct readout of it you have. The same pair that quotes half a pip during the overlap can quote several times that in the dead hours, and since the spread is a cost you pay before the trade has done anything, a strategy whose edge is measured in a handful of pips is quietly a different strategy in the two windows. Watch your own broker's quoted spread across a full 24 hours once and you will have a better session map than any published table can give you, because it is measured on the liquidity you actually get rather than on the average.

Read the first and third rows against each other, because that comparison is the point of the table. Sydney and Tokyo share six or seven hours. London and New York share four. By length alone the Asian overlap wins comfortably, and by every measure a trader cares about it loses badly. An overlap is worth what the two centres bring to it, and Sydney plus Tokyo is two regional books rather than two global ones. So the heuristic that an overlap means volatility is wrong as stated. The correct version is narrower: the London and New York overlap means volatility, and the other overlaps mean a handoff.

Which is not a reason to write the Asian overlap off. It is a reason to trade it as what it is. A long, liquid-enough, comparatively orderly stretch is a reasonable home for range strategies and a poor home for breakout strategies, and the reverse is true of the London overlap. Most of the disappointment people report about the Asian session comes from running a momentum playbook built for a deep book in a window that cannot supply the follow-through it assumes.

The Stretches Where the Tape Goes Quiet

The overlaps get all the attention and the gaps between them do most of the damage. These are the windows where a chart can look identical to a good one and behave nothing like it, because the structure is being drawn by very few participants.

One correction on the word thin, since it gets read as a synonym for calm. A thin book means fewer resting orders between the current price and the next level, so a given amount of flow moves price further, not less far. Your stop can be reached on volume that would not have touched it three hours earlier, and the spread you pay to get out of the position is wider at exactly the moment you want out. A scheduled economic release landing in a thin window is the worst version of this, because the flow arrives all at once into a book that has nothing in it.

The 22:00 UTC stretch deserves one extra note. Most brokers roll the daily candle at 17:00 New York time, which lands inside that window, and the rollover can print a wick that looks like a rejection nobody actually traded. If your setup depends on a daily high or low, check whether the extreme came from real flow or from the rollover before you build a level on it.

Which Pairs Move in Which Session

A currency sees its heaviest domestic flow when its own country is at work, which sounds obvious written down and gets forgotten constantly in practice. The pairs sort roughly like this.

Asian hours belong to JPY, AUD and NZD. If you trade USD/JPY, the Tokyo session is not the dead zone it is for European majors, it is the window where the yen side of the pair has actual participation. Japanese exporter flow and the Tokyo fix both live in there.

London belongs to EUR, GBP and CHF. The European majors do most of their daily range inside the London session and the first half of the overlap, which is why EUR/USD and GBP/USD look like two different instruments at 3am and at 9am New York time. GBP in particular tends to be the more volatile of the two, and the Asian session flatters it into looking calmer than it is.

New York belongs to USD across the board, plus the entire US data calendar. Most of the releases that reprice the dollar land between 08:30 and 10:00 ET, which is inside the overlap, which is part of why the overlap is as active as it is. Causation runs both ways there: the window is deep because both centres are open, and the scheduling of data into that window exists because the window is deep.

Gold sits in its own category, and the session logic for it is covered in the walkthrough of grading a high-probability XAUUSD setup, which reaches the same conclusion from the instrument side: the London and New York overlap is where gold trades cleanest and the Asian session is where a gold breakout is most likely to be noise. Worth adding one thing that post makes well and this one should not restate: for gold the news calendar overrides the session entirely, so a perfect window thirty seconds before a scheduled release is not a good window.

The honest caveat on all of it is that these are tendencies with exceptions that do not care about your table. A central bank surprise at 2am London time moves EUR/USD in the Asian session, and the session map will have nothing to say about it. Sessions describe the baseline. News overwrites the baseline.

What Session Timing Does an AI Chart Read Actually See?

Worth being straight about this, since the site sells a tool and session timing is exactly the kind of context people assume software has for free.

The rubric genuinely carries session logic. SnapPChart's forex analysis treats the London and New York overlap as the strongest window, treats the London open as a good momentum window, and flags the Asian session as a caution specifically for EUR/USD and GBP/USD. That is written into the grading criteria, not bolted on afterwards, and it is the same judgement this post has been making for eight sections.

What the engine does not have is a clock. It does not know what time you uploaded the screenshot, it does not know your timezone, it does not know your broker's server is running on GMT+3, and it has no idea whether today happens to be one of the weeks when London and New York are out of sync. It reads one static image. The only bridge between the rubric's session logic and your actual chart is the time axis printed on the screenshot itself.

So the behaviour is conditional, and deliberately so. When the time axis or session shading is legible in the image, the read can place the chart in a session and score it accordingly. When it is not legible, the prompt instructs the model to leave the session field empty rather than guess, because a confidently wrong session read is worse than an absent one. Crop the axis out of your screenshot and session simply drops out of the analysis. Leave it in and it becomes one input alongside the trend, the pullback, the levels and the reward-to-risk. Which chart states a single screenshot genuinely carries and which it only infers is the subject of the wider guide to how AI reads a chart, the forex upload flow specifically is walked through in the piece on uploading a forex chart picture instead of describing the setup, and a neutral description of what one chart read covers sits on the AI chart analysis page.

The limitations that stay yours: it cannot tell you a UK bank holiday has hollowed out the London session while the clock says open, it cannot know about the release scheduled in forty minutes, and it cannot verify that the timestamps on your candles mean what you think they mean given your server offset. Session context is cheap for a human to supply and impossible for an image to imply. Supply it.