What Time Does the Forex Daily Candle Open and Close?
What time does the daily candle open in forex? There is no official close, so every provider picks a cutoff. The two common picks, and how to check yours.
Spot forex has no official close, so every data provider picks its own cutoff, and the answer depends on whose chart you are looking at. The most common pick is 17:00 New York time, which most retail brokers run as a server clock of GMT+2 in winter and GMT+3 in summer. The other common pick is 00:00 UTC.
What time does the forex daily candle open and close?
There is no single answer, because there is nothing to anchor one to. The market-data vendor SiftingIO puts it plainly: "A forex daily candle covers 24 hours of a market with no official close, so every provider has to pick a cutoff."
Two conventions dominate, and SiftingIO lists a third beside them:
| Convention | UTC boundary, winter | UTC boundary, summer | What SiftingIO says about it |
|---|---|---|---|
| 17:00 New York close | 22:00 | 21:00 | "Follows the FX trading day; moves with US daylight saving" |
| 00:00 UTC | 00:00 | 00:00 | "Fixed and reproducible; splits the New York afternoon" |
| Venue-local midnight | Varies | Varies | "Common in charting tools; hardest to reproduce" |
That third row matters if your chart comes from a charting platform rather than from a broker, and it is the one SiftingIO calls hardest to reproduce.
On the broker side the same cut is expressed as a server offset rather than a New York time. TitanFX's glossary states the rule as "midnight-to-midnight server time (00:00 to 00:00)", then adds: "Most brokers set MT4/MT5 server time to GMT+2 (standard) or GMT+3 (DST) so each daily candle closes at 5:00 PM New York time." Its table:
| Period | MT4/MT5 server time | New York time |
|---|---|---|
| Standard (Nov - Mar) | 00:00 - 23:59 (GMT+2) | 17:00 - 16:59 next day |
| Daylight saving (Mar - Nov) | 00:00 - 23:59 (GMT+3) | 17:00 - 16:59 next day |
Both rows check out: midnight on a GMT+2 clock is 22:00 UTC, which is 17:00 in New York in winter, and midnight on GMT+3 is 21:00 UTC, which is 17:00 in New York in summer.
OANDA's help centre says: "MT4 daily candles close at 5 PM EST (GMT+2)." That is two clocks in one line, a city's time and a server's offset, and the sentence does not convert either into the other.
The most precise published time comes from Learn To Trade The Market: "The daily close of the Forex market takes place at 4:59:59 pm NY time", with the new day starting at 5pm as Asia-Pacific gets going.
Why does my chart show a Sunday candle, and why do some weeks have six daily candles?
Because a server running on GMT or UTC opens the trading week on Sunday evening, which is before its own midnight cutoff, so those few hours become a bar of their own.
FXNX's technical blog sets the two cases side by side. On the GMT side: "Brokers operating on standard GMT open their weekly feed on Sunday night (around 9:00 PM or 10:00 PM GMT). This creates a brief 2-to-4-hour trading session before midnight, generating a tiny 6th daily candle". It calls that bar the Sunday micro-candle. On the other: "These brokers align their server clock precisely to 5:00 PM New York. A new daily candle opens exactly as New York closes, resulting in exactly 5 equal 24-hour daily candles per week."
The GMT week reads "Sun Micro-Bar: 3 hrs", then four days of "24 hrs", then "Fri: 21 hrs", which is 120 hours. The New York close week is five bars of 24 hours, which is also 120 hours. A six-candle week is the same week cut in a different place, and it leaves that Friday bar three hours short.
SiftingIO puts the Sunday choice with whoever built the bar: under a New York close those first hours "either become a small standalone Sunday candle or get merged into Monday's bar. That single decision changes whether a week holds five daily candles or six."
How long is the Sunday candle?
Nobody agrees. The published answers run from one hour to four, and on one convention there is no Sunday candle at all.
- One hour. The Sunday candle "opens at Sunday 11:00 PM GMT and closes at 00:00 AM GMT", with the small bodies explained as "only one hour after the Tokyo session opens" (Colibri Trader).
- Two to four hours. FXNX describes "a brief 2-to-4-hour trading session before midnight".
- Three hours. The same blog's week diagram prints "Sun Micro-Bar: 3 hrs".
- None. On a New York close feed those hours may be "merged into Monday's bar" instead, leaving no separate Sunday candle (SiftingIO).
If the length matters to your rules, it is a number to measure on your own data rather than look up.
Does changing my chart's time zone move the daily candle?
Not where the data is cut. The boundary was decided when the bars were built, and whether you can move it afterwards is where the sources flatly disagree, because it comes down to who built them.
Learn To Trade The Market is categorical: "the answer is that you cannot adjust it. The time at which your daily bars close out is set by the broker offering the platform"
SiftingIO does the thing that page says cannot be done, by not taking the daily bar from anyone: "when you need a daily boundary other than midnight UTC you build it from hourly bars. That puts the cut in your own code, where it's explicit and reproducible"
Both are right about their own situation. If a provider hands you finished daily bars, the boundary arrived with them. If you hold the hourly or minute data, the cut is yours to make.
ChartLabs belongs to the first case: our daily bars arrive already cut, and the cut is at midnight UTC. Separately from that, you pick your timezone once in preferences, and charts, trade times and every analytics breakdown display in it.
Does daylight saving move the daily close?
Yes, twice a year, if you measure in UTC. SiftingIO lists it among the reasons two providers' bars can differ: "a New York close is a moving target in UTC, 22:00 in winter and 21:00 in summer. A provider that fixes the cut at 22:00 UTC year round matches the New York convention for only half the year."
It names the failure that follows: "Cutting daily FX bars at a hardcoded 21:00 or 22:00 UTC reproduces the New York close for only part of the year. Around the March and November transitions your daily closes drift one hour from any New York close reference, and the error surfaces as small differences in daily returns that are hard to trace."
On the broker side the motion runs the other way. The server offset moves so that the New York time does not: GMT+2 from November to March, GMT+3 from March to November, both mapping to the same 17:00 to 16:59 next day.
An hour twice a year matters whenever the two sides of your comparison do not move together.
How do I find out which boundary my own charts use?
Three published tests, and each needs something different from you.
| Source | The test it publishes | What you need to run it |
|---|---|---|
| Colibri Trader | Watch the clock: "the easiest way to find out is by watching when the daily candle closes." Then: "If the time is different from 5:00 PM EST, your broker uses another chart feed than the New York Close, most probably the GMT type of chart." | A live chart, and the patience to sit through a close |
| FXNX | Step one of its four-step feed audit: "Open a Daily (D1) chart on your terminal. Count the number of candles in a completed week. If you count 6 candles, your broker is running on an unaligned GMT clock." | One completed week on screen |
| SiftingIO | Resample the same hourly bars twice, once at midnight UTC and once at 17:00 New York, "and compare the close columns. They'll disagree on most days, and that disagreement is the boundary convention made visible." | Hourly data, and some code |
SiftingIO adds the check to run before you report a bad number: "check the boundary time, the Sunday rule, and whether bars are stamped at the open or the close."
A replay turns the first two tests into something you can do on history instead of waiting for a close. On ChartLabs you "step through the historical data candle by candle" (forex backtesting guide), daily charts included, so:
- Open a daily chart and count the candles in a completed week. Five or six.
- Step to a Sunday evening and see whether a short bar exists, and how long it runs.
- Compare one daily close against the 17:00 New York hourly close for the same date.
Then repeat at the start, the middle and the end of the stretch you plan to test. The answer only helps if it holds for the whole sample.
Where ChartLabs cuts the day
At midnight UTC. We use the midnight-UTC convention rather than the New York one, so our boundary does not move when the clocks do.
Our historical data comes from Dukascopy, at tick level where tick data exists, cleaned to remove anomalies and built into 1-minute candles (clean forex historical data). Depth runs to around 22 years on the oldest forex pairs, with other instruments starting later. The longer bars are aggregated up from those 1-minute candles:
- A day runs from midnight UTC to midnight UTC.
- Sunday gets a short daily bar of its own, because those hours fall inside Sunday's UTC day.
That is the six-candle week: in our data the week holds a short Sunday bar and then one bar per weekday. So you do not have to run the three checks on us to find out which convention you are on. You do still have to run them on anything you compare us against.
Our daily-based indicators handle that Sunday bar differently. Pivot Points, Central Pivot Range, Person's Pivots, Average Daily Range and the others built on daily levels also split days at 00:00 UTC, but they fold the Sunday evening hours into Monday. There is no Sunday pivot day, so Monday's levels come from Friday's full session rather than from a few Sunday hours.
Range bars on NAS100, currently our only range-bar symbol, roll on that index's real trading day and follow daylight saving instead of drifting an hour twice a year. That is one index and one bar type. It says nothing about forex daily candles.
Why the boundary matters for a backtest
Because a rule scored on a daily or 4H candle is scored on a boundary. Move the boundary and you are scoring different candles, so some of the setups your rules match stop existing and others appear.
Two pages reach the 4H case from opposite directions. In server time: "On a New York Close server, H4 candles open at 00:00, 04:00, 08:00, 12:00, 16:00, and 20:00 server time", while "On an unaligned server, the H4 bars start at different hourly boundaries, producing completely different engulfing bars, pin bars, and morning stars" (FXNX). In New York time, The Forex Trading Coach derives the same grid from the single cut: "the market opens each new day at 5 p.m. New York time", so the four-hour chart changes at "9 p.m. New York time", then 1am and 5am, giving "six four hour charts" in a day. Six H4 bars either way, counted on two different clocks.
Three more consequences the sources state:
- Candlestick patterns. A short Sunday bar has a small body, and "GMT charts are cluttered with redundant, unnecessary information (candles) which can also be confused with doji candles" (Colibri Trader).
- Daily pivots. Where the previous day was a 3-hour Sunday session, "your Monday Pivot Points will be compressed and practically useless" (FXNX). Ours are not, because our Pivot Points fold the Sunday hours into Monday, as set out above.
- Weekday and date filters. "some sources stamp a bar with its open time, others with its close date, so the same candle can appear under Tuesday in one dataset and Wednesday in another" (SiftingIO). Any test filtered by day of the week inherits that.
In 2014 a broker added a second MT4 server rather than moving the one it had, and the trade press quoted its note on why: "because we wanted to provide as much flexibility as possible while ensuring existing clients were not forced to abandon their existing automated strategies, we have chosen to add an entirely new server to support candles aligned to the New York close". Clients then chose which server to log in to. Moving the boundary on the server they were already using would have broken strategies that were already running.
Which convention is right
We do not have a position on that, and we are not going to tell you your charts are wrong.
SiftingIO, which sells neither a broker account nor a charting platform, puts it this way: "None of these is wrong. What breaks research is mixing them silently, for example validating a daily backtest built on midnight UTC bars against closes quoted on the New York convention." And on where bad-data reports come from: "Most bad-data reports about daily candles turn out to be two correct answers to two different questions."
What to write down before you test
Four lines in your notes, and the test stays repeatable:
- The boundary your data uses, as a time and a zone, and which of the three checks established it.
- The time zone your chart displays, which is a separate thing from the boundary.
- Whether the week holds five daily candles or six, and if six, how long the short bar runs.
- That the same answers hold at the start and the end of your sample, not only where you checked.
Our forex guide already asks you to write down "any filters (time of day, session, day of week)", gives a step headed "Analyze Session Performance", and asks for "at least 100-200 trades" before the numbers mean much. Every one of those filters is cut by the boundary, so the boundary belongs in the same notes as the filters.
And when two sets of results disagree, our article on backtesting accuracy tells you to "Try the same strategy on data from a different source. If the results diverge significantly, data quality is likely the issue." Check the boundary before you reach for that conclusion. Dirty data and a different cutoff look the same in a results table, and only one of them is a defect.
Check it on your own data
Pick a pair, open a daily chart, and count a week. It takes a minute, and it tells you which convention your results have been built on all along.
ChartLabs is a manual backtesting platform for forex, indices, commodities and crypto: you choose the dates and step through real historical data candle by candle. Start your free trial.
Tags
Related
Backtesting Accuracy: How Slippage and Data Quality Matter
Why backtest results and live results drift apart, and what actually closes the gap: clean tick data, honest spread assumptions, and commission and slippage modelled directly on every trade you log.
Forex Backtesting: Test Your Strategy on Real Data
A practical guide to backtesting forex strategies: why session dynamics, spread variation, and pair-specific behavior change your results, what clean historical data actually means, and a five-step process for testing a strategy before you risk real capital.
Clean Forex Historical Data on ChartLabs: Source, Cleaning, and Depth
Where ChartLabs historical data comes from, how we clean it, and how far back it goes. Sourced from Dukascopy at tick level where available, cleaned to remove anomalies, and built into 1-minute candles. Depth is per symbol: the oldest forex pairs reach back around 22 years, other instruments start later.