The best timeframe to trade gold (XAU/USD): what changes from M1 to D1 and how to choose yours

The timeframe is the length of time each candle on the chart sums up: one minute on M1, one hour on H1, one day on D1. With gold, there’s no single best timeframe for everyone. The shorter it is, the more trades you get and the more costs weigh: between August and October 2026, spread and commission could take 4.8% to 14.4% of a typical M5 candle, and only 0.2% to 0.6% of a D1 candle. The longer it is, the further away the stop goes, the smaller the lot size and the more nights you stay in the trade. The right one fits your time, your costs and your capital; with a robot, it’s the one the robot was built and tested on.

What a timeframe is in MetaTrader 5, and how to change it

Each candle sums up the price over a fixed interval: where it opened, how high and how low it went, and where it closed. That interval is the timeframe, and MetaTrader 5’s menus call it “Timeframes”. According to the MetaQuotes help, MT5 has 21 of them, from one minute (M1) to one month (MN1): M for minutes, H for hours, D1 for the day and W1 for the week. To change it:

• The Timeframes toolbar. It has buttons for the most used ones: click H1 and the chart switches to one-hour candles.
• The Charts → Timeframes menu. The most used ones, from “1 Minute” to “Monthly”, with the rest in the “Minutes” and “Hours” submenus.
• Right-click the chart → Timeframes. The same list, without going to the menu.

At the top left of the chart you’ll see the symbol and the timeframe, for example “XAUUSD, H1”. Changing it doesn’t change the market: one H1 candle holds the same prices as twelve M5 candles. What changes is what you see and, above all, what happens to anyone trading on it.

How much gold moves on each timeframe: 2026 data

We measured XAU/USD with Dukascopy’s public price data (one-minute candles of the bid price, plus the ask for the spread) from August 3 to October 30, 2026: 65 trading days. From them we built the candles for each timeframe, with the day starting at 5:00 p.m. New York time, as on many MT5 servers, and calculated the 14-period ATR (Average True Range), which measures how much each candle moves on average (we explain it in where to put your stop loss). COMEX gold futures give very similar figures.

How to read it: on H1, a typical candle moved about $16; if you pay $0.60 per ounce between spread and commission, 3.7% of that move goes on costs before you make anything. On M5 it’s 14.4%, and on D1, 0.6%. The typical candle is the median ATR(14) over the period: half the time it was smaller and half the time bigger. The cost is a cautious range of $0.20 to $0.60 per ounce per trade, spread plus commission: at the low end, a competitive account ($0.13 of spread plus a $7 round-turn commission per lot, which is $0.07 per ounce on 100 ounces); at the high end, $0.60, the median gap between the buy and sell prices in the same data. Yours depends on the broker, the account and the time of day (more in spread, slippage and commissions).

The figures change with volatility: over those weeks gold moved between $3,870 and $4,697. To see your broker’s, go to Insert → Indicators → Oscillators → Average True Range in MT5, which comes set to 14 periods by default.

Noise, costs and number of trades

The smaller the candle, the more of it is noise: swings of a few dollars that mean nothing. And costs don’t shrink with the candle: the spread is the same on M5 as on D1. If on M5 you’re after the move of a typical candle, $4.16, a $0.60 cost takes 14.4% of it; on D1, 0.6% of $94.

In return, short timeframes give you more chances: in our data, each trading day had about 276 M5 candles, 92 M15, 23 H1, 6 H4 and one D1. More candles mean more signals and, almost always, more trades, and each one pays its toll: an M5 strategy that seems to make a little at a time can end up with nothing after costs, while on D1 weeks can go by between one trade and the next. Slippage (the gap between the price you ask for and the price you get) works the same way: a few cents go unnoticed on $90 but not on $4.

Stop loss and lot size: what changes with the timeframe

The stop has to sit outside the noise of the timeframe you trade, and that sets its distance. An example to do the math, not a recommendation: a $5,000 account, $50 of risk per trade (1%), a stop at 1.5 times the typical ATR from the table, a 100-ounce contract (check it in your broker’s Specification window) and the lot size rounded down to a multiple of 0.01 (or the 0.01 minimum, if the formula gives less), using the lot size formula:

How to read it: on M5, with the stop at $6.24, you can trade 0.08 lots and, if it’s hit, you lose about $50; on D1 the stop goes $141 away and not even the smallest lot fits within that $50. On the short timeframe, the lot is bigger and costs weigh more on what you’re risking: on M5, $0.20–0.60 per ounce on 0.08 lots is $1.60–4.80, 3.2% to 9.6% of that $50; on H1, with 0.02 lots, 0.8% to 2.4%. On the long one it’s the other way around: on H4 and D1 the formula gives less than 0.01, the usual minimum lot, and on D1 that minimum already risks 2.8% of the account; for it to be 1% you’d need about $14,100. On small accounts, that forces you to risk more than you planned or to use a stop the noise will eat (the minimum lot and margin).

News, sessions, swap and weekends

News. Of the ten biggest five-minute candles in the period, six started at 8:30 a.m. ET (12:30 GMT), the usual time for US jobs and inflation data, and two came on the afternoon of the Fed decision on September 16 (2:00 and 2:30 p.m. ET). The biggest, on Friday, September 4, covered $97.16 in five minutes: 23 times a typical M5 candle and more than the typical ATR of a whole day. In a jump like that, a $6 stop is filled at the first available price (what happens to your orders when the news hits).

Sessions. On H1, the median candle from 8:00 a.m. to noon ET (12:00–16:00 GMT in this period), with London and New York both open, measured $22.36 from high to low; in the other parts of the day, between $12 and $15. And in Dukascopy’s data, the gap between the buy and sell prices was around $0.56–0.58 from 07:00 to 20:00 GMT and $0.73–0.79 from 22:00 to 06:00 GMT. On M5, the same rule can give a different result in the overlap than during Asian hours (the times are in gold’s trading sessions).

Swap. An M5 or M15 trade usually lasts minutes or hours; an H4 or D1 trade stays open overnight, and every night, when the server’s day rolls over, it pays or earns swap, with a triple charge one day a week. On gold, longs usually pay; your broker’s figure is in the symbol’s Specification (commissions and swap).

Weekends. Gold doesn’t trade from Friday’s close at 5:00 p.m. ET to Sunday’s reopen at 6:00 p.m. ET. Over the twelve reopens in the period, the median gap versus Friday’s close was $3.70, and the biggest, $18.66, came at the reopen for the week of August 31. Next to a daily candle it looks small, but it arrives all at once and the stop is filled at the first available price (the Sunday gap).

Which timeframe to choose: scalping, day trading or swing trading

There are no official boundaries: the MetaQuotes help, for example, puts scalping between M1 and M15 and intraday trading between M30 and H4. These are the most common ranges:

• Scalping, M1 to M5. Lots of trades a day, lasting minutes, with quick decisions in front of the screen. This is where costs, slippage and your connection weigh the most. For traders with free hours, a low-cost account and steady nerves.
• Day trading, M15 to M30. A few trades a day, closed by the end of the day: no swap and no weekends. It means being at the screen during a fixed window, such as the London–New York overlap.
• H1 to H4, between day trading and swing trading. Trades lasting hours or days, checked a few times a day. Costs drop to between 0.6% and 3.7% of the candle, but some trades stay open overnight and over weekends. For traders with a job who can’t look at the chart every five minutes.
• Swing and long-term trading, D1 and above. A few trades a month, which can last weeks. Costs barely show, but the stop is far away, the minimum lot can be too big for your account and swap runs every night. For traders with patience and capital.

If you trade by hand, yours is the one you can follow with a cool head in your free time (we compare both approaches in automated vs manual trading).

Multiple timeframe analysis: looking at several at once

Many traders use a higher timeframe for context and a lower one to make decisions: D1 or H4 to see where gold is heading and which levels it has respected, and H1 to enter. Used well, it helps you avoid buying on H1 against an obvious fall on the daily chart.

Two words of caution. The stop, the lot size and the costs are set by the timeframe you enter and exit on, not the one you use for context. And with three charts, there will always be one telling you what you want to hear: decide beforehand what role each one plays.

Timeframes and Expert Advisors: why you shouldn’t switch

An Expert Advisor is built and tested for one timeframe. Normally, its indicators are calculated on the candles of the chart it’s attached to, and its stops, targets and filters were tuned to how much those candles move. On another timeframe it doesn’t do the same thing faster: it does something else, with different trades, different costs and different winning and losing streaks. It’s a different strategy, and nobody has measured it.

That’s why the backtest has to be run on the same timeframe you’re going to trade: in the Strategy Tester you choose it next to the symbol (we cover it in how to backtest in MetaTrader 5). And testing a robot on five timeframes to keep the one that did best is another way of fitting it to the past (backtesting and overfitting). Use the one its developer specifies.

AUREUM is built and tested on H1. Its installation guide asks you to open the gold chart on one-hour candles: if you put it on another timeframe, it doesn’t attach, and if you change the timeframe with the robot on, it removes itself; either way, it warns you with the AUREUM E12 pop-up. Its published backtest, a simulation on XAU/USD on H1 from 2020 to 2026, has 439 trades and a maximum drawdown on balance of 21.84%. The home page also shows a live account verified by Myfxbook (what “verified” means is explained in how to read a Myfxbook account). Trading on H1 doesn’t make H1 the best timeframe for everyone: it’s simply the one AUREUM was designed and tested on.

How to choose your timeframe for trading gold

There’s no best timeframe for gold, but there is an orderly way to choose yours:

1. Start with your schedule. How many hours you can watch the chart, and when. With twenty minutes in the evening, M5 is out.
2. Do the cost math. Your broker’s spread plus commission, at the hours you trade, divided by that timeframe’s ATR(14). The higher the percentage, the more each trade has to make just to cover it.
3. Check that the lot size works. With the stop that timeframe needs, can you risk what you want with your broker’s minimum lot?
4. Watch the calendar and the clock. The shorter the timeframe, the more news and sessions affect you.
5. If you use a robot, you don’t choose it. It’s the one the robot was designed and backtested on: what you choose is the robot.

And before you put real money in, try it for a few weeks on a demo account, with your broker’s costs.

Frequently asked questions about gold timeframes