Socius NewsAsk a trader for their best trading hours and you get a clock, not a number. "London open." "The overlap." "The first hour of New York." The answer is the hour where the most happens. The best trading hours are the hours where you are most active, the reasoning goes, because that is where the volatility is. It sounds right. Your own trade history usually says something else.
Best trading hours, defined: the hours of the day during which a specific trader's closed trades show the highest expectancy, measured in R per trade over a large enough sample. The definition is personal. It is about your trades, not about the market's busiest hours. That distinction is the whole article.
What are the forex session times, in one clock?
The foreign exchange market does not have an opening bell. It has centres that wake up in sequence. The Bank for International Settlements measured 9.6 trillion dollars of daily FX turnover in April 2025. Sales desks in four jurisdictions handled 75% of it: the United Kingdom at roughly 38%, the United States at roughly 19%, Singapore at 11.8% and Hong Kong SAR at 7.0%. The Bank of England puts the UK figure at 37.8%, down from 38.0% in April 2022. Two centres, London and New York, carry more than half the flow. Where they sit on the clock decides what "session" means.
The cleanest map of that clock is 30 years old. Andersen and Bollerslev (1998) took 5-minute Deutsche Mark-dollar quotes from 1 October 1992 to 30 September 1993, 260 weekdays, 74,880 observations, and drew the intraday volatility curve. The market approaches a standstill during the Tokyo lunch, 3:00 to 4:30 GMT. Volatility rises when European trading starts around 6:00 GMT and stays elevated until about 15:00 GMT. The peak sits between 12:00 and 15:00 GMT, when it is afternoon in London and morning in New York. Then it fades. There is no spike at the New York close, unlike the U-shaped curve that equity markets show.
03:00
Tokyo lunch
3:00 to 4:30 GMT: the DM-dollar market approaches a standstill.
06:00
European trading starts
Volatility steps up and stays elevated until about 15:00 GMT.
12:00
London afternoon meets New York morning
The two most active centres trade at the same time until 15:00 GMT. The highest volatility of the day.
12:30
US data slot
US releases at 8:30 Eastern (12:30 or 13:30 GMT with daylight saving) produce the sharpest spikes.
13:30
NYSE core session opens
9:30 to 16:00 Eastern, 13:30 to 20:00 GMT in summer.
15:00
European activity fades
FX volatility declines. No elevated volatility at the New York close.
For stocks the map is written down. The NYSE core session runs from 9:30 a.m. to 4:00 p.m. Eastern Time, with early sessions from 7:00 a.m. (4:00 a.m. on NYSE Arca) and late sessions to 8:00 p.m. That is the session clock. It tells you when the market is busy. It tells you nothing about when you are good.
When is the London New York overlap, and why does it matter?
The London New York overlap is the 12:00 to 15:00 GMT window, 14:00 to 17:00 in Paris in summer, 8:00 to 11:00 a.m. in New York. It matters for one measurable reason: it is where the two centres that carry roughly 57% of global turnover are both fully staffed. Andersen and Bollerslev found that the US macroeconomic releases at 8:30 Eastern, which land inside this window, are the source of the most pronounced volatility spikes of the day. If you want the market to move, this is when it moves.
That is the argument for the overlap. Here is what it leaves out. Volatility is a property of the market. Expectancy is a property of your trades. The overlap gives you more movement per hour. It does not give you more edge per trade. A move that goes 40 pips against you in the overlap costs the same R as a move that goes 40 pips against you at 7:00 GMT. What changes is how often you trade and how fast you have to decide.
Are the best trading hours the busiest hours?
For the market, the busiest hour is the most liquid hour. For you, it is the hour with the most trades in your journal. Two different sentences, usually read as one. The best time to trade forex, for you, is not a market fact.
Barber and Odean (2000) studied 66,465 households at a large US discount broker from 1991 to 1996. The average household turned over 75% of its portfolio a year. The households that traded most earned a net annual return of 11.4% while the market returned 17.9%. The paper is about stocks and years, not hours. The mechanism transfers: more activity, more cost, more chances for the same bias to fire. If your busiest hour is busy because the market is loud, the hour inherits the loudness. The trading expectancy of that hour has to survive the extra trades, not just the extra pips.
The second reason is how you count. Most traders judge an hour by P&L, and P&L by hour rewards the hour where you trade the most size and the most often, because that is where the biggest numbers live, positive or negative. The hour with 60 trades will always have a bigger bar than the hour with 15, in one direction or the other. That bar shows where your money moves, not where your decisions are right.
Why does your P&L by hour lie to you?
Because it mixes three things: how many trades, how big, and how good. Only the third one is about you.
Take the same trade history and count it in R-multiples instead of currency. Each closed trade becomes a ratio: result divided by the risk you defined at entry. A trade that made 1.5 times its initial risk is 1.5R. A trade that hit its stop is -1R. Size disappears. Frequency is still there, but now it is a divisor: R per trade, not R in total. The busiest hour stops winning by volume.
Then the curve changes shape. An hour that looked strong in currency because you sized up during the overlap can turn flat or negative in R per trade. An hour that looked boring can turn out to be the one where your setups work. You do not need a theory. You need the count.
Three rules for the count. First, the same risk definition on every trade, or the R is fiction. Second, a minimum sample per hour before you conclude anything: 15 trades in one hour is a hint, not a fact. Third, clock time in one timezone, and the hour of entry, not of exit. A trade opened at 12:55 GMT that closes at 15:10 belongs to 12:00, because that is when you decided.
What does the afternoon do to a trader who lost in the morning?
The clock also changes you. Coval and Shumway (2005) looked at 1,082 proprietary traders in Treasury bond futures on the Chicago Board of Trade during 1998, more than 5 million transactions. Traders who lost money in the morning were more likely to take above-average risk in the afternoon than traders who had a profitable morning: 31.2% against 27%. They placed more trades, larger trades, and held more inventory. They were about 15% more likely to place trades that moved the price.
Probability of above-average afternoon risk, CBOT T-bond futures traders, 1998
Source: Coval and Shumway (2005), The Journal of Finance
These were professionals on an exchange floor. The bias is not about experience. It is about sequence: a loss early in the day changes the risk taken later. Your afternoon hours carry your morning results. That is why an hour can look bad in your journal for reasons that have nothing to do with that hour's market. It is the hour where your revenge trading lands.
So when you read expectancy by hour, read it twice: once for all trades, once for trades taken after a red morning. If the 14:00 hour is negative only on days where the 8:00 hour was negative, the problem is not 14:00. The problem is what you bring to it.
“The market has a best hour. So do you. They are rarely the same hour, and only one of them is in your journal.”
What your journal would show
Here is what the by-hour view looks like on a connected account. All figures below are illustrative demo data: 205 closed trades over 10 weeks, one instrument, risk defined at entry on every trade, clock in GMT, hour of entry.
64
trades in the 13:00 hour
illustrative demo data
-0.18R
expectancy in the 13:00 hour
the busiest hour
0.31R
expectancy in the 07:00 hour
18 trades, sample too small
-0.42R
expectancy in the 14:00 hour
52 trades, worst of the day
Expectancy by hour of entry, R per trade (illustrative demo data)
Source: Illustrative demo data, Socius Trades
Read the two blocks together. The 13:00 hour is where this trader lives: 64 trades, almost a third of the sample. It is also negative, -0.18R per trade. The 14:00 hour is worse, -0.42R on 52 trades, and it is the hour right after the busiest one. The 07:00 hour is the best of the day at 0.31R, but on 18 trades, which is not yet enough to build anything on. The honest reading is not "trade at 07:00." It is "stop bleeding at 13:00 and 14:00, and give 07:00 another 40 trades before you decide."
A currency P&L chart of the same account would show 13:00 as the tallest bar of the day and 07:00 as a rounding error. The trader would keep following the tall bar.
On Socius Trades this view builds itself. You connect a cTrader or MetaTrader account, the platform imports the executed trades in read-only mode, and the analytics tab shows results by session, instrument, hour and day, in R. Socius AI answers the question in plain language: "what is my expectancy by hour after a losing morning?" There is nothing to type into a spreadsheet and nothing to remember to log. The plans start with a Free tier; Essential is 129.99 euros a year with a 14-day trial. We look. We never touch.
What can you actually do about it?
None of this is a forecast or a promise. It is a set of counting rules that turn a clock into a number you can see.
Count in R, not in currency. Define the risk on every trade at entry, and stick to it. Without that, an hour with a big win and a big size will always look like your best hour, whatever the decisions were.
Group by hour of entry, in one timezone, and set a minimum sample. Below 30 trades in an hour, write "to be confirmed" next to the number and keep collecting. An hour with 18 trades at 0.31R is a candidate, not a conclusion.
Split every hour in two: trades taken after a green morning, trades taken after a red one. If an hour is negative only in the second group, the fix is not the hour. The fix is a rule about what happens after a red morning, for example no new entries for 60 minutes after the daily loss crosses -2R. The rule is yours to set. The journal only shows you whether you kept it.
Compare your busiest hour with your best hour, once a month. If they are the same hour, you have something rare. If they are not, you have a decision, and it is a boring one: fewer trades in the loud hour, and no size increase because the market is moving.
Do not move to the quiet hour because this article said so. It was good for one demo account. Yours is in your broker history. Read it before you change anything.
Trading involves risk of loss. Socius Trades is an analytics tool, not investment advice.