Take a working breakout strategy on EURUSD and run it around the clock. It will make its money between 08:00 and 17:00 London time and give most of it back overnight. Same rules, same symbol, same chart — the only variable is the clock. Session timing is not a detail; for many strategies it is the difference between an edge and a donation.
Three sessions, three different markets
Tokyo is a range market: modest volatility, JPY and AUD pairs at their most active, EURUSD often drifting sideways in a 20-30 pip corridor. London is the volume king — trends start here, ranges break here, and roughly the widest EURUSD moves of the day happen in its first two hours. New York overlaps London for four hours, inherits the momentum, then thins out badly after London closes. The late New York afternoon and the rollover hour are the emptiest, most expensive time of the trading day: spreads widen several-fold exactly when liquidity is thinnest.
Why the same signal means different things
A breakout of the Asian range at 08:00 London has institutional order flow behind it. The same breakout pattern at 04:00 server time is usually noise: there is no volume to carry it, so it travels ten pips and dies. The mirror image holds for mean reversion — fading a move works in Tokyo's ranges and gets steamrolled by a genuine London trend. Neither approach is wrong; each is wrong somewhere. When we test a strategy, the hour-of-day profit table is often more revealing than the equity curve.
The server-time trap
MT5 charts run on broker server time, and most brokers use GMT+2 in winter and GMT+3 in summer so that the daily candle closes with New York. Two consequences bite automated traders. First, "London open" is not a fixed server hour year-round: US and European DST shift on different dates, so for a few weeks each spring and autumn every hard-coded session filter is one hour off. Second, if you move an EA between brokers with different server offsets, a time filter that was profitable becomes a random hour filter. Check the offset, do not assume it.
Practical rules
1. Split every backtest by hour and by session before trusting it. An edge concentrated in one session is normal; an edge spread evenly across all 24 hours is suspicious.
2. Trade breakouts where volume arrives (London, early New York) and mean reversion where it does not (Asia) — or filter the mismatched hours out.
3. Avoid the rollover hour entirely unless your strategy explicitly prices in 3-5x spread.
4. After any broker change or DST switch, re-verify your session filter against actual GMT.
One free helper here: session activity is visible in real time as currency strength — when GBP and EUR lines wake up, London is open regardless of what your server clock claims. Our free meter: https://www.mql5.com/en/market/product/190005
Our products and free utilities: https://www.mql5.com/en/users/app.develop.sk/seller
Our measured backtest data for every EA (profit factor, equity drawdown, trade count, year-by-year results) is published at fxea365.com/ea/ranking
