A custom symbol can reproduce an instrument’s prices without reproducing the hours when its underlying market accepts trades. In MetaTrader 5, trading sessions are part of a custom symbol’s specification, separate from its imported bars and ticks. That distinction matters when an Expert Advisor is tested against a synthetic instrument and later connected to a broker’s live CFD.
For mt5 developers, user-defined sessions are not merely labels on a chart. They establish a model of market availability that must be consistent with timestamps, historical prices and the strategy’s execution assumptions. Five mismatches can distort the results.
Trading Sessions and Quote Sessions Serve Different Purposes
A custom symbol can have quote sessions describing when prices are available and trading sessions describing when orders may be accepted. The two schedules need not coincide. An instrument may publish indicative quotes during a period when its exchange is closed, for example.
If an EA treats every incoming tick as permission to trade, it can generate entries that would be rejected on the live instrument. Session information should be queried explicitly, rather than inferred from the presence of a price update. Custom symbols are primarily for analysis and testing; configuring their sessions does not create a broker-executable market.
Server-Time Assumptions Can Shift the Opening Window
Session definitions use the symbol’s trading schedule as represented by the terminal, while historical data must be aligned to the intended time basis. Importing timestamps from another time zone without conversion can shift the apparent market open. Daylight-saving transitions complicate the mapping when the underlying exchange changes clocks but the data source does not.
An algorithm designed to trade the first fifteen minutes of a session may consequently operate on pre-opening or late-session prices. Compare imported timestamps with an independently known exchange calendar, including days when the market closes early.
Missing Overnight Gaps Can Make Stops Look More Reliable
Imagine a hypothetical regional electricity-sector index CFD with a live trading session ending at 16:00 and reopening at 09:00. A custom symbol is mistakenly configured as tradable around the clock. Its imported series contains an indicative 17:00 quote of 2,480 after a 16:00 close at 2,500, and the next executable opening price is 2,455.
A backtest of a long CFD worth $4 per point might record a stop exit at 2,480, showing an $80 loss from 2,500. In live trading, a stop triggered at the reopening could fill at 2,455, a $180 loss, excluding costs. The $100 difference comes from assuming executable liquidity during a closed session, not from a faulty stop calculation.
Session Changes Can Alter Indicator Inputs
Two datasets with identical daily closing prices can produce different intraday signals if one includes out-of-session ticks and the other excludes them. Moving averages, opening ranges and volatility estimates may respond to additional bars or differently grouped sessions. An apparently more complete dataset can therefore be less representative of the market the strategy intends to trade.
For mt5 backtests, review whether imported off-hours observations should remain visible as quotes, be excluded from indicator calculations or be handled by explicit EA logic. The correct choice depends on the underlying instrument, not simply the number of available ticks.
Custom Schedules Cannot Override Live Broker Restrictions
A custom symbol’s sessions are user-maintained metadata. A live CFD’s trading hours, holidays, maintenance breaks and order restrictions come from the broker’s instrument specification. Matching a custom schedule to ordinary weekdays does not guarantee matching behavior during an unscheduled halt or exceptional holiday.
Before opening a CFD position based on a custom-symbol test, compare the custom quote and trading sessions with the live symbol’s published hours, inspect the timestamp conversion, and rerun the strategy across at least one market closure and reopening. Confirm that its simulated exits occur only at prices that could reasonably have been executable.