An hourly candle can show the correct high and low while concealing the sequence of prices that produced them. For an automated strategy, that sequence determines whether a stop was reached before a target, whether a spread widened at entry, and whether a signal existed long enough to trade. Missing ticks can therefore change the apparent economics of a backtest without visibly altering its chart.
In mt5, selecting a real-tick test is not a guarantee that every moment in the requested interval contains original tick observations. The tester’s journal and history quality deserve as much attention as the final profit figure.
Missing Intrabar Moves Can Reverse the Order of Exits
A bar’s high and low do not disclose which came first. When original ticks are unavailable and a test relies on generated ticks for affected intervals, simulated movement follows a modeling procedure rather than the actual sequence. A strategy with both a stop and a target inside the same bar may consequently record a different exit.
Imagine a hypothetical orange-juice CFD representing 1,000 pounds per contract. A long position opens at 320 cents per pound, with a stop at 316 and a target at 326. The hourly bar reaches 315 and 327 cents. If the actual market touched 315 first, the stop would imply a $40 loss per contract at 316, assuming execution there. A simulated path reaching 326 first instead records a $60 gain. The $100 difference comes from event ordering, not from a disagreement over the candle’s extremes.
Absent Bid-Ask Changes Understate Entry Costs
Real ticks can capture changing bid and ask quotes, including temporary spread expansion. When that detail is missing, a reconstructed sequence may not reproduce the spread actually available when a market order would have been sent. A backtest can then attribute a favorable fill to a price that was visible on the chart but not executable on the appropriate side of the quote.
The distortion is especially relevant around session transitions and thin trading periods. Compare the test’s spread assumptions with available historical quotes, and treat unusually consistent transaction costs as a reason to inspect the underlying observations.
Tick-Triggered Signals May Never Have Existed
Some Expert Advisors react to a brief crossing of a threshold rather than waiting for a completed candle. Removing intermediate ticks can erase that crossing; synthesizing a path can introduce one that was never observed. Either change affects the number and timing of trades.
An apparently improved win rate may therefore reflect fewer losing entries rather than better strategy logic. Examine trade timestamps against the available tick record, particularly when the algorithm depends on rapid reversals or tick-by-tick indicator updates.
Quiet Gaps Can Distort Time-Based Trade Management
An EA may close a position after a specified elapsed period, update a trailing stop on each tick, or evaluate a condition only when a new quote arrives. During a gap in observations, those actions may occur later than their nominal schedule. A tester that generates replacement ticks can produce a different pattern of updates.
For mt5 strategies using tick-driven timers, distinguish elapsed clock time from opportunities for the program to execute. Review whether the logic relies on OnTick() alone or has separate timer handling, and inspect behavior around data gaps.
Mixed Data Quality Can Make Optimization Rankings Fragile
One parameter set may trade mostly during well-covered periods while another depends on intervals with reconstructed movement. Comparing their profits as though both faced identical historical evidence can reward sensitivity to the data gaps. Longer backtests do not necessarily solve this problem: adding years with uneven tick coverage can make the comparison less reliable.
Before opening a CFD position based on such a test, inspect the tester journal for real-tick coverage and generated-tick notices, isolate affected dates, and rerun the strategy on a verified continuous interval. Compare exit order, spreads and trade counts before accepting the backtested position size.