Modeling quality of 90 percent and what it hides
Posted: Sat Oct 31, 2026 8:53 pm
The MT4 strategy tester prints a modeling quality percentage. A figure of 90 percent sounds reassuring, and many traders treat it as a stamp of approval.
In reality, the number reflects how much of the test period was covered by real tick data versus interpolated data. A 90 percent figure can still contain long gaps and artificial ticks.
Worse, the quality number says nothing about the accuracy of spreads, slippage, or the realism of fills. A perfect tick stream with a fixed spread still gives unrealistic results.
For scalping, the weak spot is the data source. Much of the tick data for MT4 is generated, not recorded, and its microstructure does not match live markets.
Anyone relying on modeling quality should check the source of the data, whether it contains real bid and ask, and how gaps are filled.
A cross-check is comparing a handful of backtest trades with what actually happened in the live feed at the same time, for the period where both exist.
How much trust goes into a modeling quality number, and what else gets checked?
In reality, the number reflects how much of the test period was covered by real tick data versus interpolated data. A 90 percent figure can still contain long gaps and artificial ticks.
Worse, the quality number says nothing about the accuracy of spreads, slippage, or the realism of fills. A perfect tick stream with a fixed spread still gives unrealistic results.
For scalping, the weak spot is the data source. Much of the tick data for MT4 is generated, not recorded, and its microstructure does not match live markets.
Anyone relying on modeling quality should check the source of the data, whether it contains real bid and ask, and how gaps are filled.
A cross-check is comparing a handful of backtest trades with what actually happened in the live feed at the same time, for the period where both exist.
How much trust goes into a modeling quality number, and what else gets checked?