DAY 9: Backtesting and Forward Testing — Differences and How to Use Them

Last time, I talked about the mental challenges that arise when operating an EA and how to address them.
This time, let’s focus on the "verification process" before actually running the EA.
In particular,backtesting andforward testing differences and the advantages and disadvantages of each verification method will help minimize gaps after you start operating.

What is Backtest (Historical Verification)

Backtesting is a method that simulates what trades an EA would have made using past market data.

1. Advantages

   • Verification can be done quickly
     Because you can test long periods—1 year, 2 years, or even longer—in a short time, you can efficiently try many parameter adjustments.

   • Roughly gauge the "hit or miss" of logic with market changes
     Test across different market phases, such as trending vs ranging periods, to understand strengths and weaknesses.

2. Caveats / Disadvantages

   • Overfitting (over-optimization)
     Logics tuned too tightly to past data may not perform well in future markets.

   • Spreads and fill speeds often differ from real conditions
     Backtests in MT4/MT5 assume ideal fills based on historical data, which can lead to discrepancies with real trading environments.

What is Forward Test (Future Verification)

Forward testing is running the EA on a demo account or a small real account,operating the EA in real rate fluctuations and observing the results.

1. Advantages

   • Easily reflect real fill conditions
     Elements such as spreads, slippage, and order delays that can occur in actual operation are included in the test results.

   • Can verify how the logic responds to real-time market changes
     In realistic scenarios where trends suddenly shift or economic indicators cause volatility, you can gauge the true value of the EA.

2. Caveats / Disadvantages

   • Takes time
     You cannot accelerate past data; you must wait in real time, so longer verification periods take more time to obtain results.

   • Limited samples in certain market regimes
     In short periods, the EA's strengths and weaknesses may not be fully apparent. You'd like to observe for at least 1–3 months, but even then you may not capture overall trends.

Complementary Relationship: "Backtest x Forward Test"

In practice, the common flow is tonarrow down the logic with backtesting and then grasp real operation feel with forward testing.

1. Firm up the basics with backtest

   • Test across multiple periods (long and short) and check maximum drawdown and win rate

   • Including different market conditions (rising market, falling market, range-bound) makes it easier to understand the logic’s characteristics

2. Use forward testing to measure real-world deviations

   • Watch for unexpected slippage or order rejections (requotes)

   • How did the EA react to economic indicators and unexpected news?

   • If there is a large difference from backtest results, consider adjustments like lot sizing or parameter tuning

When the Difference Between Backtest and Forward Test Is Large

If you feel performance differs significantly between verification and live operation, check factors such as the following.

• Differences in spreads and commissions
  If the backtest uses fixed spreads, there will naturally be divergence from actual variable spreads.

• Actual market conditions changed rapidly
  In cases of increased volatility or financial shocks not seen during the test period, the logic may fail to keep up.

• Over-optimization of parameters
  Fitting perfectly to past scenarios can cause poor performance under other market conditions. The variety of test patterns may have been too limited.

Today's Summary and Next Episode

• Backtests allow quick, large-scale verification, but they often diverge from actual markets

• Forward tests reflect a more realistic environment and risk, but they take time

• Ultimately, by combining backtest and forward test to evaluate an EA, you canmanage risk in preparation for operation.

Next time (DAY 10), we will delve further into “Items to be especially careful about in backtesting—Spread and Slippage.” We will explore fine-tuning settings for effective backtesting and how to bridge gaps with reality.

Introduction to the EAs I Sell

If you’re considering an EA, please also take a look at the EA I sell.
https://www.gogojungle.co.jp/users/147322/products

Comparing backtest results with forward-test outcomes can help solidify a realistic operation image.

In the next article, we will look more closely at elements that are easy to overlook in backtesting.
Be sure to press "Read more" to deepen your understanding.