ForexEAsMall
Newbie
- Messages
- 2
- Likes
- 0
Hey everyone,
I’m Jepoji. I’ve been working with Forex EAs for a few years now, and like many of you, I started out thinking backtesting was just clicking a few buttons and looking for high profits on the report.
Turns out, there’s a bit more to it. 😅
I wanted to share how I personally backtest EAs in MT4/MT5 – not in a super technical or textbook way, but how I actually do it, with all the small things I’ve learned through trial and error.
Step 1: I Always Start With Clean, Realistic Data
I used to run backtests straight from MT4 without thinking twice. Now I always double-check the data quality. If I can, I use Tickstory or something similar to get more accurate tick data (especially for scalpers).
In MT5, it’s a bit better because the built-in strategy tester already supports real ticks. If you're using MT5, take advantage of that.
Step 2: Use Visual Mode (Yes, It Takes Longer, But...)
When I test a new EA or even just a new setup, I turn on Visual Mode.
It slows things down, but it helps me see how the EA is actually opening/closing trades, and whether it reacts weirdly to news spikes or sudden volatility.
Sometimes a backtest looks good on the report but watching the trades visually tells a different story. One EA I tested had amazing stats, but when I watched it trade, it would open 10 positions right before NFP. No thanks.
Step 3: I Look at the Equity Curve First, Not the Profit
I made this mistake early on: focusing too much on net profit. Now, I zoom in on the equity curve and drawdown.
A smooth, stable curve with low drawdown is way more important than a high profit number.
If the curve looks like a roller coaster, I pass – even if it ends in profit.
Step 4: I Backtest Over Multiple Years
If an EA only works in one year or one market condition (like trending only), I know it’s probably curve-fit.
I test at least 2–3 years, and sometimes break it down by quarter to see how the system holds up in different environments: ranging, trending, high news volatility, etc.
Mistakes I’ve Learned From (So You Don’t Have To)
Anyway, just wanted to share this because I see a lot of people backtesting just to find the "best result" – but that’s not the goal. The goal is to find something stable, logical, and that you can trust when the market gets crazy.
If anyone here has tips or their own routines for testing EAs, I’d really love to hear them too. We all have different angles, and I’m still learning every day.
Cheers,
Jepoji
I’m Jepoji. I’ve been working with Forex EAs for a few years now, and like many of you, I started out thinking backtesting was just clicking a few buttons and looking for high profits on the report.
Turns out, there’s a bit more to it. 😅
I wanted to share how I personally backtest EAs in MT4/MT5 – not in a super technical or textbook way, but how I actually do it, with all the small things I’ve learned through trial and error.
Step 1: I Always Start With Clean, Realistic Data
I used to run backtests straight from MT4 without thinking twice. Now I always double-check the data quality. If I can, I use Tickstory or something similar to get more accurate tick data (especially for scalpers).
In MT5, it’s a bit better because the built-in strategy tester already supports real ticks. If you're using MT5, take advantage of that.
Step 2: Use Visual Mode (Yes, It Takes Longer, But...)
When I test a new EA or even just a new setup, I turn on Visual Mode.
It slows things down, but it helps me see how the EA is actually opening/closing trades, and whether it reacts weirdly to news spikes or sudden volatility.
Sometimes a backtest looks good on the report but watching the trades visually tells a different story. One EA I tested had amazing stats, but when I watched it trade, it would open 10 positions right before NFP. No thanks.
Step 3: I Look at the Equity Curve First, Not the Profit
I made this mistake early on: focusing too much on net profit. Now, I zoom in on the equity curve and drawdown.
A smooth, stable curve with low drawdown is way more important than a high profit number.
If the curve looks like a roller coaster, I pass – even if it ends in profit.
Step 4: I Backtest Over Multiple Years
If an EA only works in one year or one market condition (like trending only), I know it’s probably curve-fit.
I test at least 2–3 years, and sometimes break it down by quarter to see how the system holds up in different environments: ranging, trending, high news volatility, etc.
Mistakes I’ve Learned From (So You Don’t Have To)
- Overfitting parameters – I once had an EA that performed insanely well on one pair with magical settings... only to blow the account live.
- Ignoring spread/slippage – In backtest, it’s easy to forget these. In real life, your broker’s conditions matter a lot.
Anyway, just wanted to share this because I see a lot of people backtesting just to find the "best result" – but that’s not the goal. The goal is to find something stable, logical, and that you can trust when the market gets crazy.
If anyone here has tips or their own routines for testing EAs, I’d really love to hear them too. We all have different angles, and I’m still learning every day.
Cheers,
Jepoji
