EA Recommendations ⑧: Thoughts on EA parameter settings, explaining common NG settings beginners tend to make from a developer's perspective
Good morning, everyone. I’m Renmi.
This is the eighth article. In sections ⑥ about choosing an EA and ⑦ about backtesting and forward testing, I’ve talked about it. For those of you who think, “Okay, I’ve chosen an EA. I understand backtesting. Let’s get it running right away!” please read this content.
The previous article is here.
https://www.gogojungle.co.jp/finance/navi/articles/117632
When you actually run an EA, the first thing you face is “parameter settings.” If you do this haphazardly, you might end up with no results at all despite buying a good EA… This time, from a developer’s perspective, I’ll explain common NG (Not Good) settings and the reasons.
What exactly are parameters?
An EA has several "settings values." These are the parameters.
Typical ones include,
Lot size（how many lots per trade）
Slippage tolerance（how many pips of deviation to tolerate）
Filter conditions（trading times, days of the week, etc.）
Stop loss/take profit width（where to cut or take profits in pips）
There are more. The number of settings varies by EA, but every EA certainly has at least a lot size setting.
NG settings beginners tend to make
NG① Increasing the lot size too much
This is the most common pattern.
The thinking is, “The recommended setting is 0.1 lots, but I want to earn more, so I’ll set it to 0.5 lots.” I get the feeling, but this is very dangerous.
Raising the lot size meansnot only the returns but the risks expand by the same multiple. Even an EA with a maximum drawdown of 10% under the recommended setting would have a drawdown of 50% if you multiply the lot size by five. Losing half of your account is mentally very hard.
There is a reason developers recommend certain lot sizes. The correct order is to run with the recommended settings first and then examine real performance before considering adjustments.
NG② Changing parameters after a small loss
This pattern is, “The results have been poor recently, so I’ll tweak the parameters.”
EA cannot determine good or bad from short-term results. To make statistically meaningful judgments, you need at least several dozen to several hundred trades as samples. Changing parameters just because you’ve had 10 losses in a row is like flipping a coin 10 times and deciding this coin is more likely to come up heads because you saw 7 heads.
Every time you tweak parameters, a reset occurs and you’ll never be able to make an accurate evaluation for a long time.Resist the urge to tweak—that’s the correct approach.
NG③ Using default settings without understanding their meaning
This is surprisingly easy to overlook.
When you purchase an EA, default values are already loaded, but that doesn’t mean they’re optimal for any account or margin. In particular,the lot setting must be adjusted to match the margin.
For example, with 100,000 yen margin, 0.1 lot is different from 1,000,000 yen margin in terms of risk proportion. If the default settings were created assuming a certain margin, you must adjust to your own funds; otherwise risk will be too high or too low.
Make it a habit to check the recommended margin before purchase on the sales page.
NG④ Over-optimizing (parameter overfitting)
A bit advanced, but important too.
MT4/MT5 Strategy Tester has an “optimization” feature that automatically adjusts parameters to maximize performance. However, it’s dangerous to use the exact “optimal values” that come out of that.
Overfitting issues can occur not only in the EA itself but also in parameter settings. Parameters tuned to fit past data often do not work in future markets.Remember: the parameter that yields the best backtest result is not necessarily the best parameter for live operation.
So, how should you set it?
Simply put,start with the developer’s recommended settings as they are.
The developer knows the EA best. The recommended settings encapsulate long-term testing and real-world experience. First, run with them and observe real performance over months, and then, if necessary, consider careful adjustments.
By the way, for my EA (GTX, etc.), I clearly state the recommended lot and recommended margin on the sales page. I also publish real-trade results, so you can use them as参考 for settings.
GTX real-trade results https://real-trade.tech/accounts/52392
GTX sales page https://www.gogojungle.co.jp/systemtrade/fx/34527
Summary
Start with the recommended settings for lot size. Increase only after seeing real performance
Don’t tweak parameters after a small loss. You need a sample size that allows statistical judgment
Always check the assumptions of the default settings (such as the recommended margin)
Using values from optimization tools as-is is dangerous
Even if you choose a good EA, misconfiguring it can ruin everything. Conversely, if you set it correctly, you can unleash the EA’s true power. Don’t rush—operate according to the rules.
See you next time!
Renmi’s EA List
https://www.gogojungle.co.jp/users/189446/products