FX discretionary trading after which I learned about EA

“This time, I should be returning.”
I entered with that belief, but the red numbers kept increasing again.

It went well in the demo, but in the real thing emotions moved first.
I had set a cut-loss line, yet my fingers pressed to buy more with “just a little more.”
In FX, for some reason, after I feel I can do it, it’s not about steady gainsbut a world where a large amount of capital can be lost in one shot.

That happened to me as well.
Having tasted the difficulty of discretionary trading and the fear of money dwindling every day,
I thought there was nothing I could do about it anymore

Then,I learned about EA (Expert Advisor＝automatic trading program).
It isn’t a magic box, butthe idea of shifting “judgment from human fingers to the rules in code” looked like a single path for me after the crushing defeat by emotions.

In this article, I will explain the mechanism of EA in concrete terms, and also
“Can EA become a side business?”, based on publicly available information.
Then I will summarize considerations for commercial EA, self-made, AI-assisted, and operations
(Note: I do not guarantee profits or specific results).

https://www.gogojungle.co.jp/re/EovF5gdbktyC3bU" rel="nofollow">Prosperitas Nexus 30https://www.gogojungle.co.jp/re/img/NsoEwjVKQjDc9gy" alt="30の戦略知能が、一つの判断へ。 | GogoJungle"/>" target="_blank">What sparked it? The limits of discretionary trading and meeting EA

When I started FX, small profits continued, and I easily fell into the illusion that I had a talent for winning..
SNS screenshots, friends’ success stories, the smoothness of demo accounts
They do not teach the real fear of a live run..
When the market once betrayed my expectations, my body moved first.
I knew in my head the stop-loss,
but in front of the screen, knowledge does not lower body temperature.

After losing substantial capital, opening the chart became scary itself.
The heaviness of suddenly closing off all life options leaves a mark deeper than numbers.
Only then did I realize that the loss wasn’t just the method, but something broader.
I, who acted on emotions, also lost—
loss aversion, the rush to recover, the impulse not to leave the screen.
The difficulty of discretionary trading lies not in reading charts but inthis circuit.

When I thought there was nothing I could do,I learned about automated trading. It repeats orders and settlements according to the rules while people are asleep.
Like a drowning person grabbing a branch, I read and reread the manuals.

https://www.gogojungle.co.jp/re/EovF5gdbktyC3bU" rel="nofollow">Prosperitas Nexus 30https://www.gogojungle.co.jp/re/img/NsoEwjVKQjDc9gy" alt="30の戦略知能が、一つの判断へ。 | GogoJungle"/>" target="_blank">

How does EA work? Specifically, what does it do?

EA is a trading program that mainly runs onMetaTrader 4 / 5 (MT4/MT5) (other platforms such as cTrader exist).
Written in the language of MQL4 / MQL5, when attached to a chart,
it executes trades based on predefined conditions.

In short, it comprises the following elements.

Entry conditions: when to buy/sell
(e.g., moving average cross, time window, breakout)
Exit conditions: take-profit and stop-loss ranges, trailing stop
Money management: lot size per trade, number of concurrent positions, risk % of account balance
Filters: skip when spread is wide, pause before major announcements,
and the flow of moves proceeds as follows.

1. Read market data every tick (or bar)
2. If conditions are met, place an order via OrderSend-equivalent process
3. While holding, monitor exit rules and close when conditions are met
4. Log it and wait for the next signal under the same rules

EA is not a crystal ball that predicts the future.
It is a tool to bring the professional practice of “rule-based, repeatable, emotion-free trading” into a personal account size.
If the rules are poor, losses accumulate automatically.
Therefore, EA is not a guaranteed win; it is safe to understand as moving judgment into the system.

Can EA become a side business?

Reality seen through research

In short,
“passive income by leaving it alone” is unlikely, but it can be a side business that designs how it is operated
This is the sense on the ground as of 2025–2026.

Reasons it can happen

Less time wasted: you don’t stay glued to charts; you rely on rules, logs, and stopping lines

Reproducibility: same conditions executed repeatedly
(less affected by day-to-day mood)

Second-tier income: selling or renting self-made EAs on MQL5 Market etc. is officially supported
(Developer guide: https://www.mql5.com/en/articles/385 )

Realistic aspirational target: in operational articles and practitioner summaries, with proper risk management

Around 3–5% monthly is often described as a realistic target
(sales that promise high yields come with fraud risk warnings)

Reasons it is unlikely / risky

Principal loss risk remains: whether as side business or investment, you should separate living funds
Past beautiful backtests do not guarantee the future
(curve fitting, ignoring spreads in validation, etc.)

Trading costs: even with high win rate, profits are eroded by number of trades × spread
(high-frequency scalping tends to be less favorable)

Market environment changes: an EA built in a trending market can collapse in a ranging market

Unregistered operators and withdrawal troubles: Consumer Affairs Agency and Financial Services Bureau repeatedly warn about trading with unregistered FX brokers (withdrawal issues, etc.)
(https://www.caa.go.jp/policies/policy/consumer_policy/caution/caution_034 )

In other words, EA side jobs are not about the fantasy of earning easily, but about “monitoring and rules as a side job.”
If you design to avoid cutting into your main job, set maximum monthly loss and stop when exceeded
If you cannot design up to this level, turning it into a side business tends to fail early.

EA and side business: two meanings organized

It is important not to confuse “EA and side business.”
There are at least two meanings.

① Side business in investment (operating a trading account with EA)
Run auto-trading alongside your main job.
Source of income istrading profits
What you need isstable VPS/PC,broker selection,log keeping,compliance with stop rules.
Not chart artistry, butoperational managementbecomes the work.

② Tool-based side business (selling or renting self-made EA) 
Source of income issoftware sales.
MQL5 Market has frameworks for product registration, verification, and support.
Even good logic may not sell; you must handle explanations, version control, complaints, and disclosure of backtests.

For most people, it is practical to grasp the feel with approach ①, and if it doesn’t fit, move to ② or self-madein a two-step approach
If you expect “passive income” from approach ① alone, market changes will break your heart first.

Approach ② is a test of your ability to “create, fix, and explain” even if you’re not good at investing.

Characteristics of Commercial EAs: “What you get when you can use it immediately”

Features of commercial EA (MQL5 Market, broker distributions, personal sales sites, etc.) are as follows.

Can try immediately: works with installation and parameter setup
Reviews and rankings give a feel for it
(though live results may diverge)
Some products have support.
Source code often not disclosed
More people using the same EA can thin the edge and cause broker environment to show differences
Price range is wide (free to expensive); expensive doesn’t guarantee safety
Market size (reference): MetaQuotes official introduction (2018) shows about 14,000 products in total, among them robots
(EA) about 4,800 ( https://www.mql5.com/en/articles/5194 )
The 2020s saw further increases in product numbers; even official stores have thousands of EAs
When adding non-public, self-made, and other platforms, the realistic count istens of thousands. It’s not a world where you can simply pick one.

The matching difficulty between commercial EAs and “the EA that fits you”

“Highly rated EA” does not equal “EA that fits your account.”
Just like clothing size,deciding without trying is dangerous.

Common obstacles include the following.

A beautiful backtest curve may indicate overfitting to the past.Results change with spread, slippage, and order execution quality
If account balance and leverage don’t match, you’ll get unexpected drawdowns
Gaps between demo and live trading (reviews vs. actual operation)

Compatibility with your lifestyle rhythms, acceptable losses, and currency pairs.
Many who fail with commercial EAs end up not from buying but from
deploying to live without verification.
If it doesn’t fit, instead of “keep searching,” many people move toward
narrowing down to a personal solution (self-made or customized).

The option of making your own EA

For MetaTrader, you can write EAs in MQL4 / MQL5.
The flow is simple.

1. Verbalize the conditions (e.g., only in Tokyo time, stop loss of X pips)
2. Translate into code for orders and position management
3. Backtest with a strategy tester
4. Try with a demo, then small live lots

Benefits includethe ability to add one more step: stop-loss width, time window, news pauses, and drawdown during losing streaks.
Many complaints that cannot be fixed with commercial products are resolved here.

The drawbacks areverification responsibility, bug risk, and the time to learn and operate, which all return to you.
Starting from scratch is a learning costgreater than another side job.

We are in an era where you can “make it” without coding

In the 2020s,natural language explanations of logic, with AI providing draft codehas become common.
Tell ChatGPT etc. “for USD/JPY, 1-hour chart, short-term moving average crossing the long-term moving average to buy, 30 pips stop, 20 pips take profit,” and generate a rough MQL blueprint; a human fixes compile errors, tests with a tester, and so onhybridis practical.

What matters is not “who wrote it,” butwhether it breaks.
Common accidents where AI integration goes wrong include the following.

Bugs that reference future data(repaint)
Errors in order logic(stop orders, overlapping orders)
The trap of optimization(parameters that only fit the past)
Completely leaving it to automation is dangerous, butthe burden of becoming a programmer has decreased, which is true. Writing in plain language like a specification about “how you want it to move” becomes the gateway to self-making.

Advanced self-made EA: “one step up in design”

Session splitting: switch logic between Tokyo, London, and NY
Volatility filter: skip when ATR is below a threshold
Automation of money management: halve lot size after N consecutive losses, stop when monthly loss limit is reached
News filter: prohibit new entries around indicator releases
Diversification across multiple currencies and multiple EAs: allocation to prevent one failure from ending the entire account

At this level, EA is not just “automatic entry,” butcoding one’s own trading constitution.

If you’re motivated, you can also consider AI-equipped EA

Going one step further,EA that uses AI (machine learning) for signals and filtersis technically possible.
Example: learn from past data to score the probability of entry; however, the difficulty increases dramatically.

Data quality and quantity, overfitting management
Real-time inference latency, VPS specifications
Accountability (can you explain why you bought?)
“If you’re motivated” to build it, it is not an exaggeration but a cost discussion.
Starting with rule-based EA and then adding AI as a predictive layer is often safer in practice.

Notes on self-made EA? If you skip these, you’ll get a “Bang” again

Self-made is free but responsibility returns 100% to you.

Over-optimization: do not pick parameters that only look good in the tester
Bugs: unintended consecutive orders, opposite-direction double positions
Validation that ignores costs: factor in spreads and slippage
Over-leverage: EA can be robust but you can still blow up with large lots
Brokers and regulations: residents of Japan should verify they are with registered brokers (as per Financial Services Agency unregistered broker lists)
Sales promising high yield: purchasing an EA can become a secondary harm

Those who have suffered a big loss are more prone to jump to another belief: “this time it’s automatic, so safe.”Another illusion.
Self-made is a tool to dispel that illusion, but it can also be a breeding ground for new overconfidence.

How to operate using EA: “practical workflow”

Operation isn’t “set EA and leave it,” but a realistic rhythm like the following.

“Initial run” 
At least 1–3 months in a demo with logs. In live, split funds from living expenses into a separate account, starting with small lots.

“Daily routine” 
Every day: check profit/loss, drawdown, order errors (no need to be glued to the screen; monitoring can be brief).
Weekly: limit rule changes to once a week, avoid tinkering too much.
Monthly: when hitting the maximum loss, stop by force and write down the cause.

“Environment”
If running 24 hours on a VPS, verify the provider, latency, and maintenance windows. If plugged directly into a PC, understand risks of power outages and line cuts.

“Record”
Make it possible to trace back why volatility widened, what indicators caused the loss, etc.

[Summary this time]

EA is not a “way out,” but a restart of design

Experiencing a large loss in FX is embarrassing.

But what you can learn is that it wasn’t the market that was bad, but you were fighting with your emotions.

EA isn’t a magic that makes your emotions go away.

However,moving judgment into rules and protecting them with tests and stop lines is one of the best tools for a restart in practice.

To summarize, the following:

- EA can be a side job that automates entry, exit, and capital management, butwhether it’s a trading-side business ora sales-side business

- Commercial EAs are convenient, but matching to yourself is difficult; you cannot fix a black box
Self-made is freedom tied to responsibility.

- AI can be used for code drafting or advanced signals, but verification is done by humans
Operating is the rhythm of monitoring, stopping, and recording

- The next step may be less about chasing flashy EAs and more aboutwriting down acceptable losses and monthly stop ruleson paper.
From there, adding automated trading becomes a movement from fear to a sense of design.

Your“Bang”— a mechanism to never repeat it —
That is the essence of how to face EA.