About me

Hi. I'm Yevhen — some people call me Eugene. I trade US stocks with programs I wrote myself. Every trade, every day, no guessing. It took me almost seven years of losing money to get here. This is the honest version of how that happened.

I'm not a financial advisor, and nothing here is financial advice. Everything I write about is my own experience, with my own money. Trading is risky — you can lose all of it. Past results tell you nothing about future results. Do your own research, and if you need advice, talk to someone licensed.

Where it started

I started at the end of 2019, just before the world shut down. I had a full-time job as a software developer, and one salary felt like standing on one leg. I wanted a second source of income. That's all — no dream of yachts. Nobody around me knew anything about markets, so I did what everyone does: I searched the internet.

The slot machine

The first thing I found was called binary options. Spoiler: it wasn't options at all. It was a bet on whether a price would be above or below a line — in five minutes, or one minute, or thirty seconds.

I know what that was now. It was a slot machine with a chart on it. You place the bet, you get the little hit of dopamine, then you lose and you want it back right now. So you click again. And again. Thirty seconds at a time. Nobody can predict thirty seconds, so the ending is written in advance.

The site had one more feature: if you won a few bets in a row, your payout on the next bet got smaller. At the time I thought that was a detail. Now I understand what it was.

Forex, and a brilliant idea

That took me about a year. Then I heard about forex. I still knew nothing about stocks — I typed "trading" into Google and Google gave me CFDs. Currencies. So that's what I traded.

And I had a brilliant idea. I'd read ten or fifteen morning news reports every day. If they all said buy the euro, I'd buy the euro, then sell when there was some money in it. I didn't know how to size a position. I didn't know what a stop loss was. I didn't know what a lot was.

The day it doubled, and the day it died

One day in 2020 I put my whole account into one trade. By the evening it had doubled. I remember thinking: that's it, I found it, I'll be a millionaire soon. So I withdrew half, to feel safe.

A few days later I did exactly the same thing. Read the news. Whole account. Buy. Then I sat there all day and watched the price walk slowly against me. In the evening the stop took me out and the account was gone. One day. My hands were shaking. I didn't understand what had happened — all the news said buy.

Half a year away, then the rabbit hole

I quit for about half a year. I decided the whole thing was rigged: the broker wanted my money and regular people had no business being there. Then the interest came back, and this time I found a video about something called technical analysis. I thought I'd watch two or three videos and understand what to do. That was the entrance to a hole I spent years in.

The midnight candle

Somewhere in there I found an Australian trader called Nial Fuller. I bought his course and studied it properly. It was inspiring, everything looked clean and simple, and I don't doubt it works for him. But his approach is built on daily charts, and in my time zone the new daily candle opens at midnight. Sitting at a chart at midnight, every night, is not a life.

After a few months I admitted it wasn't for me. So I had a thought: if you can make some number on daily bars, then inside the day you can make ten times that. Right? Obviously this guy just hadn't figured out day trading. And I would. Because I'm clever.

Almost six years in one market

That was the beginning of the longest part of the story. Years of intraday forex. Support and resistance. Hundreds of indicators. Brokers changed, courses changed, and thousands of hours of YouTube went into looking for anything worth automating.

And I only touched currencies — CFDs on currencies, nothing else. I believed you have to pick one niche and beat on it until it opens. Stubborn is right, by the way. But sometimes you also need to let yourself go around.

Doing it by hand, in the evening, after a full working day, didn't work. So I moved towards automating it.

The lone wolf

There's one more piece to this, and it's the least flattering one. I've always been a lone wolf. I wanted to work everything out myself — my own way, my own thing.

And when somebody who really knew the subject told me something, I couldn't take it in. It wasn't that I thought they were wrong. I could see the person knew what they were talking about. I just couldn't let it land.

That's probably what cost me the years. Not the market. Not the broker. Me, not listening to anyone.

Writing the robots myself

I have a background in software development, and that turned out to be the most useful thing I owned. Hiring somebody to code your ideas costs money, costs time, and you have to pour your whole head into another person. So I learned MQL5 and wrote the robots myself. And I tested. And tested.

By the sixth year I was deep in the sunk cost trap. Years gone, real money spent on courses, and I kept telling myself I had to finish what I started — because look how much I'd already put in. It was all useless. That's the honest summary.

The disloyal detour

Then I let myself do one small disloyal thing. The weight of those years was getting hard to carry, so I decided to go around it, just to see.

There was one stock strategy I'd read about. I knew nothing about stocks — only that my CFD broker had CFDs on them. I coded the rules and ran the test out of curiosity. Even with that broker's brutal costs, it came out around seven to nine percent a year across the four or five years of stock data the broker had. That's a backtest, not a forecast: short period, one broker's data, and I'd want much more than that today.

But the number wasn't what got me. What got me was how boring it was. It just kept working, quietly, after costs.

Stocks, properly

So I went looking for what people actually use to test stock strategies. I found the specialised software, rewrote that strategy properly in it, and it held up. A few hundred tested strategies later I knew two things: I wasn't going back to CFDs, and I wasn't going back to intraday.

And I finally learned the thing I'd had backwards for years. Markets are not interchangeable. I used to assume a real edge should work everywhere. It doesn't. What works on stocks isn't what works on currencies.

What I actually run now

Several systematic strategies on US stocks, end of day. No screen watching, no gut calls, no overrides.

The main group is mean reversion: when something gets stretched too far down inside an uptrend, it tends to snap back. That's the whole idea. There's also a breakout sleeve, for when markets just trend and the snapback never comes.

Every strategy gets built and validated in AmiBroker first, then ported to my own Python engine — and the port has to reproduce the AmiBroker trade list exactly (same trades, same dates, same prices) before it goes anywhere near real money. On the strategies running now that's 111 out of 111 trades matched, 195 out of 195, and 563 out of 563.

The data is end-of-day and survivorship-bias-free, with point-in-time index membership, so the backtest can't buy a company before it joined the index and can't quietly forget the ones that went bust. Execution is automated through Interactive Brokers: every position gets a profit target and a stop attached, positions are reconciled against the broker every day, and a monitor compares real fills against what the backtest expected, so drift shows up as a number instead of a feeling.

I paper traded first. Real money started on 13 July 2026.

The part most sites leave out

Now the honest weaknesses — I'd rather say them myself than have somebody find them.

• My live record is short. A few weeks of real money proves the machinery works. It does not prove the edge is real. Those are two different claims.
• The account is small.
• The strategies come from published books, not secret research.
• Mean reversion has an ugly shape: lots of small wins, and every so often a loss that hurts. That's the nature of it, not a bug I can remove.

I've linked a third-party verification service, so from around January 2027 there'll be a record I don't control. Until then you have my word — which is worth exactly what anybody's word is worth on the internet.

What this site is

I write and record what I test, what breaks, and what the numbers actually say — including all the times I was wrong, which is most of the archive.

I also post the trades I take in a private channel, as I take them. That's a paid thing I run, so factor that in when you weigh anything I say here. It's a journal of what I did with my own money. It isn't advice, and it isn't a service that tells anyone what to do.

Short version of me? A developer who lost money for almost seven years, got stubborn about the wrong market, and eventually built something boring that works.

If you want new posts as they go up, subscribe below — it's free.