My Quant Journey – From 2021 University Trading to Systematic Risk Management

So let me tell you something interesting about myself... I've been trading since 2021. My first year of university. I had my first allowance, some saved up pocket money, and the kind of confidence you only get when you don't know anything. That was five years ago. I'm graduated now (December 2025), and I'm still at it. Still losing sometimes. Still learning. That's the journey.

2021 To Now. Years Of Losing

My final year of university wasn't my first year trading. It was actually year four or five. At that point I had already experienced enough to know two things: I could make money sometimes, and I could lose even more of it consistently. The winning was sporadic. The losing was consistent. That's the worst kind of relationship with trading. The occasional win keeps you hooked just long enough to lose more.

Some years I made profit. Some years I broke even. Most years I lost. And every year I told myself next year will be different. It wasn't. It was the same mistakes, different markets. The same overconfidence, different indicators. The same disasters, different excuses.

I wasn't a trader. I was a gambler with extra steps. And I didn't even know it. That's the worst part. You don't realize you're losing until you've already lost so much you have to make a choice. Stop completely, or figure out what you're actually doing wrong.

What I Learned In University

Here's where my actuarial background comes in. In university, I learned statistics. Probability theory. Stochastic processes. Time series analysis. Differential equations. All these fancy concepts that sound great in textbooks. And I tried to use them. Of course I did. Because I'm that guy. If I learn something, I want to apply it immediately. Sometimes that works. Sometimes it doesn't.

ARIMA and Time Series Forecasting

I tried forecasting price movements using autoregressive integrated moving averages. Looked amazing on historical data. Completely useless in live markets. The assumptions behind ARIMA don't hold in financial markets. Markets aren't Gaussian. They're not even close. The stationarity requirement? Forget about it. Markets are non-stationary by nature. They evolve. They adapt. They break your models. And they laugh at your pretty mathematical assumptions.

GARCH for Volatility Modelling

GARCH models should capture volatility clustering. And it does capture some of it. But capturing volatility isn't the same as predicting direction. I could know volatility was high and still not know which way the price would go. Useful information? Yes. Tradeable edge? No. It's like knowing it's going to rain but not knowing if you should bet on the pool filling up or drying out.

Copula Theory for Dependency

In class I asked my lecturer to go over it in detail, people decided to move on, so i took the time to learn it. I tried using copulas to model the dependency between different currency pairs. Gaussian copula, t-copula, Clayton copula. The theory is decent. The implementation is a nightmare. Correlation between assets changes over time, especially in crisis. The copula assumes a static relationship. Markets don't do static. They do chaos. And chaos doesn't fit in your copula. It's like trying to use a map from last year to navigate today's city. The roads have changed. The traffic has changed. Your map is useless.

Jacobian Multipliers and Continuity

In university, I learned about Jacobian multipliers and continuity conditions for differential equations. The math is beautiful. The application to trading? Limited at best. These concepts matter for theoretical finance, for pricing derivatives, for understanding the mathematical foundations. But for actually making money in the market? Not directly. It's background knowledge. It's context. It's not a signal. It's like knowing how an engine works but not knowing how to drive.

The Brutal Truth About Academic Statistics

Most of what I learned in statistics class doesn't apply directly to trading. Not because the concepts are wrong. But because markets are different from the textbook problems. The assumptions don't hold. The distributions aren't normal. The relationships aren't stable. Academic statistics prepares you to think about risk. It doesn't prepare you to trade it. Those are different skills. It's like being taught to swim in a pool and then dropped in the ocean.

What Actually Makes Sense

But not everything from my degree was useless. Some things actually do translate. Risk management concepts. Probability of ruin calculations. Expected value thinking. Kelly Criterion for position sizing. These aren't just academic exercises. They're the foundation of systematic trading.

Monte Carlo simulation. Yeah, that's the one that actually stuck. Not for predicting the future, but for understanding what could happen. Running thousands of simulations to see the distribution of outcomes. Not to know exactly what will happen, but to know what might happen. That's the whole point.

Regime detection. Markov chains for state transitions. These are stochastic process concepts that actually work. Not perfectly. Not predictively. But as a way to think about market states. To categorize what's happening now rather than guessing what happens next.

The difference between useful statistics and useless statistics. Useful statistics help you understand risk and make decisions under uncertainty. The aim of modelling isn't to eliminate uncertainty, that's impossible. It's to make informed decisions in the face of uncertainty. Useless statistics try to predict the unpredictable. One makes you a better risk manager. The other makes you a dreamer with a spreadsheet. I've been both. I prefer the first one.

My Honours Thesis

My honours thesis was titled "Teaching Old Models New Tricks: Comparative Analysis of Classical and Machine Learning Models for FX Forward Contract Evaluation Through Future Spot Rate Prediction Across Developed and Emerging Markets."

Long title. Complex topic. What did I learn from it? Machine learning models don't automatically beat classical models. The classical models have economic theory behind them. ML models just optimize for the training data. Sometimes classical wins. Sometimes ML wins. It's not about which is better. It's about understanding when each approach makes sense.

The research taught me something important. ML is a tool. Not a magic wand. You need domain expertise to use it properly. You need to understand what the model is doing and why. Otherwise you're just throwing data at an algorithm and hoping something sticks. That's not research. That's guessing. That's like praying to the RNG gods and hoping they answer.

ML Is A Lie. Especially If You're Not An Expert

Let me be direct about this. I tried ML. I tried it hard. And it's a lie. Not completely. There's real stuff there. But the narrative around ML in trading is completely overblown.

Overfitting Is The Default

When you have thousands of features and limited data, your model will find patterns that don't exist. Backtests look incredible. Forward tests fail. The reason is simple. The model memorised the past instead of learning the future. And the future isn't the past. It never is. It's like memorizing the answers to a test you already took and thinking you'll ace the next one. You won't. The questions are different.

Data Leakage Is Everywhere

Look backwards in your training data, accidentally include future information, build a model that seems to predict perfectly. Then deploy it and lose money. It's that easy to fool yourself. I did it multiple times. Each time I thought I'd found something. Each time I was wrong. It's like looking at a photo of someone after you already know what happened to them and thinking you "predicted" it. You didn't. You cheated. And the market doesn't accept cheaters.

Feature Engineering Is Guesswork

You don't know what features matter. You try hundreds, keep the ones that work in backtests, and call it research. That's not science. That's selection bias with extra steps. The features that worked historically might work going forward. Or they might not. You have no way to know. It's like picking lottery numbers based on which ones came up last week. Technically a strategy. Technically useless.

You Need Real Expertise

You need domain expertise to do ML properly. And I don't mean "know some trading." I mean deep expertise. Years of experience understanding how markets behave, what features actually capture edge, how to avoid overfitting, how to validate properly. If you're not that person, you're not doing ML. You're doing high-tech gambling with extra steps. I wasn't that person. Most people who try ML in trading aren't that person either.

The ML enthusiasts online don't show you their backtests that failed. They don't show you the models that worked in simulation and collapsed in live trading. They show you the one that worked and say "look, it works!" That's survivorship bias. Possibly even selection bias. That's not reality. That's like showing someone your one winning trade and pretending you're a profitable trader. We both know that's not how it works.

So I stopped trying to predict the market with ML. Not because the idea is bad. But because I'm not good enough to do it properly. And neither are most people who try. That's an honest assessment. It took me years to make it.

Why I Stopped Discretionary Trading

Here's what I want to tell you. If you're still trading discretionary, if you're still looking at charts and making decisions based on how you feel, stop. I'm not saying that to be harsh. I'm saying it because I was you. And I lost a lot of money doing it. A lot.

Discretionary trading sounds appealing. It feels like you have control. You look at a chart, you see a pattern, you make a call. You feel smart. You feel in control. But that feeling is a trap. The market doesn't care how smart you feel. The market will take your money just the same whether you feel like a genius or like an idiot.

The problem with discretionary trading is consistency. You might make good decisions sometimes. But you also make bad decisions sometimes. And you can't tell the difference between the two until the trade is closed. By then, it's too late. You're either celebrating a win or licking your wounds. There's no system. There's no process. There's just you, guessing, hoping, praying.

And emotions. Don't get me started on emotions. Every discretionary trader says they can control their emotions. They can't. Nobody can. Not really. The moment you see red on your screen, your brain does things you didn't plan for. It tells you to exit early. It tells you to double down. It tells you to revenge trade. You think you're different. You're not. I'm not. Nobody is.

That's why I moved to systematic trading. Not because I'm smarter. But because I know I'm not smart enough to beat the market with feelings and intuition. I know my brain will lie to me. I know I'll convince myself of things that aren't true. So I built a system that doesn't care what I think. The system doesn't have emotions. The system doesn't second-guess itself. The system follows rules. Unlike me, apparently. Also invest in your own server it helps.

Systematic Trading With Hypothesis Testing

This is the approach I take now. It's not about predicting the market. It's not about being right. It's about testing hypotheses and letting the data decide. Wild concept, I know.

Every strategy I build starts with a hypothesis. Not a feeling. Not a pattern I think I see. A hypothesis. Something I can test. Something with clear rules. Something that can be proven wrong. If you can't prove it wrong, it's not a hypothesis. It's just a hope. And hopes don't make money. Trust me, I've tried.

Hypothesis: Markets that are ranging will mean revert within a certain time frame. Test: Build a system that trades mean reversion in ranging markets. Run it on historical data. See if it works. If it doesn't, the hypothesis is wrong. Move on. Hypothesis: Trend following works in strong trends. Test: Build a system that trades with the trend when certain conditions are met. Run it. See if it makes money. If it doesn't, the hypothesis is wrong.

This is how proper research works. You form a hypothesis. You test it. You accept the results. If the data says your hypothesis is wrong, you don't double down. You don't tweak the parameters until it works. You accept it and move on. That's the difference between trading and gambling. Gamblers double down when they're wrong. Researchers move on when their hypothesis fails. I'm still learning to be a researcher, not a gambler.

The system I run now is built on this approach. It has regime detection. It uses Monte Carlo simulation for stop placement. It has Markov chain filtering for regime transitions. It uses Kelly Criterion for position sizing. It has hard risk limits. 1% max risk per trade. 3% daily limit.

Does it work all the time? No. Nothing works all the time. But it works consistently. It loses less when it's wrong. It survives through drawdowns. And most importantly, it's testable. I can run the same tests again. I can validate the results. I can improve it over time.

That's the point. Systematic trading with hypothesis testing isn't sexy. It's not exciting. You won't feel like a genius when you make a trade. But you also won't feel like an idiot when you lose one. Well, you might still feel like an idiot sometimes. But at least it's a systematic idiot with a process. You'll just follow the system. You'll collect the data. You'll test the hypotheses. And you'll keep improving. That's the process.

Life Philosophy. Why I Keep Going

Here's the thing. Trading without passion is exhausting. When you're only in it for the money, every loss feels like a failure. But when you're in it for the learning too, every loss is just data. Every failure is feedback. Every drawdown is information. The emotional weight changes when you're doing research, not just chasing profit.

And honestly. That's the only way I've been able to stick with it. If this was just about money, I would have quit years ago. The losses would have been too discouraging. But because it's also about building something interesting, about understanding markets, about applying what I learned in school in a way that actually works, there's a different kind of motivation there.

This might be one of my longest commitments. Right up there with Overlord. And that's saying something. That show has been my thing for years. But this quant journey? It's become more than just trying to make money from trading. It's become a genuine research interest. A technical challenge. Something I'm actually curious about beyond just the money. And that's saying a lot coming from a guy who really likes money.

Still Learning. Long Haul Ahead

I'm still in the early stages. Still building the system. Still learning what works and what doesn't. Still making mistakes. Still "losing money" sometimes. But now I lose less. And now I understand why. That's progress. That's what this journey is about.

The code is private now. Not because it's special. Just because I haven't made it public yet. But the system lives in the quant branch running on my server somewhere No more loadshedding. It's not finished. It might never be. But it's honest. It's systematic. It's mine.

What If I Started A Quant Research Firm

Sometimes I think about this. Actually starting a quant research firm. Not a hedge fund. Not a prop trading desk. A research firm. Focused on systematic approaches to trading. Focused on the research itself, not just the trading.

What would that mean. For me. It would mean taking everything I've learned. The failures, the lessons, the system I've built. It would mean committing to this path not as a side project, but as a career. As an identity. As something I actually do for a living, not just something I do in my spare time while answering emails for my day job. Not that I have a day job yet. Fingers crossed.

It would mean accepting that I'm not just a trader. I'm a researcher. A systematic thinker. Someone who builds things that try to solve a hard problem. The problem isn't just making money. The problem is understanding how markets work, how risk works, how to survive in an environment that's designed to take your money.

What would it mean to me. Everything, honestly. I've been building toward something like this without naming it. The trading system, the research, the learning, the failures. All of it is preparation for something bigger. Not just profit. But a contribution. A real contribution to how people think about systematic trading, possibly investing and risk management.

There are a lot of people out there selling trading courses. Selling signals. Selling indicators. Selling the dream of quick money. I don't want to be that. I gatekeep and I want to be the person who actually does the research. Who builds something honest. Who shares what's real, not what's profitable to sell. There's no money in honesty, but there's definitely money in gatekeeping .

Maybe a quant research firm is the way to do that. Or maybe it's just a fancy way to describe what I'm already doing. Either way, it's on my mind. It's one of the things that keeps me going. The idea that this could be more than just trading. This could be a life's work.

I'm not there yet. I don't have clients, I don't have capital, I don't have a team. What I have is a system that's getting better, a way of thinking that's maturing, and a willingness to keep learning. Maybe that's enough to start. Maybe that's where the journey leads next.

The Brutal Truth

Let me end with the brutal truth.

Algorithmic trading won't make me invincible. It won't make me rich yet. It won't make the market stop doing what it does.

But it will make me consistent. It will make me survive longer. And most importantly, it will make me lose less when I'm wrong. That's the actuarial promise. Not you won't lose. But you'll lose less than you would have otherwise. And over time. Over time, that's everything.

I'm in it for the long haul. Just like everything else I've ever cared about. Including that time I binge watched 8 seasons of Naruto... But that's another story.

PS: If I am making money from this I wont tell you. If am earning good money you wont know... But you can trust me to smile and gatekeep.