I lived off football betting for 5 years. Then I built this.

In 2019 I was 24, broke, and stuck in a job I hated. I had been quietly tracking football matches in spreadsheets since university — not for fun, but because I noticed patterns the bookmakers consistently mispriced. Closing odds drifting against the public side. Mid-table teams over-rated at home. The same dumb premiums on "big name" clubs week after week.

I had no money to risk, so I treated it like a research project. Six months in, my bankroll had grown from 80 thousand HUF to nearly 600 thousand. I quit the job two months later.

For five years, this was my full-time job. Wake up, pull match data, build models, scan markets, place bets, log everything. Every single result — win or loss — went into the same spreadsheet. No cherry-picking. No "this one doesn't count." Just data.

Some months were brutal. I had a stretch in late 2022 where I lost 14% of my bankroll over six weeks. I almost quit. But the math was right — the variance was just doing what variance does. Three months later I was up 38%.

I learned the most important lesson in this game: the edge is real, but the discipline to ride it out is what 99% of people don't have.

By 2024 my models had become too big for spreadsheets. I was processing form data, BTTS rates, xG splits, head-to-head context across 8 leagues, every single day. Manually. It took 4 hours every morning.

I started writing code to automate it. The code got better than my manual analysis within months — because it didn't miss things, didn't get tired, didn't skip a fixture because I was hungover.

Once I had the system running, I noticed something: my friends kept asking me which match I was on. Strangers in the betting groups I lurked in were paying 60-120k HUF a month for picks that were objectively worse than what my code spat out for free. The "tipster" market was a circus of confidence men and zero accountability.

So I made a choice: package the system, publish every single result publicly, charge a fair price, and let the data speak for itself. That is Novah Labs.

Every pick comes with a confidence rating, the model's reasoning, the published odds, and — within 24 hours of kickoff — the closing market odds so you can verify the value (CLV) yourself.

Every result lands in the public Track Record. WINs and LOSSes both. No edits, no deletions. The model had a 2/7 day in October. You can find it in the log. The model also had a 14-game winning streak in March. That's in the log too.

I'm not selling you the dream of getting rich on football. I'm selling you a piece of a system I built for myself, that I still use every day, that has fed me for five years. That's the whole pitch.

If you can't afford the monthly subscription, don't take it. Try the 3-day pass. If the model doesn't convince you in 72 hours, walk away — no hard feelings.

And if you do subscribe and the model goes negative for 30 days? The next month is on me. That's the deal. I trust the math enough to put my own revenue behind it.

Welcome to Novah Labs.