The turning point wasn't a win.
It was a bug.

I'm Judah Ahiaba. I trade automated systems on gold, FX, and index CFDs, and I do the research and engineering behind them myself: the statistics, the backtest engines, the risk code, the 2 a.m. watchdog alerts.

I started where most people start: discretionary charts, conviction, and a results curve that depended heavily on which month you asked about. The habits that eventually mattered came from an uncomfortable observation: I couldn't tell you, with numbers, whether anything I was doing worked. So I started testing it. Properly: point-in-time data, real costs, every belief given a fair chance to fail.

Most of them failed. A daily pattern I'd trusted for years turned out to be a coin flip across 116,199 measured instances. Later, the first backtest I was genuinely proud of turned out to be reading the future: a subtle time-slicing bug that accounted for 70% of its edge. I could have shipped it and blamed the market when it bled. Instead, that correction became the foundation of everything here: the system you trade must be the system you validated, and the number you plan on must survive your own attempts to kill it.

Today the operation is three uncorrelated systems (one live on a funded account, two held back by their own written deployment gates) plus a standing audit practice that treats my code the way a risk desk would treat a stranger's. The style is deliberately unglamorous: small fixed risk, deflated planning numbers, published negative results, kill-switches in code.

The same pipeline is available to others: as build and audit work, and as an eight-week mentorship for traders making the same shift I did. Not because the shift is easy, but because it's testable, and that changes everything.