It's an empirical process. The results themselves are what you use to construct an idea of probability and risk-reward. Nothing is ever certain which is why position size rules are a must. If an extremely specific pattern continues to show up on one specific stock on days with specific conditions (huge multiples of daily volume, for one), it's likely a significant observation representing an edge rather than noise. Unless you think it's totally random for hundreds of independent traders to exploit a specific pattern on a specific stock on multiple repeated dates and make money repeatedly. It's not a poker/blackjack or a casino game where the odds are fixed and known.

Ask yourself, if you flipped heads on what you thought was a fair coin 500 straight times, did you REALLY just observe an ultra rare event? Or is it more likely another phenomenon at work (like a rigged coin)? Whether you can fully explain it or not doesn't matter.

Why do you care so much what I do with my money? Or how a firm chooses to allocate its money? This wasn't client money or institutional money, it's the money of a few guys (partnership type of structure) who used to be or still are daytraders themselves.

(one last edit: if you're too thick to see it, I deliberately tried to showcase my overconfidence to show how things can go wrong easily. guy makes money and wants to make more, guy wins money and thinks he's a champ -- it's called the human condition. I deviated from normal execution rules and position size rules and paid the price)