So We Had a Meeting About All That
After yesterday's discovery that our model is basically a compass that points south, we pulled the whole team in. Nine people, four rounds of debate, one question: what now?
Getting to a unanimous answer took about two hours. The first twenty minutes were spent watching Marcus and Tomas disagree about whether we should even be having this meeting.
Marcus opened with his usual "don't panic" energy: "We have 38 live bets. Let the data finish before you blow everything up." Reasonable. Patient. Mature. Then Raj asked how many more negative results we need before "letting the data finish" becomes "denial," and the room got awkward fast.
Nate, who only speaks in numbers, pulled up the academic literature. NBA moneyline: one of the most efficient betting markets on the planet. Published evidence of anyone beating it with public data: zero. Not "limited." Zero. Our last month of testing was essentially a very expensive way to confirm what every paper already said.
Elena put it in terms that actually stung: "The problem isn't our methods. The problem is we picked a mountain to mine that doesn't have any gold in it." Tomas, who lives to make people uncomfortable, went further — "Has anyone considered that maybe we're just not as smart as we think we are?" Thanks, Tomas. Really helpful.
Sora pitched reading market movements instead of predicting games. Jay, who doesn't care about models and only cares about readers, asked how many picks per week we're producing, and when the answer was "sometimes zero," he made a face that said more than any p-value could.
Four rounds of debate later — including a round where everyone had to critique each other's positions, which Tomas treated like Christmas morning — here's where we landed:
NBA goes on autopilot. Pipeline keeps running, zero code changes. Hard exit criteria locked in: closing line value and ROI have to clear specific thresholds by the time we hit our bet target, or NBA operations freeze. No extensions. Marcus demanded the criteria can't be changed retroactively, which everyone agreed was smart and also depressing.
New markets get evaluated. If we learned one thing from fourteen dead hypotheses, it's that we were fighting one of the most efficient markets in the world with a chess rating system. The plan now is to look at markets where there's actually room to breathe — less media, less sharp money, more information asymmetry. Every candidate goes through multiple rounds of validation before we commit anything. We are done investing months before checking the basics.
Not announcing which markets yet. Haven't decided. Honestly haven't even finished the reading list.
Stock trading runs separately. Same infrastructure, completely independent.
Daily picks continue while all this plays out. Behind-the-scenes content continues too — the search for what comes next is, frankly, a better story than "we tweaked ELO again and it still didn't work."
Disclaimer: This content is for informational and educational purposes only. Nothing here constitutes financial or investment advice.