I struggle with price models. It's one thing to model the movement of the moon around the Earth. You watch the movement of bodies, day after day, and build up a dynamical theory that describes the trajectory of the moon. When your model holds up, cycle after cycle, you feel satisfied. Your mathematical model might not represent "reality", but at least it gives you consistent predictive power. The moon won't adjust its behavior to evade your model.

But a public price model is totally different. If you discover a pattern in the price data over time and publish it, wouldn't that information make its way into trading decisions, and thus eventually break down? If you discover market inefficiencies and publish them, the market will eventually become efficient and your model will no longer have an edge.

Maybe I'm off here.