[lbrito]

Until the day they don't.

This kind of thing is inherently un-provable. You can have a method that works stupidly well for decades, and then it stops working. Building a model that provenly (in the mathematical sense) works with X% efficiency would be akin to building a free energy device.

[nightski]

It's very likely you put your life on the line every day with things that can't be proven. That is engineering after all.

[logifail]

> your life on the line

Your life at 0.00000x% risk is one thing. I'm fairly relaxed about that.

Your capital at - say - 50% risk of total loss is another. I don't play that game.