You acknowledge a non-trivial chance of losing a non-trivial amount of money. The confusion is that I thought your previous statement that a "smart bias-educated person can beat the prediction markets fairly easily" would preclude this.

There are arbitrage opportunities, but they're not what I'm thinking of.

An analogy: knowing about biases and how to play optimally is important to play poker at any high level; but that still doesn't mean you're going to win every hand. I might correctly call an election for Obama, but that's not going to help me as a trader if he abruptly dies of a heart-attack or Sarah Palin stages a coup with a crack unit of Alaskan hunters - I'll still lose my money. I don't see any contradiction here.