Everytime someone loses money on a bad bet in a prediction market, it's an opportunity for them to learn something about ~~counter-party risk~~ adverse selection. You could easily make the argument that the more people are losing in prediction markets, the more learning is happening.

Every time a drug addict uses, they’re actually learning about biology and the sociology of addiction.

Counterparty risk would be more like the betting site going belly-up and not getting money out despite having a profitable trade. This is just regular risk.

Betting venues can make money on transaction volume rather than holding an edge against their customers on the actual bets.