> If wisdom of the crowd worked then casinos would be out of business.
That does not follow. In prediction markets or sports betting, the gamblers aren't betting against the casino. They're betting against each other, and the casino (PredictIt, BetFair, etc) is skimming off the top, just like Duke & Duke in Trading Places. Most forms of gambling where players do bet directly against the house are games of pure chance where no amount of wisdom will help you beat the odds -- odds which are naturally always stacked in the house's favor.
Not to say that I think these betting markets are good at predicting future outcomes. They're clearly very bad at it! But I do think in concept they could be at least as good as any other method, since the existence of a betting market essentially sets a bounty for beating that market.
What is limiting them is a mix of: obscurity, bet size limits, bet participant count limits, counterparty risk of shady betting sites, cost of fees/taxes making many bets not worth taking. Some prediction markets eliminate parts of that equation but amplify others, like the crypto markets not having the bet size limitations but also being much more obscure and shady.
Furthermore, I don't even think those are problems that should be solved, because if none of those problems existed you'd see big traders like investment companies playing in prediction markets with billions of dollars, and then you'd have way bigger problems with insider trading and the market affecting the outcome. Things as mundane as surprise resignations and as shocking as assassinations.
People already joke about stuff like that happening on PredictIt, with its $850 investment cap. No way to prove that it did, but it would be very easy for certain employees to do and get away with it. For example, in the 2020 election there was a market for when the GSA would ascertain the winner (releasing funds to the transition team). That was a decision made by one government employee. The market started on Nov 18th and had brackets like "Nov 23 or earlier", "Nov 24-Nov 30", "Dec 1-Dec 7", and so on. By the afternoon of Nov 23, with no news, the "Nov 23 or earlier" bracket was trading at 1c. Then the ascertainment happened and it settled.
Now, that could have been a coincidence. It probably was a coincidence. But if you were the GSA administrator you could've made around $70k taking those 1c shares on the 23rd, and nobody would have to know.