Comment by bit-anarchist

The existence and severity of the harm, as argued by OP, relies on incentives. But incentives don't exist on a vacuum, nor are they individually absolute over human behavior. Simply stating that something has a bad incentive and that'll result in evil, on its own, isn't much better than assuming spherical cows.

The reason I nitpicked on the arsonist example is to elaborate on that: the arsonist does indeed have an incentive to put fire onto someone's house due to the bet, but the bet itself increases the probability of the fire being prevented, the culprit being caught, etc., which are all disincentives that may negate the original misaligned incentive. When we escape from this particular scenario onto others, this pattern remains. If the disincentives didn't prevent said harm, there are only two causes:

- culprit is unreasonable beyond saving: this one is trivial - they are beyond saving. Banning gambling would simply cause them to commit crimes in a different way;

- the mechanisms behind the disincentives are malfunctioning or aren't enough: that warrants an investigation onto said mechanisms, not a ban on prediction markets.

Perhaps this has made it clearer that, at least from my perspective, these "irrelevant details" aren't irrelevant.