Hacker News Reader

Comment by BurningFrog | Hacker News Reader

I haven't seen Hanson address the "destructive insider trading" scenario where I buy a prediction that California air quality will be bad, and then spend the summer starting wildfires in California forests.

Today, people have no incentive to do that other than "wanting to watch the world burn". But if you can get rich burning forests, more forests will burn.

I don't doubt Hanson has thought about this. I just haven't seen it addressed.

View on HN

My guess is that if a particular prediction were really so easy (low cost and risk) to manipulate then you wouldn't expect much trade volume. As an easier example, you wouldn't expect many people to bet money that no one will tweet a cartoon of a cat wearing a pink hat today, because it would be obvious to everyone that it would be trivial for an insider to manipulate this outcome.

As for your specific example, starting wildfires is an extremely serious crime that I expect the authorities would spend significant resources investigating, so I don't know how bettors would estimate the ease of manipulation for that market. Even if the prediction market were implemented in a perfectly anonymous system where you're guaranteed to be able to collect (and spend) your money without authorities knowing you were the one who profited from that prediction market, it seems like you have a pretty high risk of getting caught simply for the act of arson itself (not to mention the "risk" of firefighters controlling the fires that you start).

I'm also not sure how many people there are out there who refrain from starting wildfires only because there is no direct financial incentive, and not for other reasons (like not wanting to destroy massive amounts of ecosystem and property and potentially kill many people).

Air quality is something that is hard to predict and really quite important for many reasons.

I think you underestimate organized crime. Starting wildfires is easy and skillful criminals won't get caught. The people that pay them take on an even slimmer risk.

I'm not estimating organized crime at all, because I'm very ignorant of it. I'm only saying that I would expect the potential bettors in that prediction market to make some estimation of how difficult manipulating the market would be, and not many people would bet on a market they deemed to be easily manipulable. Another mechanism that could exist is that the organizer of the prediction market could itself judge whether a particular market was manipulated and have policies for what to do in such a situation.

But in reality, that just doesn't and didn't historically happen in stock markets. Buyers didn't accurately estimate the likelihood of manipulation. The government had to step in.

I mean, someone can buy stocks short and then burn down a company's HQ/factory today.

This isn't a reason to not allow equities markets.

Reply on news.ycombinator.com