Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

There’s a standard solution for this problem, namely numeric probabilities and prediction markets. If an internal prediction market had given an 80% chance of this happening, that’s meaningful in a way that individual ambiguous naysayings are not.


I'm aware of the existence of prediction markets, but I'm not aware of much evidence that they are actually good oracles, especially for events that will happen in the medium to long term. Do you happen to have any evidence you'd like to share about this?

Practically speaking, I don't think this can really work, because I don't think employees will necessarily feel good about betting on things like this. You also have some pretty severe moral hazards when you have employees who stand to gain financially if things go badly for the company. If I take a significant short position on NYC accepting an Amazon headquarters, I now have an incentive to leak negative things about the employer or otherwise sabotage the deal in order to get my contracts to close in my favor.


Are there any examples of corporations running internal prediction markets? That sounds fascinating and I'd love to read about exactly how they made it work, because I can see about 100 problems with this being the "standard" solution.


Google ran one internally for some time maybe 10 years ago. (Maybe still does, for all I know.) IIRC it was for points, maybe transferable for massage credits or something, so no huge economic incentive. I have no idea if it was in any way effective, actionable, or even seen by executives with decision making power. Based on the quantity of discussion on the mailing list I suspect it was mostly used by a pretty small set of people, so I'd have some doubts about its predictive power.


How would you get enough liquidity in the prediction market? Moreover, how do you prevent it from being entirely driven by insider trading?


A prediction market is just that, a market of folks with diverse interests on all sides of a given deal. That is not helpful to the leader of a company, whose interest is one-sided.


Then you're only as good as your source data, which unsurprisingly can go back to expert appraisals, i.e. "this won't work".




Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: