Hacker News new | ask | show | jobs
by somenameforme 18 days ago
I'm not really sure what you're trying to say. Perhaps I don't know what you mean by "make a difference." But providing accurate odds on a wide array of topics isn't yelling fire in a crowded theater and the results of the election did reflect the polymarket odds.
1 comments
freejazz 18 days ago
But you already conceded the odds were not accurate.

> But providing accurate odds on a wide array of topics isn't yelling fire in a crowded theater and the results of the election did reflect the polymarket odds.

I didn't say they were. You are making the broad meaningless distinction of implying that the "information" provided by "prediction markets" has any value. If it does not have value, what is the point of having the markets?

>Perhaps I don't know what you mean by "make a difference."

If the availability of the information makes no difference societally, then what is the value of the information? I'm not sure if you are being so obtuse on purpose, or what. But you keep saying it provides information... so what? As I pointed out... white noise is information.

> and the results of the election did reflect the polymarket odds

No they did not. Trump did not win by twice as much and certainly did not in the popular vote.

link
somenameforme 17 days ago
Oh boy. You don't even know how prediction markets or even odds work, at all, yet are abnormally strongly opinionated on the topic, and just assume you know everything without bothering to even check once. Of course. Prediction markets are binary yes/no. When Polymarket prices Trump as a 2:1 favorite, that does not mean he's expected to win 66% of the vote, but that if you ran the election 3 times, he'd be expected to win twice.

Polymarket's accuracy (as defined by an event corresponding with the majority position) is 90% a month out from an event, and 94% 4 hours out from an event. I suspect a weighted analysis (e.g. a 90% event should happen 9 of 10 times, not 100%) would show even better results. That is dramatically higher than you get from things like pollsters, and it's all freely available to the public.

link