Molly Whitethe betting platforms can't open markets on Trump's death for obvious reasons, but Kalshi just so happened to open up a "Trump out as President" market on Saturday
Unlike Polymarket, on Kalshi these markets do not resolve to "yes" if the person in question dies (they settle at the last traded price). I suspect many bettors don't read the fine print, though.
Polymarket opened their own "Trump out as president" market on September 1. unlike Kalshi, this resolves to "yes" if he dies
Donald Trump Jr. is an adviser to BOTH Kalshi and Polymarket, btw
(this is deeply weird for all of the many reasons having Don Jr. as an adviser is deeply weird, AND also because the companies are the two major competitors in the prediction markets space)
Luka Rubinjoni@molly0xfff "prediction markets", lol
AaronThey aren't perfect, but there is research to support their relative accuracy.
https://en.wikipedia.org/wiki/Prediction_market#Accuracy
It makes sense, when you consider the concept of "ensemble learning" from machine learning. Errors and biases tend to average out.
@rubinjoni @molly0xfff No argument with that
@rubinjoni @molly0xfff It's just that the gamblers' bad habits are being used for prediction
Avery 
@hosford42 @rubinjoni @molly0xfff the number of real-world tests seems pretty small. I made easy money on a prediction market in 2024 by betting it would be the hottest year ever, which was something easily predictable if you had been seen the news about sulphur dioxide
@avery @rubinjoni @molly0xfff Yeah, there's also the consideration that making it about gambling, as opposed to using a very specific structure for the question, doesn't seem to improve accuracy. There are other ways to get people to be accurate, in other words. Gambling is more of a motive to participate than an actual source of accuracy, I think.
My personal interest in it isn't in the gambling side of it so much as the adaptation of machine learning techniques to effectively coordinate human decision making. I think it's a fun idea.