Jordi Hays
Co-founder and host of TBPN, a daily technology talk show covering startup news, funding rounds and the businesses behind Silicon Valley.
“The prediction market's not that helpful if it looks one way for months and months, and then in the two hours before results are in, it flips completely. This is the trick.”
TBPN · Startups & Venture · November 2026
Context
Giving a mixed verdict on prediction markets after the midterms, Hays objects that a market is of limited use if it only swings at the last minute.
Transcript
Overall, they called almost everything right.
Yeah.
I thought that obviously Texas was a bit of a surprise. Like the prediction market's not that helpful if it looks one way for months and months and months. And then in the two hours before results are in, it flips completely. This is the trick, right? But, I mean, again, and it's hard to process that, you know, if something has a 30% chance of happening. That's
one in three. I'll take those odds any day. Anyways, 50% chance. So yeah, some
surprises, but overall, like they seem to be pretty effective. It was interesting. There was a number of people, both in tech and tech adjacent, that were basically taking the stance of the polls are wrong and the prediction markets are wrong. Like just every indicator that we have is wrong. And…
Caul Shi came in at 95% accuracy when the polls opened. So you can always look at a prediction market and it's like, oh, right before the mainstream media calls the election, the prediction markets usually have it pretty dialed right before then. But the question is, what's the predictive power a…
More from TBPN
