4) After a _lot_ of back and forth, we realized what was going on:
People wanted to be able to separate accounting.
We had subaccounts for that! But it took an extra 3 clicks to switch subaccounts, and that was annoying.
So we put a bar on trading pages to switch.
Conversation
6) Going into the election today, had Trump at 10% to win.
Prediction markets had him around 35%.
was not impressed with prediction markets:
Quote Tweet
1
1
6
8) It's unclear.
On the one hand, I think he had a bit of a point: the people trading on prediction markets was *not* representative.
Crypto Twitter is ~60% Trump supporters, but the US is ~45% Trump.
So there were more buyers than sellers of ftx.com/trade/TRUMP on FTX
1
7
It’s far too late here for me to have a coherent argument, but I’ll note for a later date that I’ll challenge you on this. I think Silver et al, are still on track and haven’t seen much off. I think it’s a lot of misunderstanding people have of political models vs finance etc.
1
2
Replying to
I don't think the miss was egregious--but I do think it's likely that 538 will have underestimated Trump here, and that their error bars will end up seeming a bit small here.
Tbf the only state I think they’ve predicted off was Florida. Other states are tracking right. I think once we get an even distribution of vote types counted this will show it’s a stronger Biden night than ppl thought. So I think their electoral numbers are decent
1
3
Replying to
Yeah that could end up being true, but I'd guess Florida won't be the only one--seems like it's also true of Ohio, Texas, a few senate seats, etc.
1
Show replies