Mantic Monday: Twitter Chaos Edition

Plus FTX charges, scandal markets - and oh yeah, wasn't there some kind of midterm recently? https://astralcodexten.substack.com/p/mantic-monday-twitter-chaos-edition Twitter! This is all going to be so, so obsolete by the time I finish writing it and hit the "send post" button. But here goes: 395 traders on this, so one of Manifold's biggest markets, probably representative. The small print defines a major outage as one that lasts more than an hour. See here for a good explanation of why some people expect Twitter outages. https://www.metaculus.com/questions/embed/13499/Polymarket is within 2% of Manifold. Metaculus here has slightly stricter criteria but broadly agrees. 71 traders, still pretty good, but I find it meaningless without a way to distinguish between "everything collapses, Elon sells it for peanuts to scavengers" vs. "Elon saves Twitter, then hands it over to a minion while he moves on to a company building giant death zeppelins". Oh, here we go. 20 traders, they think Musk will stay in charge. 23 traders. Twitter was profitable in 2018 and 2019, then went back to being net negative in 2020 and 2021 (I don't know why) . I don't think it's been very profitable lately, so it would be a feather in Musk's cap if he accomplished this. 24 traders. Twitter's mDAU have consistently gone up in the past. DAU is slightly different and I think more likely to include bots. 26 traders. One thing I like about Manifold is that it lets you choose any point along the gradient from "completely objective" (eg Twitter's reported DAU count) to "completely subject" (eg whether the person who made the market thinks something is better or worse). This at least uses a poll as its resolution method. But the poll will be in the comments of this market, which means it will mostly be by people who invested in this market, who'll have strong incentives to manipulate it. Maybe Manifold should add a polling platform to their service?

There are no comments yet.

Please log in to write the first comment.