Interesting thing I noticed about my calibration graph.
May 6, 2023

Almost all of my NO bets in resolved markets were correct, while my YES bets were a bit more scattered - significantly more of them were incorrect. I wonder if that says something meaningful about my predictions - maybe I'm too optimistic/naive, maybe I err on the side of "YES" because I generally expect things to happen more often than they don't happen, etc. - or if it's just noise because I've only been here for about six weeks and the vast majority of the markets I've traded in (~90%) have yet to resolve. My Metaculus binary calibration (102 resolved, been using the site for a few months now) is "Overconfidence 0%" i.e. overconfident by less than one percent, with an average Brier score of 0.087, so I expect that it's probably noise, but we'll see. Maybe my predictions are less accurate than I think they are.

Maybe the consequence of all those Tears of the Kingdom markets closing and resolving over the next few weeks will be that my calibration gets less skewed, since I've placed a lot of YES bets that I'm very confident in. I'm going to be so, SO vindicated if the Zelink and time travel markets resolve yes like I expect - I was betting YES on those when they were below 40%.

© Manifold Markets, Inc.Terms + Mana-only TermsPrivacyRules