Open jg-you opened 5 months ago
(I noticed this while adding NUMBER markets. I will put in a PR for this, too.)
The calculation of probabilities seems wrong for full_history.. from what I can gather from the manifold code, the mechanism now involves a parameter p which depends on initial antes
.
Free responses were changed to multiple choice with "addAnswersMode"
This breaks several tests.
I am writing a PR, and I'm happy to contribute the code, though I'm curious about your attitude towards backward compatibility.
My proposal is not backward compatible, it
If that's all fine with you I'll put in the PR.