Skip to content
This repository has been archived by the owner on Mar 24, 2021. It is now read-only.

Distributed prediction option #26

Open
msmith25 opened this issue Mar 26, 2012 · 0 comments
Open

Distributed prediction option #26

msmith25 opened this issue Mar 26, 2012 · 0 comments

Comments

@msmith25
Copy link

I would love to have the option of putting in several predictions for the same event with different confidence levels associated with each prediction. For instance, I just made a prediction about when I would arrive to a destination I'd never been to before. Because I'm trying to calibrate over time, I find I always want to pick a time that I can have 70% confidence in. But I'd like to be able to make six predictions that depend on the same outcome: one at 50% confidence, one at 60%, one at 70%, one at 80%, one at 90%, and one at 95%. That way I can calibrate across several different confidence levels at once whenever I'm dealing with a continuous random variable. To be even more concrete, here's how an example might look:

"I predict that this meeting will be done by..."
5:45pm (50% confidence)
6:00pm (60% confidence)
7:00pm (70% confidence)
8:30pm (80% confidence)
10:00pm (90% confidence)
12:00am (95% confidence)

And then later I could come back and put in when the meeting actually ended, automatically updating my ongoing calibration score at all six confidence levels.

Currently the only way to approximate this function would be to make five completely independent predictions about the same thing. It'd be nice to be able to put in just one answer and have all these confidence levels updated at once.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant