Intro to Forecasting (EA San Francisco + Stanford EA discussion event)
cross-posted fromFacebook. (I will check both this thread and the FB event for comments)
Date: 2020/04/29
Time: 19:00-21:00 (PDT)
Location: Zoom videoconferencing (recommend web portal)
You might have heard of forecasting by now. Many of the cool kids* are doing it, using fancy terms like “brier score,” “metaculus”, “log-odds”, “calibration” and “modeling.” You might have heard of superforecasters: savvy amateurs who make robustly better forecasts on geopolitical events than trained analysts at the CIA. What you might not have learned is that these skills are eminently trainable: In the original Good Judgement Project, researchers have found that a short 1-hour training course can robustly improve accuracy over the course of a year!
Next Wednesday, EA SF is collaborating with Stanford Effective Altruism to host an introductory event on forecasting. Together, we will practice (super)forecasting techniques: the skills and ways of thinking that allowed savvy amateurs to make better forecasts on geopolitical events than trained analysts at the CIA.
Here are some of the topics we’ll try to cover and practice in small groups, time permitting:
Base Rates: Outside View vs. Inside View Credence References: Thinking in Credible Intervals Controlling for Scope: Consider the probability distribution across different outcomes than posed by the question, such as longer/shorter timeframes Analytics: Fermizing, assessing signal vs. noise. Controlling for biases and fallacies Comment: Making explicit rationales to prevent hindsight bias and share information Compare: Explain your reasoning, benefit from viewpoint diversity, and accelerate learning Update: Revise your forecast as new information comes in or you change your view
Structure: We’ll meet together briefly to go over the details and then split into smaller groups with 3-6 group members each, including a group leader. Each group will be given a discussion sheet that they can copy and group leaders will be given an answer key.
We’ll be using Zoom as our platform as that allows the most seamless split into smaller groups. For security reasons, we recommend using the Zoom Web portal over the Mac App.
We expect many of the attendees to be new to forecasting, but also several people who’re very experienced forecasters and/or otherwise quite plugged in into the current state-of-the-art of forecasting. Depending on who shows up, it might also make sense to have a Q&A in addition to the small group discussions.
As this is a virtual event, all are welcome. However, we only have limited small group leader capacity, so in the (very fortunate!) world where many more people show up than we expect, groups may be asked to nominate their own group leaders instead of having an appointed one with prior experience managing EA discussions.
Intro to Forecasting (EA San Francisco + Stanford EA discussion event)
cross-posted from Facebook. (I will check both this thread and the FB event for comments)
Date: 2020/04/29
Time: 19:00-21:00 (PDT)
Location: Zoom videoconferencing (recommend web portal)
You might have heard of forecasting by now. Many of the cool kids* are doing it, using fancy terms like “brier score,” “metaculus”, “log-odds”, “calibration” and “modeling.” You might have heard of superforecasters: savvy amateurs who make robustly better forecasts on geopolitical events than trained analysts at the CIA. What you might not have learned is that these skills are eminently trainable: In the original Good Judgement Project, researchers have found that a short 1-hour training course can robustly improve accuracy over the course of a year!
Next Wednesday, EA SF is collaborating with Stanford Effective Altruism to host an introductory event on forecasting. Together, we will practice (super)forecasting techniques: the skills and ways of thinking that allowed savvy amateurs to make better forecasts on geopolitical events than trained analysts at the CIA.
Background
https://en.wikipedia.org/wiki/The_Good_Judgment_Project
https://en.wikipedia.org/wiki/Superforecasting
http://www.academia.edu/download/37711009/2015_-_superforecasters.pdf
https://aiimpacts.org/evidence-on-good-forecasting-practices-from-the-good-judgment-project-an-accompanying-blog-post/
Here are some of the topics we’ll try to cover and practice in small groups, time permitting:
Base Rates: Outside View vs. Inside View
Credence References: Thinking in Credible Intervals
Controlling for Scope: Consider the probability distribution across different outcomes than posed by the question, such as longer/shorter timeframes
Analytics: Fermizing, assessing signal vs. noise. Controlling for biases and fallacies
Comment: Making explicit rationales to prevent hindsight bias and share information
Compare: Explain your reasoning, benefit from viewpoint diversity, and accelerate learning
Update: Revise your forecast as new information comes in or you change your view
Structure: We’ll meet together briefly to go over the details and then split into smaller groups with 3-6 group members each, including a group leader. Each group will be given a discussion sheet that they can copy and group leaders will be given an answer key.
We’ll be using Zoom as our platform as that allows the most seamless split into smaller groups. For security reasons, we recommend using the Zoom Web portal over the Mac App.
We expect many of the attendees to be new to forecasting, but also several people who’re very experienced forecasters and/or otherwise quite plugged in into the current state-of-the-art of forecasting. Depending on who shows up, it might also make sense to have a Q&A in addition to the small group discussions.
As this is a virtual event, all are welcome. However, we only have limited small group leader capacity, so in the (very fortunate!) world where many more people show up than we expect, groups may be asked to nominate their own group leaders instead of having an appointed one with prior experience managing EA discussions.
Hope to see you there!
*and some of the uncool kids