#29 - Some Scattered Thoughts on Superforecasting

Increments

We're back! Apologies for the delay, but Vaden got married and Ben was summoned to be an astronaut on the next billionaire's vacation to Venus. This week we're talking about how to forecast the future (with this one simple and easy trick! Astrologers hate them!). Specifically, we're diving into Philip Tetlock's work on Superforecasting.

So what's the deal? Is it possible to "harness the wisdom of the crowd to forecast world events"? Or is the whole thing just a result of sloppy statistics? We believe the latter is likely to be true with probability 64.9% - no, wait, 66.1%.

Intro segment:

"The Sentience Debate": The moral value of shrimps, insects, and oysters

Relevant timestamps:

• 10:05: "Even if there's only a one in one hundred chance, or one in one thousand chance, that insects are sentient given current information, and if we're killing trillions or quadrillions of insects in ways that are preventable or avoidable or that we can in various ways mitigate that harm... then we should consider that possibility."
• 25:47: "If you're all going to work on pain in invertebrates, I pity you in many respects... In my previous work, I was used to running experiments and getting a clear answer, and I could say what these animals do and what they don't do. But when I started to think about what they might be feeling, you meet this frustration, that after maybe about 15 years of research, if someone asks me do they feel pain, my answer is 'maybe'... a strong 'maybe'... you cannot discount the possibility."
• 46:47: "It is not 100% clear to me that plants are non sentient. I do think that animals including insects are much more likely to be sentient than plants are, but I would not have a credence of zero that plants are sentient."
• 1:01:59: "So the hard problem I would like to ask the panel is: If you were to compare the moral weight of one ant to the moral weight of one human, what ratio would you put? How much more is a human worth than an ant? 100:1? 1000:1? 10:1? Or maybe 1:1? ... Let's start with Jamie."

Main References:

• Superforecasting: The Art and Science of Prediction - Wikipedia
• How Policymakers Can Improve Crisis Planning
• The Good Judgment Project - Wikipedia
• Expert Political Judgment: How Good Is It? How Can We Know?: Tetlock, Philip E.: 9780691128719: Books - Amazon.ca

Additional references mentioned in the episode:

• The Drunkard's Walk: How Randomness Rules Our Lives
• The Black Swan: The Impact of the Highly Improbable - Wikipedia
• Book Review: Superforecasting | Slate Star Codex
• Pandemic Uncovers the Limitations of Superforecasting – We Are Not Saved
• My Final Case Against Superforecasting (with criticisms considered, objections noted, and assumptions buttressed) – We Are Not Saved

Use your Good Judgement and send us email at [e-mail].

Support Increments