Hide table of contents

Forecasting tools and prediction markets allow you to state your credence in something precisely and view others' credences. This can be an important way of gathering information from a group or sharing information precisely between individuals. 

This post will introduce a few tools for generating and sharing predictions. Many of these tools are excellent; they've been grouped together because of their similarity and number. 

One way I think forecasting tools can be useful for collaborative truth-seeking is that spending some time using them can improve both your calibration and your ability to precisely define questions. I think this can then greatly improve your ability to communicate your implicit beliefs explicitly in conversation or text. I'd be interested to hear other people's opinions on this view! 

Forecasting tools and putting precise numbers on predictions in general is important because it lets a community approach truth over time better by letting people have a more transparent track record to learn who is in fact good at predicting.

This post will discuss the following tools:

  • Metaculus 
  • Manifold Markets
  • Pastcasting
  • Confido
  • PredictionBook
  • Elicit

When can forecasting/prediction tools be useful?

  • Practice predicting so that you're better able to communicate your beliefs verbally (especially PredictionBook, Pastcasting)
  • Get community predictions relevant to deciding where to donate, e.g.:
    •  AI milestones
    • Cultured meat
    • Factory farming
    • Global poverty
    • Politics 
  • Make a private question to ask people in your organisation to predict project outcomes or engagement (Confido)
  • Add a prediction question on your post to see if people's views change after reading (Elicit)
  • Find out how people feel about a controversial topic (with people who feel more strongly able to influence the result more; e.g. here)
  • Learn to forecast (especially Pastcasting)
  • Perform curious self-referential research on prediction markets[1] (Manifold Markets)

Metaculus

Metaculus is a reputation-based site for soliciting and aggregating predictions.

The official Metaculus tutorials are suitable for beginners as a way to introduce forecasting, probability distributions, and a basic understanding of how to behave under uncertainty. 

Tutorial #1 is over-explaining for those with a basic understanding of the forecasting, but still a good introduction to using Metaculus specifically. 

For those familiar with forecasting, tutorials #2 and #3 are more useful as a (small) amount of calibration practice and building familiarity with the Metaculus interface. 

Examples

Note that you can embed metaculus questions in forum posts by posting the direct link into the editor:

Manifold Markets 

Manifold Markets is a play money prediction market with user-created questions. The play money can be redeemed as a charitable donation. Because it works as a market, it can be used in lots of interesting ways-- you can ask people to predict which opportunities your organisation will fund (as one factor to influence your decision, maybe), you can play chess, and you can even implement futarchy.

A text tutorial is available on the wiki here

Examples

Note that you can embed Manifold Markets questions into forum posts by pasting the direct link into the editor:

Other Tools

Pastcasting is a forecasting platform using resolved questions from other sites (e.g. Metaculus) to shorten the feedback loop and practice forecasting, particularly over long time horizons. The official FAQ is a great introduction. In partcular, I'd like to highlight Pastcasting's feature Vantage Search:

Vantage Search is a custom search engine that only shows results from before the vantage point to prevent information leaks from the future.

Confido is tool that provides an instance of a closed forecasting platform. Forecasts are entered similarly to on metaculus, but questions are intended for a single group or organisation. This is useful for questions requiring insider knowledge like "How will the EA conference we are planning go" or "How many employees will we have by the end of the year". Trying the demo version is a good way to understand the tool.

PredictionBook is simple tool for prediction by entering a % credence only. You write can write a question, put the date when you should know the true answer by, and enter a credence. On the date you entered, you'll be notified to give the true answer. 

Predictions can be public or private; I find it's particularly useful for producing personal questions "in 2023, will I go 60 days without allergy symptoms?" as a way of improving your calibration (including an automatic calibration graph[2]). 

Elicit Forecast is a tool that lets you easily set up  questions people can put credences on, and shows an aggregate of them, but is currently unsupported[3]. At present, only binary questions (with a % chance input) work as far as I can tell; date or number estimate questions are currently non-functioning. There is no guarantee any functionality will continue. 

Binary elicit forecasts currently work as embeds:

1%
2%
3%
4%
5%
6%
7%
8%
9%
10%
11%
12%
13%
14%
15%
16%
17%
18%
19%
toribrk (15%)
20%
21%
22%
23%
24%
25%
26%
27%
28%
29%
30%
31%
32%
33%
34%
35%
36%
37%
38%
39%
40%
41%
42%
43%
44%
45%
46%
47%
48%
49%
50%
51%
52%
53%
54%
55%
56%
57%
58%
59%
Saul Munn (56%)
60%
61%
62%
63%
64%
65%
66%
67%
68%
69%
70%
71%
72%
73%
74%
75%
76%
77%
78%
79%
80%
81%
82%
83%
84%
85%
86%
87%
88%
89%
90%
91%
92%
93%
94%
95%
96%
97%
98%
99%
1%
Will the US retail price of lemons be above $2.10 on 01-12-23?
99%

Personal Experience

I don't honestly use prediction or forecasting tools much in my day-to-day life[4]. If this were to change I think I'd be most likely to use PredictionBook to improve my calibration on everyday questions that I need to be able to predict, or manifold markets simply for fun. 

I think if you're regularly making big decisions like grant making or running projects in AI, biosecurity, etc., forecasting tools like Metaculus or Manifold could be useful, particularly the comments (where other people explain their reasoning). Similarly, I think Confido would be useful for a sufficiently-sized organisation. 

Elicit is useful because of its ability to be embedded & interacted with in forum posts. If somebody picks up supporting the tool & fixes the current issues, I imagine I'd add it to a lot more forum posts. 

Try it Yourself!

You can access the below market to buy shares yourself here.

I'd also suggest making an account on PredictionBook and formalising 3 predictions that you're currently making implicitly! Here are some suggestions to get the ball rolling:

  • The completion date of a current project
  • How much money you'll donate to charity before 31/12/2023
  • How many people will read your next forum post
  • How successful you will be with your new year's resolution
  • Your total spending this month
  • What time you'll get home today

Tomorrow at 6pm GMT we'll also be running a short event  in the EA GatherTown to discuss forecasting tools & do a short exercise about them!

Also, join us at 6pm GMT today in the GatherTown to discuss the use of Squiggle!

 

  1. ^
  2. ^

    i.e. a graph of "% you assigned to an outcome" against "% chance that that actually happened" (which would be the line y=x if you were perfectly calibrated)

  3. ^

    Private communication, Ought

  4. ^

    Though we did get feedback from people who do on various drafts of this post, don't worry!

  5. Show all footnotes
Comments


No comments on this post yet.
Be the first to respond.
Curated and popular this week
 ·  · 16m read
 · 
At the last EAG Bay Area, I gave a workshop on navigating a difficult job market, which I repeated days ago at EAG London. A few people have asked for my notes and slides, so I’ve decided to share them here.  This is the slide deck I used.   Below is a low-effort loose transcript, minus the interactive bits (you can see these on the slides in the form of reflection and discussion prompts with a timer). In my opinion, some interactive elements were rushed because I stubbornly wanted to pack too much into the session. If you’re going to re-use them, I recommend you allow for more time than I did if you can (and if you can’t, I empathise with the struggle of making difficult trade-offs due to time constraints).  One of the benefits of written communication over spoken communication is that you can be very precise and comprehensive. I’m sorry that those benefits are wasted on this post. Ideally, I’d have turned my speaker notes from the session into a more nuanced written post that would include a hundred extra points that I wanted to make and caveats that I wanted to add. Unfortunately, I’m a busy person, and I’ve come to accept that such a post will never exist. So I’m sharing this instead as a MVP that I believe can still be valuable –certainly more valuable than nothing!  Introduction 80,000 Hours’ whole thing is asking: Have you considered using your career to have an impact? As an advisor, I now speak with lots of people who have indeed considered it and very much want it – they don't need persuading. What they need is help navigating a tough job market. I want to use this session to spread some messages I keep repeating in these calls and create common knowledge about the job landscape.  But first, a couple of caveats: 1. Oh my, I wonder if volunteering to run this session was a terrible idea. Giving advice to one person is difficult; giving advice to many people simultaneously is impossible. You all have different skill sets, are at different points in
 ·  · 10m read
 · 
Regulation cannot be written in blood alone. There’s this fantasy of easy, free support for the AI Safety position coming from what’s commonly called a “warning shot”. The idea is that AI will cause smaller disasters before it causes a really big one, and that when people see this they will realize we’ve been right all along and easily do what we suggest. I can’t count how many times someone (ostensibly from my own side) has said something to me like “we just have to hope for warning shots”. It’s the AI Safety version of “regulation is written in blood”. But that’s not how it works. Here’s what I think about the myth that warning shots will come to save the day: 1) Awful. I will never hope for a disaster. That’s what I’m trying to prevent. Hoping for disasters to make our job easier is callous and it takes us off track to be thinking about the silver lining of failing in our mission. 2) A disaster does not automatically a warning shot make. People have to be prepared with a world model that includes what the significance of the event would be to experience it as a warning shot that kicks them into gear. 3) The way to make warning shots effective if (God forbid) they happen is to work hard at convincing others of the risk and what to do about it based on the evidence we already have— the very thing we should be doing in the absence of warning shots. If these smaller scale disasters happen, they will only serve as warning shots if we put a lot of work into educating the public to understand what they mean before they happen. The default “warning shot” event outcome is confusion, misattribution, or normalizing the tragedy. Let’s imagine what one of these macabrely hoped-for “warning shot” scenarios feels like from the inside. Say one of the commonly proposed warning shot scenario occurs: a misaligned AI causes several thousand deaths. Say the deaths are of ICU patients because the AI in charge of their machines decides that costs and suffering would be minimize
 ·  · 32m read
 · 
Authors: Joel McGuire (analysis, drafts) and Lily Ottinger (editing)  Formosa: Fulcrum of the Future? An invasion of Taiwan is uncomfortably likely and potentially catastrophic. We should research better ways to avoid it.   TLDR: I forecast that an invasion of Taiwan increases all the anthropogenic risks by ~1.5% (percentage points) of a catastrophe killing 10% or more of the population by 2100 (nuclear risk by 0.9%, AI + Biorisk by 0.6%). This would imply it constitutes a sizable share of the total catastrophic risk burden expected over the rest of this century by skilled and knowledgeable forecasters (8% of the total risk of 20% according to domain experts and 17% of the total risk of 9% according to superforecasters). I think this means that we should research ways to cost-effectively decrease the likelihood that China invades Taiwan. This could mean exploring the prospect of advocating that Taiwan increase its deterrence by investing in cheap but lethal weapons platforms like mines, first-person view drones, or signaling that mobilized reserves would resist an invasion. Disclaimer I read about and forecast on topics related to conflict as a hobby (4th out of 3,909 on the Metaculus Ukraine conflict forecasting competition, 73 out of 42,326 in general on Metaculus), but I claim no expertise on the topic. I probably spent something like ~40 hours on this over the course of a few months. Some of the numbers I use may be slightly outdated, but this is one of those things that if I kept fiddling with it I'd never publish it.  Acknowledgements: I heartily thank Lily Ottinger, Jeremy Garrison, Maggie Moss and my sister for providing valuable feedback on previous drafts. Part 0: Background The Chinese Civil War (1927–1949) ended with the victorious communists establishing the People's Republic of China (PRC) on the mainland. The defeated Kuomintang (KMT[1]) retreated to Taiwan in 1949 and formed the Republic of China (ROC). A dictatorship during the cold war, T