Forecasting is a research field & practical community devoted to improving our judgement about the future. The best forecasters — called “superforecasters” — make probabilistic predictions on falsifiable questions, then update their forecasts as new evidence comes in. Through deliberate practice, forecasters improve by tracking their calibration and applying specific strategies (like Bayes’ rule, base rates, reference classes, etc).

Predicting the future is useful for a variety of reasons. Imagine having a better understanding of election outcomes, or the impacts of climate change, or where the next global pandemic would strike, or where the economy will be 5 years from now. Forecasters make, track, and improve on their predictions in a wide range of topics. (Scroll through Metaculus to get a good sense!)

Forecasting is also a skill that can concretely and materially improve your own life — you might predict your grade on a test, or if you’ll be offered an internship. Having calibrated, clear-headed judgements about your own future can help you navigate important decisions and worry less.

Think 1-day hackathon/olympiad/debate tournament, but for forecasting the future — teams predict on topics ranging from geopolitics to celebrity twitter patterns to financial asset prices, and the best forecasters get thousands of dollars in cash prizes and access to exclusive internships.

Day of, teams give probabilistic predictions for a couple hours on about 30 given questions (with breaks for lunch & speakers). Teams are 3-5 competitors, and we’ll place you on a team if you don’t already have one in mind. A few months after the tournament, all questions resolve and winners are announced/awarded; until then, you can track how your team is doing in real time.

Are you:

1. A university student*

Great! You’re eligible to compete.

*In some circumstances, we allow high schoolers and grad students to compete. Please reach out if you’re unsure.

Not much!

A tiny bit of politics, stats, and economics can be helpful — if you know “the United Nations,” “skewed distribution,” and “gross domestic product,” you’ll be fine. (And if you don’t know what those mean, that’s fine too!)

This is a list of the previous questions we’ve asked — you’ll be forecasting on questions similar to those.

Great question! There are a few important tools to learn, but after that, it’s mainly just practice. Fortunately, forecasting can be pretty fun!

You can learn how to forecast through this video series from Alex Lawsen and get started on Metaculus using these interactive tutorials. To practice estimation strategy, you can play the Calibration game on Quantified Intuitions; you can also practice forecasting & prediction (with slightly slower feedback loops) on Metaculus and Manifold.

TLDR: A few thousand dollars, exclusive internships, and a whole lot more. Depends a bit on your region.

Thousands of dollars of cash prizes awarded to the most accurate forecasting teams, and exclusive internships at our sponsors to outstanding competitors. There are sometimes also smaller prizes for high-quality forecasting rationales and for continued predictions after the competition. All prizes depend slightly on the details of the tournament you’re competing in.

OPTIC usually awards about $1,000-$3,000 in cash prizes for the first competition; the exact prize amount will depend on the exact competition you’re competing in.

Teams are 3-5 competitors, and we’ll place you on a team if you don’t already have one in mind. Fill out the registration form as an individual, and let us know who you’d like to be on a team with when you arrive at the competition.

Technical: log scoring, with some extra details — contact us if you have any technical questions.

Nontechnical: the scoring system rewards confidence in proportion to accuracy. Highly confident forecasts that end up happening are rewarded highly; highly confident forecasts that end up not happening are penalized heavily; forecasts made with low confidence that end up happening are rewarded slightly; forecasts made with low confidence that end up not happening are penalized lightly.

After the forecasting period on the day of the tournament ends, we wait for the actual outcome of each event to be known. All of the questions resolve within a couple months, after which the predictions are scored. At that point, the most accurate forecasters are declared the winners, and are made public & receive their prizes. Also, note that different competitions can have different timelines — make sure to check your competition, or contact us if you have any questions.

Awesome! We really want to help you — we’d love to provide outreach materials, instructional docs, funding & sponsorshops, and anything else you might need. Check out opticforecasting.com/clubs for more info!

Send us a message, we’d love to talk to you! Please have a pretty low bar for what counts as a good reason to email us — feel free to just say hi :)

You can contact us here.