How can more people learn about rationalism and adjacent topics (expected value, risk management, probabilistic thinking)? This is a question we've asked ourselves for a long time, and we now think we have found a solution: a forecasting tournament.
For example, let's take the question of "Who will win the Nobel Prize in Physics in 2025?".
One could of course first look if there are already some markets there (one could do arbitrage!), no, there aren't. So maybe we can look if some more physics experts have already voiced their opinion on this matter (and if so, how much to trust those individuals and their suggestions, what weigh should we assign their suggestions, have they proposed concrete probabilities or vaguely named some individuals, do they have a good track record of maybe predicting the winners in previous years?).
And then we already have some initial data points, we can then look at what work the Nobel Prize Committee rewarded in previous years, does that help us? Are there some articles in journals or newspapers speculating about who might receive a Nobel Prize?
And since we are living in a modern age, what do the AI systems say? Is their output in line with what we previously have researched or the complete opposite? Should we trust them? Which model gives better information? Are there maybe, if it overall is not really accurate given our own research, are there at least some valuable hints we might want to follow through with and investigate?
We might also want to read up and watch a lot of videos on the most interesting discussions and research in physics of the previous years to make a better informed decision at the end. And could we not maybe even create new information that isn't anywhere available, and thus can't be know by the other participants or AI systems, by asking the top physicists of this world for an interview or just answering our poll we mailed them on who they think will win the Nobel Prize in 2025?
Then we have maybe made a model which assigns X a probability of 20% to win the Nobel Prize, but the current probability is 15%. That sounds good, but wait, how confident are we? How much of our points should we allocate to this forecast, well, luckily there is something called the Kelly Criterion which can help us.
Now, we've not only learned about a multitude of wonderfully interesting concepts in Physics which might be rewarded with a Nobel Prize this year, but also became a better decision-maker and understood many key concepts of rationalist thinking.
We believe that through forecasting tournaments we can make this world more rational, it might never before has been that needed in our world. We also learn about how much we should trust AI systems (which one to trust more? on what topics?) very intuitively because well, when we make the wrong forecasts often enough because AI has misled us, then we'll update our priors and see the utility of not blindly trusting it.