💹Prediction markets
What are prediction markets and why they are great!
What are prediction markets?
Overview of prediction markets and their history.
Prediction markets are a type of open market that allows people to predict specific outcomes rewarding them for correct predictions in accordance with some predefined reward function.
Prediction markets have a long history, going back as early as the 16th century. However, their benefits were first clearly outlined by the Austrian School economist Friedrich Hayek in the 1940s. Robin Hanson, the inventor of the so-called Logarithmic Market Scoring Rule, has likewise contributed significantly to popularising the concept from the 1990s onwards as a way to improve policy decision making.
Notable off-chain prediction markets include the Iowa Electronic Markets (US election markets), Futuur (US election markets) and Manifold (general topic markets).
More recently, on-chain prediction markets have begun to gain popularity. As the name suggests, they are implemented on blockchain and hence offer more transparency to their users both in the prediction process (making sure that you given a fair amount of outcome shares for the money paid) and resolution process (making sure that the correct outcome is determined in the most objective way possible). The first such market, Augur launched in 2015. Presently, the biggest on-chain prediction market is Polymarket (launched in 2020).
What makes on-chain prediction markets so special?
On-chain prediction markets are special for several big reasons:
Enable accurate estimation of the likelihood of events.
Allow us to tranparently rank and reward people based on their predictive ability.
Why use on-chain prediction markets for on-chain reputation?
As mentioned in the previous section, prediction market are a transparent method to rank and reward individuals based on their predictive ability. And predicting is precisely what crypto content creators and influencers do on a daily basis.
Some do it more implicitly cherry picking interesting projects and write about their strengths, weaknesses, and future prospects or more explicitly in tweeting market forecasts or shilling certain coins.
On-chain prediction markets therefore are an excellent format to judge the quality of crypto content creators. They are a simple to understand and transparent methodology, unlike some other off-chain reputation models out there that use Machine Learning and web scraping to collect information on influencers/content creators. At the same time, they also allow for a broader assessment of knowledge than social trading platform that are only really geared for price prediction. Finally, by their nature, prediction markets offer crypto content creators an environment they are used.
Pythia's prediction market design.
Pythia's prediction markets have three major components to them, the user, the markets themselves, and the oracles that resolve the said markets. The user (a content creators in our case) makes predictions and accumulates reputation. Content creators also have the ability to create new markets.
Decentralized oracles are another key components of the architecture. Decentralized oracles (Chainlink and Kleros arbitration courts) resolve markets, that is identify the correct outcome, in the most objective way possible. Chainlink accumulates and averages information from a variety of sources to arrive at an answer, Kleros leverages game theoretic mechanisms to source public wisdom to determine the outcome. In the case of Kleros, Pythia (Pythia's market contract to be specific) routes its requests through reality.eth, which then queries Kleros courts. Decentralized oracles ultimately determine whether the user earns reputation or not.
On the whole, Pythia's market design is similar to what you would expect to see in vanilla on-chain prediction markets. However, there is one key difference: Pythia makes content creator prediction private before resolution (i.e. before the correct outcome is determined by a decentralized oracle). We shall cover the rationale behind this in the following section, but briefly the reason for this additional architectural element is to protect the value of the predictions. Making predictions private, in other word, ensures that we prevent prediction copying and identify those who truly understand market and industry trends.
Last updated
