A Smart Contract Oracle for Approximating Real-World, Real Number Values

Abstract

A key challenge of smart contract systems is the fact that many useful contracts require access to information that does not natively live on the blockchain. While miners can verify the value of a hash or the validity of a digital signature, they cannot determine who won an election, whether there is a flood in Paris, or even what is the price of ether in US dollars, even though this information might be necessary to execute prediction market, insurance, or financial contracts respectively. A number of promising projects and research developments have provided a better understanding of how one might construct a decentralized, binary oracle - namely an oracle that can respond by one of two possibilities, typically "yes" or "no", even while not requiring the interaction of a trusted third party. In this work, we extend these ideas to construct a general-purpose, decentralized oracle that can estimate the value of a real-world quantity that is in a dense totally ordered set, such as R. In particular, this proposal can be used to estimate real number valued quantities, such as required for a price oracle. We will establish a number of desirable properties about this proposal. Particularly, we will see that the precision of the output is tunable to users' needs.