A recent report from Canada’s central bank discusses “cheating” on a blockchain through double spending and likens the potential issue to classic incentive compatibility constraints within game theory.
Researchers Jonathan Chiu and Thorsten V. Koeppl from the Bank of Canada, the country's central bank, published a working paper[1] detailing their model to test "cheating" on a distributed ledger such as a blockchain. Chiu and Koeppl set up a Cournot[2] game of mining to represent a blockchain's proof-of-work (PoW), such as Bitcoin's consensus protocol.
In applying their model to two examples – a securities settlement and a goods transaction – the researchers assert that double spending, or the risk that a cryptocurrency may be spent twice, can be represented by incentive compatibility[3] constraints as part of traditional mechanism design[4] (contained within the fields of economics and game theory).
In other words, to sustain a PoW-based blockchain and prevent double spending on it, miners must have a sufficient reward (e.g., cryptocurrency payments) to motivate them to accurately and honestly maintain the blockchain.
Through this lens, the researchers note that "sufficient overall mining activities help discourage dishonest behavior" and that "increasing rewards will increase the effort and, hence, the computational investment by miners, making it harder to double spend." Although the idea of issuing miners rewards to detract potentially dishonest actors is by no means novel, framing the issue of double spending as an economic design problem allows Chiu and Koeppl to analyze it in a different light. For instance, they argue that cryptocurrencies would work "best for a fairly homogeneous group of transactions with small incentives to double spend, such as retail payments."
The researchers also suggest that longer confirmation lags (which occur between the moment when a
