A smart contract architecture for automated decentralized risk markets based on the Kelly Criterion designed for long term capital sustainability.

Abstract

A key blocker to scalability in derivative AMMs is the risk of non survival for the LP. Many existing architectures are based on Black-Scholes-Merton pricing which is known to inadequately characterize crypto instruments and risk ruin for LPs. In this paper we introduce the Kelly Machine - an automatic market maker architecture for risk management based on the Kelly Criterion and survival constraints.

While the Kelly Criterion is traditionally used to implement optimal risk sizing strategies, here we present an extension of the use of the Kelly Criterion into the realm of optimal combined risk pricing & sizing strategies. The result of this optimization are so-called “Kelly Optimal bonding curves which allow the creation of automated Kelly Criterion based trading strategies.

The Kelly Machine was designed in the hopes of contributing towards higher levels of market stability and risk management within the decentralized ecosystem, and will be released as a public good.

1. General system overview

An automatic market maker (AMM) is a set of smart contracts that autonomously manage capital on behalf of Liquidity Providers (LPs), in order to create automated on-chain services to buyers.

https://lh3.googleusercontent.com/oUuQsSi3qivmtmtDBKHNyM7zHn7zpI03yjTRqoxQTRLNTAOY0mmjhaDR9YAxfibfpJv4ItyMtlFE0rxv-0kan-gi3PL3YS0BxmOAHFPOd3ckY2BV41Y1DtS4DTdYmPo99a9-WtKN

The Kelly Machine is a new class of AMM designed to create on-chain risk instruments such as asset price insurance. For example, a fully collateralized insurance on the price of Ethereum at $3000 for a duration of 1 month.

Participants

Potion’s Kelly Machine intermediates two sets of users: Liquidity Providers (LPs), and Insurance Buyers.

LPs act as insurance providers - they provide capital to underwrite and collateralize the insurance policies being created and sold by the AMM. LPs charge a premium every time they collateralize an insurance contract. LPs participate in the hopes that premiums will more than offset losses from insurance claims, and that in the long run they will experience capital growth.

Insurance buyers buy insurance contracts which are securely collateralized on chain. At contract expiration time, the policy automatically assigns the correct insurance payout across the buyer and the seller of the contract.  Buyers participate because they want to protect themselves against risk (hedge application), or because they want to trade optionality in hopes the market crashes and they collect a large payout from a small premium (up to 10X and even 100X+  payout).

Pricing heuristic