Abstract: To become the premier decentralized liquidity layer the @aptos trading engine needs to serve the needs of professional traders, institutional investors, high-frequency traders (HFT), and retail investors. In this talk, we will focus on the inner workings of HFT to reach a deeper understanding of the role and needs of HFT in a decentralized setting. We will argue that CLOBs are the only empirically backed price discovery mechanism, and briefly review the evidence that current AMM designs may not be viable in the long term. We will then revisit the optimal market-making model of Avellaneda and Stoikov (2008) and point out that, while insightful and celebrated, this model suffers from several shortcomings, among them the absence of price ticks and priority queues, the fixed order sizes, and other unrealistic assumptions regarding the price dynamics. We will then introduce alternative optimal market-making models, some of which have closed-form solutions, and are also easier to adapt to the multi-asset case. We will close with a discussion of the practical considerations that the stochastic price dynamics of the published market-making solutions of the Hamilton-Jacobi-Bellman equation often neglect, such as quote sniping by informed traders, and provide heuristic descriptions of what market makers do in practice to reduce their risk. When: July 16th at 11 am Where: aptos hq