Do Algorand MEV searchers use optimized inputs for network-level backruns?

Determine whether MEV searchers on the Algorand blockchain who perform network-level backruns select profit-maximizing (optimized) trade inputs for their arbitrage strategies during block construction.

Background

The paper studies MEV on Algorand, a first-come-first-served (FCFS) blockchain where transaction ordering hinges on network latency rather than fees. The authors design an arbitrage detection algorithm and evaluate it on finalized block states, finding that most profitable opportunities are executed within the same block, leaving little for post-block-state discovery.

Because their analysis is based on finalized blocks rather than mempool-level observations during block construction, the authors cannot directly observe the input choices used by live MEV searchers when executing backruns. This leads to an unresolved question about whether searchers’ intra-block arbitrage executions are input-optimized, which has implications for understanding searcher efficiency and the competitiveness of MEV strategies on Algorand.