Whether Algorand network-level backrunning MEV searchers use optimized inputs

Ascertain whether MEV searchers that execute network-level backrunning arbitrages on Algorand choose optimized input parameters (e.g., profit-maximizing trade sizes) in their transactions during block construction.

Background

The paper compares arbitrage opportunities discovered from finalized block states with realized arbitrage profits and finds that most opportunities are exploited within the same block, suggesting that competitive MEV strategies operate at the network (mempool) level.

Because the analysis uses finalized block states rather than mempool-level data, the authors cannot determine whether searchers’ executed backruns use profit-maximizing inputs, leaving open the question of optimization in observed strategies.