Application-controlled execution (ACE) allows applications to dictate the order of transactions interacting with them, with Hyperliquid’s cancel prioritization serving as a key example. In this model, cancel orders are prioritized over take orders within transaction batches. The article outlines four mechanisms for implementing ACE, focusing on their ability to support cancel prioritization and analyzing their respective benefits and drawbacks. The first mechanism discussed is app-chains, exemplified by Hyperliquid, which enforces order directly on-chain at the protocol level. This method offers fine control over transaction sequencing but raises concerns about censorship resistance. Proposers could potentially exclude cancel orders to favor take orders, undermining the mechanism's integrity. The article also highlights the isolation of app-chains from broader crypto markets, where bridging transactions can introduce delays. Next, the article examines batching on general-purpose chains, specifically through Asynchronous Message Queues (AMQs). This design allows applications to queue transactions in one block and execute them in the next, creating a batch without altering the underlying consensus rules. While this approach can maintain some level of composability, it introduces complexity in managing the queued transactions, and the enforcement of cancel prioritization relies on application-level logic rather than on-chain consensus. The analysis also touches on the general implementation of ACE within smart contracts, where consensus enforces transaction order at the state transition level. This method presents challenges, such as increased complexity in block building and reduced composability across applications. By evaluating these designs, the article aims to clarify the landscape of ACE, demonstrating that effective cancel prioritization can be achieved through various approaches, each with its own trade-offs.