You are the Lead Architect for a High-Frequency Trading (HFT) platform. You need to design a distributed data store that holds current market prices. Analyze this system through the lens of the CAP theorem: 1. Between Consistency (C), Availability (A), and Partition Tolerance (P), which combination must you prioritize and why? 2. If a network partition occurs, describe the specific behavior your system must exhibit to prevent financial loss or erroneous trades. 3. Compare and contrast the choice of using an AP system (like Cassandra or DynamoDB) with tunable consistency versus a CP system (like HBase or MongoDB with strong consistency) for this specific use case. 4. Propose a conflict resolution strategy for writes that occur during a partition, specifically addressing the issue of 'last write wins' in a financial context where order matters.