### 5.1 Performance Posture The dataset exhibits a straightforward performance profile anchored by the **fighters** dimension table (2,760 rows) and a smaller **popular_matches** reference table (152 rows). At this scale, query performance remains well within acceptable bounds for predictive modeling workflows, with no immediate scaling concerns. The absence of validated joins between tables suggests each operates independently—**fighters** likely serves as the primary feature source for outcome prediction, while **popular_matches** may provide supplementary context or validation cases. Current storage is adequate for analytical workloads, though text-based fighter lookups will benefit from targeted optimization as query complexity increases. ### 5.2 Key Optimizations | Target | Optimization Type | Recommendation | Strength | |--------|-------------------|----------------|----------| | fighters.name | Text Search | keyword | High | Implementing keyword indexing on **fighters.name** will accelerate fighter-specific queries essential to pre-fight feature extraction and model training pipelines. When analysts filter training data by specific boxers or validate predictions against historical performance, this optimization eliminates full-table scans and reduces query latency by an estimated 60–80%. Given that outcome prediction hinges on rapidly assembling fighter attributes across multiple bouts, this single enhancement directly supports the core analytical goal while maintaining infrastructure simplicity appropriate to the dataset's modest footprint.