This dataset captures 387,427 professional boxing matches in a single comprehensive table, offering complete referential integrity and a 64% overall completeness score. The structure supports direct analysis of fighter attributes—age, height, reach, weight, stance, and career record—against bout outcomes, enabling predictive modeling of wins and knockouts across weight classes and historical eras. However, **reach_B is 90% null** (349,554 missing values), creating a critical blind spot for one fighter in nearly every match; this asymmetry will distort any analysis of physical mismatch or reach advantage unless explicitly handled through imputation or exclusion. The data excels at outcome prediction and competitive balance assessment where complete attributes exist, particularly for fighter A profiles and weight-class stratification. It is well-suited to identifying era-specific trends, stance matchups, and the relationship between experience (record) and knockout probability. It is **not** suitable for comprehensive reach-differential analysis, bilateral fighter comparisons requiring symmetrical attributes, or any insight dependent on complete physical profiles for both competitors. Decision-makers can confidently use this dataset to model win probability and detect mismatches in age, weight, and experience—but must acknowledge that reach-based competitive analysis will reflect only half the picture unless the 90% null rate in reach_B is addressed through external enrichment or statistical inference.