```xml <result> <column_name>opponent_2_round_ko_percentage</column_name> <family>additive</family> <subtype>intensive_additive</subtype> <confidence>0.85</confidence> <requires_user_confirmation>false</requires_user_confirmation> <sum_meaningful>false</sum_meaningful> <signed>false</signed> <reducers> <primary_reducer>mean</primary_reducer> <valid_reducers>mean, min, max, median, distribution</valid_reducers> <excluded_reducers>sum (summing percentages across entities is meaningless), product (not a ratio/factor)</excluded_reducers> <average_semantics>arithmetic_mean</average_semantics> </reducers> <inner_outer_measure> <inner_measure_reducer>min</inner_measure_reducer> <outer_measure_reducer>max</outer_measure_reducer> </inner_outer_measure> <multiplicative_partner_derivable>false</multiplicative_partner_derivable> <reasoning> DIMENSION table with no sequence axis strongly gates against multiplicative. The column name contains "percentage" but the values are in [0, 100] scale (not ratios near 1.0), representing an absolute performance metric level per opponent entity. This is an intensive property: the percentage describes a characteristic of each opponent, and averaging across opponents is meaningful while summing is not (you cannot add "10% + 10.32% = 20.32% of something"). The lack of sequence axis and DIMENSION grain confirm this is a snapshot level, not a ratio that compounds over time. </reasoning> <table_type_conflict>false</table_type_conflict> </result> ```