<result> <column_name>f_boxer_ranking</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</valid_reducers> <excluded_reducers>sum (summing ranking points across entities is meaningless), product (not a ratio)</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> Table type is DIMENSION with no sequence axis, making multiplicative highly implausible. The column name "f_boxer_ranking" and sample values showing "points" indicate a ranking score or rating level assigned to each boxer entity. This is an absolute intensity measure (a level) where summing across boxers is meaningless—averaging is the natural aggregation. The dtype is object (likely due to formatting with "points" suffix), but the numeric values (20.15, 25.73, 29.65) are absolute levels centered away from both 0 and 1, consistent with intensive_additive. </reasoning> <table_type_conflict>false</table_type_conflict> </result>