<result> <column_name>vRk</column_name> <family>additive</family> <subtype>ordinal_additive</subtype> <confidence>0.6</confidence> <requires_user_confirmation>false</requires_user_confirmation> <sum_meaningful>false</sum_meaningful> <signed>false</signed> <reducers> <primary_reducer>sum</primary_reducer> <valid_reducers>sum, mean, min, max, median, last, rank, normalized_mean</valid_reducers> <excluded_reducers>product (not a ratio), geomean (not a ratio)</excluded_reducers> <average_semantics>arithmetic_mean</average_semantics> </reducers> <inner_outer_measure> <inner_measure_reducer>n/a</inner_measure_reducer> <outer_measure_reducer>n/a</outer_measure_reducer> </inner_outer_measure> <multiplicative_partner_derivable>true</multiplicative_partner_derivable> <reasoning> The table is PATH with a Date sequence axis, so the full analysis applies. The name "vRk" suggests a rank (for example, a volume rank), and the data fit this. Values are positive integers from 1 to 2252, with a low distinct fraction (~0.9%) and a median of 273 that is far from 1.0. There are no ratio naming tokens and no center near 1.0, so the column is not multiplicative. Rank values are only ordinal along the sequence axis, so ordinal_additive fits better than real_additive. Summing ranks across timestamps is meaningless, so sum_meaningful is set to false. Confidence is moderate because the name is cryptic and no reconstruction signals are available. Ranks may also be treated as intensive levels, where mean would be the primary reducer. </reasoning> <table_type_conflict>false</table_type_conflict> </result>