```xml <result> <column_name>docs_indexed</column_name> <family>additive</family> <subtype>ordinal_additive</subtype> <confidence>0.85</confidence> <requires_user_confirmation>false</requires_user_confirmation> <sum_meaningful>true</sum_meaningful> <signed>false</signed> <reducers> <primary_reducer>sum</primary_reducer> <valid_reducers>sum, mean, min, max, median, count, last</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>min</inner_measure_reducer> <outer_measure_reducer>max</outer_measure_reducer> </inner_outer_measure> <multiplicative_partner_derivable>true</multiplicative_partner_derivable> <reasoning> Table is PATH with sequence axis, so ordinal_additive is plausible. Column name "docs_indexed" strongly signals a count/quantity (real_additive naming), but on a PATH table this represents an instantaneous state at each sequence step rather than a cumulative total. The large magnitude (max ~14M, mean ~643K), strictly non-negative values, and presence of zeros all confirm additive family. The key distinction: this is a snapshot count at each timestamp, not a delta or flow - summing across entities at a fixed time is meaningful, but along the sequence axis it should use ordinal reducers (last value, or normalized aggregation). Classified as ordinal_additive rather than real_additive because it's a level measurement on a PATH table. </reasoning> <table_type_conflict>false</table_type_conflict> </result> ```