```xml <result> <column_name>Previous Points</column_name> <family>additive</family> <subtype>intensive_additive</subtype> <confidence>0.72</confidence> <requires_user_confirmation>true</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 snapshot point values 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 gate: STATE with no sequence axis makes multiplicative highly implausible. The column name "Previous Points" indicates a snapshot of accumulated points at a prior state—an absolute level, not a rate or ratio. The wide range (0 to 17469) and lack of center near 1.0 confirm this is not multiplicative. This is a snapshot balance/level: summing "previous points" across different entities is meaningless (you cannot add Alice's 100 points to Bob's 50 points to get a meaningful 150), but averaging to understand typical prior achievement is valid. Classified as intensive_additive (a level measure on a STATE table), though it shares characteristics with out-of-scope tensorial snapshot values. </reasoning> <table_type_conflict>false</table_type_conflict> </result> ```