<result> <column_name>high</column_name> <family>additive</family> <subtype>intensive_additive</subtype> <confidence>0.6</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 price levels 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>The table is STATE with no sequence axis, so multiplicative is implausible and no cumprod or cumsum reconstruction is possible. The name "high" (likely the high of an OHLC-style price bar) has no ratio token. The values are strictly positive, with a median of about 36 and a range from 0.009 to 306,500, so they are far from 1.0 and look like absolute price levels. The large log_symmetry_gain just reflects a skewed price distribution and does not override the gate. A posted price level is an intensive quantity, so mean is the natural rollup and sum is excluded. Confidence is moderate because the column could also be a generic extreme value with unclear semantics.</reasoning> <table_type_conflict>false</table_type_conflict> </result>