<result> <column_name>2SP</column_name> <family>uncertain</family> <subtype>none</subtype> <confidence>0.4</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>product (values are not a ratio centered on 1.0), 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>false</multiplicative_partner_derivable> <reasoning> The table is PATH with a Date sequence axis, so the full analysis applies. Multiplicative is strongly ruled out: the mean (28.7) and median (27) are far from 1.0, the range runs to 112, zeros are present, the name carries no ratio token, and log-symmetry gain is weak. The values look like non-negative integers (very low distinct fraction), which is consistent with either a count-like real_additive quantity or a level-like intensive or ordinal quantity. The name "2SP" is opaque and there are no units or reconstruction results to separate these subtypes, so I abstain on the subtype. I listed conservative reducers that are safe for any additive reading and excluded sum until the meaning is confirmed. </reasoning> <table_type_conflict>false</table_type_conflict> </result>