```xml <result> <column_name>latency_ms</column_name> <family>additive</family> <subtype>ordinal_additive</subtype> <confidence>0.92</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, p50, p95, p99, distribution, sum (cross-entity only)</valid_reducers> <excluded_reducers>product (not a ratio), geomean (not a ratio), sum (along sequence - summing instantaneous latencies is meaningless)</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 is PATH with sequence axis, making ordinal_additive plausible for physical quantities. Column name "latency_ms" indicates an instantaneous measurement (duration in milliseconds) at each sequence step. Center of mass (mean=8228, median=6635) is far from both 0 and 1, ruling out multiplicative. This is an intensive physical quantity - summing latencies across time steps is meaningless, but averaging them or taking percentiles is the natural aggregation. Cross-entity sum at a fixed sequence step would be valid. Classified as ordinal_additive rather than intensive_additive because it's on a PATH table where the sequence-axis behavior differs from cross-entity behavior. </reasoning> <table_type_conflict>false</table_type_conflict> </result> ```