diff --git a/runs/frontier-fidelity-envelope-v1/analyze_qwen235_fixed_pd_state.py b/runs/frontier-fidelity-envelope-v1/analyze_qwen235_fixed_pd_state.py
index 934605b..103d176 100644
--- a/runs/frontier-fidelity-envelope-v1/analyze_qwen235_fixed_pd_state.py
+++ b/runs/frontier-fidelity-envelope-v1/analyze_qwen235_fixed_pd_state.py
@@ -695,10 +695,16 @@ def analyze(
     if decision_contrast is None:
         verdict = "STOP: real Running proxy has insufficient exact simulator support"
     elif decision_contrast["total"] < 0:
-        verdict = (
-            "State-composition mismatch is insufficient: after reweighting Frontier "
-            "to measured real decode composition, it still predicts TP8 faster."
-        )
+        if exact_contrast is not None:
+            verdict = (
+                "State-composition mismatch is insufficient: after reweighting Frontier "
+                "to measured real decode composition, it still predicts TP8 faster."
+            )
+        else:
+            verdict = (
+                "Active-batch-count mismatch is insufficient: after reweighting Frontier "
+                "to the frozen-real Running proxy, it still predicts TP8 faster."
+            )
     else:
         verdict = (
             "Coarse active-batch state can flip Frontier's ordering, but iteration-level "