iter48-卡1：卡①(状态合并累加制批): 六行stack_policy refresh→accumulate+schema词表换代+容器累加臂(_apply_accumulate只加时长,层数快照不变)+roll_multiplier五段式探针(V1/V2/V3)+table_structure_guard V4+damage_controller标签换代+甲案裁决随动(2026-11-02用户裁): _ineffective_row载荷deferred自描述键(非行为改动)+round_flow _row_phase分类补deferred→战斗段——相位单调断言归绿,EV三锚0.9148/1.0012/1.1948逐位不变

game/core/damage_controller.gd

@@ -688,4 +688,4 @@ func _target_for(rule: String) -> Variant:
 
 func _ineffective_row(t: Dictionary, reason: String) -> void:
 	emit_row_cb.call({"type": "reward_ineffective", "source_class": String(t.source_class), "origin_cell": t.origin_cell, "highlight": String(t.highlight)}, int(t.round_index),
-		"%s 未生效（%s）" % [String(t.display), reason], 0, {"batch_key": String(t.context.get("batch_key", ""))})
+		"%s 未生效（%s）" % [String(t.display), reason], 0, {"batch_key": String(t.context.get("batch_key", "")), "deferred": true})