Record reproducible plateau campaigns and selected-direction probes

experiments/plateau_control_cases.json

@@ -0,0 +1,47 @@
+[
+  {
+    "id": "n1_J0p05_no_central_R40",
+    "arguments": [
+      "--index",
+      "1",
+      "--angular-momentum",
+      ".05",
+      "--no-central",
+      "--restart",
+      "40",
+      "--max-newton",
+      "8"
+    ],
+    "notes": "Matched central-phase removal control; baseline solver policy."
+  },
+  {
+    "id": "n1_J0p05_no_central_R160",
+    "arguments": [
+      "--index",
+      "1",
+      "--angular-momentum",
+      ".05",
+      "--no-central",
+      "--restart",
+      "160",
+      "--max-newton",
+      "8"
+    ],
+    "notes": "Only Krylov restart differs from preceding model."
+  },
+  {
+    "id": "n1p5_J0_no_central_R160",
+    "arguments": [
+      "--index",
+      "1.5",
+      "--angular-momentum",
+      "0",
+      "--no-central",
+      "--restart",
+      "160",
+      "--max-newton",
+      "8"
+    ],
+    "notes": "Higher-index live control before initial-guess safeguard."
+  }
+]