diff --git a/app/src/androidTest/java/org/libremediaconverter/FailsOnEmulatorApi37.kt b/app/src/androidTest/java/org/libremediaconverter/FailsOnEmulatorApi37.kt
index 7ff9511..8db3c5a 100644
--- a/app/src/androidTest/java/org/libremediaconverter/FailsOnEmulatorApi37.kt
+++ b/app/src/androidTest/java/org/libremediaconverter/FailsOnEmulatorApi37.kt
@@ -54,10 +54,13 @@ annotation class FailsOnEmulatorApi37
  * **The fourth carrier is the one to read that sentence carefully for.**
  * `pickingAFileThroughTheSystemPickerFillsInTheFileCard` was marked on 2026-09-05 for aborting
  * `system_server` rather than for failing (#108), and on the gating leg it passed two runs of
- * four. On the advisory leg it runs after the rotation test has already taken the framework down,
- * which is why the count still holds there — measured, not assumed, and the measurement is the
- * reason this line did not have to become two numbers. If it ever starts reporting three failures
- * out of four, read that as this test having got lucky rather than as an image that improved.
+ * four. The count still holds on the advisory leg, and that was **measured rather than assumed**:
+ * `api37-debug.yml` run 34008889182, dispatched with this annotation as its filter, reports
+ * `expected: 4, received: 4, failed: 4` (and `completed cleanly: no`, which is this job's normal).
+ * It fails there because it runs alongside the rotation test, which takes the framework down first
+ * — so the reason this line did not have to become two numbers is a property of the advisory leg,
+ * not of the test. If it ever reports three failures out of four, read that as this test having
+ * got lucky rather than as an image that improved.
  *
  * So: adding or removing a [FailsOnEmulatorApi37] means changing this number, in this file, in
  * the same diff. The report says so on the run itself if you forget — it prints the tree's own
diff --git a/docs/api-37-emulator-crash.md b/docs/api-37-emulator-crash.md
index d8905a3..911043a 100644
--- a/docs/api-37-emulator-crash.md
+++ b/docs/api-37-emulator-crash.md
@@ -492,6 +492,18 @@ input entirely, so the picker's process can be removed from a task no key press
 `pickTheFixture`'s whole-picker retry — which exists for exactly this — becomes reachable again.
 That is a fix to the test on every level, not to API 37.
 
+**It was made to bite before it was believed.** On a local API 36 emulator, with the walk cut short
+so the picker is left open and in front and with `device.pressBack()` removed, so that nothing but
+the force-stop can close it:
+
+| | result |
+|---|---|
+| with `forceStopThePicker()` | **passes** — `ActivityManager: Force stopping com.google.android.documentsui ... from pid 5334`, `Killing 5269:com.google.android.documentsui (adj 0)`, a second `PickActivity` opens, the retry completes the pick |
+| with the one call removed | **fails** — `the system picker would not close: after 4 back presses ... com.google.android.documentsui is in front`, which is the API 37 failure verbatim |
+
+The unmutated class passes on that emulator either way, which is the point of running the mutation
+at all: the recovery path is unreachable on a healthy device, so a green suite says nothing about it.
+
 #### The correction that produced that table
 
 **The first version of this section said both tests failed, and put the marker on the class.** The