diff --git a/CLAUDE.md b/CLAUDE.md
index 0dc37c0..707e495 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -211,9 +211,12 @@ event's length. It adds **no schema field**, so the feed contract is untouched.
 - **A repeating event the reader marked done leaves the strip.** They have said there is nothing
   left to do; keeping a tickable chip for it is the app arguing with them. Their logged days are
   untouched, so unmarking it brings the chip and the streak straight back.
-- **Detection is a default, not a verdict.** The reader can mark any event as repeating, or unmark
-  one detection got wrong (`progress.daily`, resolved by `resolveDaily`), and `prefs.detectDaily`
-  switches the guessing off entirely. Store an override only when it *disagrees* with detection —
+- **Detection is a guess, not a verdict — and it ships off.** `prefs.detectDaily` defaults to
+  `false` and the control is labelled experimental: wording is a weak signal and gets it wrong in
+  both directions, so a new reader opts in rather than out. The default moves nothing for an
+  existing reader, whose stored `prefs` wins. The reader can mark any event as repeating, or unmark
+  one detection got wrong (`progress.daily`, resolved by `resolveDaily`), whether the guessing is
+  on or off. Store an override only when it *disagrees* with detection —
   recording agreement would freeze today's guess and stop a better parser from ever reaching that
   event. Neither control ever deletes a mark or a logged day, so both are reversible.
 
diff --git a/docs/DATA-MODEL.md b/docs/DATA-MODEL.md
index ecfd0fe..ba45953 100644
--- a/docs/DATA-MODEL.md
+++ b/docs/DATA-MODEL.md
@@ -293,7 +293,14 @@ It adds no schema field, so nothing about the feed contract or the event ID chan
 for one event and wins outright (`resolveDaily`); absent means they have not said, so detection
 stands. `prefs.detectDaily` switches the guessing off altogether, leaving only events they marked
 themselves — it silences detection rather than deleting anything, so every mark and every logged day
-survives and switching it back on restores exactly what was there. An override that merely
+survives and switching it back on restores exactly what was there.
+
+**Detection is off by default and labelled experimental in the UI.** Wording is a weak signal and it
+is wrong in both directions, so a new reader starts with only the standing `dailies:<game>` chores
+and whatever they mark themselves; opting in is one checkbox. The default applies to new readers
+only — a stored `prefs` keeps whatever value it has, because turning it off under a reader who has
+been ticking auto-detected checklists would pull those chips out of the strip with no explanation.
+An override that merely
 agrees with detection is **not stored** (`dailyOverride`) — freezing today's guess into their data
 would stop a later parser improvement from ever reaching that event. This is the only field in
 `progress` that changes what the app *shows* rather than recording what the reader did, which is
diff --git a/src/client/components/Controls.tsx b/src/client/components/Controls.tsx
index 43d86eb..6cb2220 100644
--- a/src/client/components/Controls.tsx
+++ b/src/client/components/Controls.tsx
@@ -87,9 +87,9 @@ export function Controls({
           </label>
 
           {/* Detection reads the source's wording and is wrong in both
-              directions. Off leaves only the events the reader marked, and
-              discards nothing — every mark and logged day survives, so it can
-              be switched back on. */}
+              directions, so it ships off and says so. Off leaves only the
+              events the reader marked, and discards nothing — every mark and
+              logged day survives, so it can be switched back on. */}
           <label className="flex cursor-pointer select-none items-start gap-2 text-xs text-muted">
             <input
               type="checkbox"
@@ -99,9 +99,13 @@ export function Controls({
             />
             <span>
               Spot daily events automatically
+              <span className="ml-1.5 rounded-full border border-hairline px-1.5 py-0.5 align-[1px] text-[0.5625rem] font-medium uppercase tracking-wider text-faint">
+                Experimental
+              </span>
               <span className="mt-0.5 block max-w-xs leading-relaxed text-faint">
-                Off, only events you mark yourself get a checklist. Your ticks
-                and streaks are kept either way.
+                Guessed from what the source wrote, so it misses some and
+                invents others. Off, only events you mark yourself get a
+                checklist. Your ticks and streaks are kept either way.
               </span>
             </span>
           </label>
diff --git a/src/client/state/usePrefs.ts b/src/client/state/usePrefs.ts
index 3f0cf99..00c468b 100644
--- a/src/client/state/usePrefs.ts
+++ b/src/client/state/usePrefs.ts
@@ -23,6 +23,10 @@ export interface Prefs {
    * Whether to guess which events repeat daily from what the source printed.
    * Off leaves only the ones the reader marked themselves; it never discards a
    * mark or a logged day, so it is reversible.
+   *
+   * Off by default: the guess reads source wording and is wrong in both
+   * directions, so a reader starts with only the dailies they chose. Readers
+   * who already switched it on keep it — stored prefs win over this default.
    */
   detectDaily: boolean;
   showCompleted: boolean;
@@ -40,7 +44,7 @@ function defaults(): Prefs {
     hiddenGames: [],
     focusGame: null,
     sort: "ending",
-    detectDaily: true,
+    detectDaily: false,
     showCompleted: true,
     showIgnored: false,
     regionConfirmed: false,