diff --git a/docs/library.md b/docs/library.md
index 9ef59c4..59a8bfe 100644
--- a/docs/library.md
+++ b/docs/library.md
@@ -40,7 +40,7 @@ Diese Woche (the last 7 days), Diesen Monat (30 days), then the months — after
 Kino feed. In „Folgen“ (sort „Nach Zeitraum, seltene zuerst“, the default) and „Neu“
 the podcasts that seldom come go first within a group: `interval_days` of each feed
 in `/podcasts/state/`, the mean gap between its newest 10 episodes (one episode: its
-age, at least 30 days, the same when all came on one day; kept an hour). Heard is a filter chip, not an order.
+age, at least 30 days, the same when all came within an hour; kept an hour). Heard is a filter chip, not an order.
 
 ### Shared code: `static/js/folders.js` (diora-web, every platform)
 
diff --git a/podcasts/tests.py b/podcasts/tests.py
index ee4cf8b..d1fb5a1 100644
--- a/podcasts/tests.py
+++ b/podcasts/tests.py
@@ -817,6 +817,13 @@ class EpisodeListTests(TestCase):
                                           pub_date=_at(1, 1) + timedelta(minutes=k))
         feeds = {x['id']: x for x in self.client.get('/podcasts/state/').json()['feeds']}
         self.assertGreater(feeds[f.id]['interval_days'], 30)
+        hourly = PodcastFeed.objects.create(user=self.user, rss_url='https://example.org/h', title='Stündlich')
+        for k in range(10):
+            PodcastEpisode.objects.create(feed=hourly, guid=f'h{k}', title=f'H{k}', audio_url=f'https://example.org/h{k}.mp3',
+                                          pub_date=_at(1, 1) + timedelta(hours=k))
+        cache.clear()
+        feeds = {x['id']: x for x in self.client.get('/podcasts/state/').json()['feeds']}
+        self.assertAlmostEqual(feeds[hourly.id]['interval_days'], 1 / 24, places=2)
 
     def test_all_newest_first_without_date_last(self):
         data = self.get()
diff --git a/podcasts/views.py b/podcasts/views.py
index 4ca92d7..10c00e2 100644
--- a/podcasts/views.py
+++ b/podcasts/views.py
@@ -816,7 +816,7 @@ INTERVAL_SINGLE_DAYS = 30
 
 def _intervals(user):
     """How often each feed brings an episode, in days: the mean gap between its newest
-    INTERVAL_EPISODES. One episode only (or all on one day): its age, at least INTERVAL_SINGLE_DAYS. The lists
+    INTERVAL_EPISODES. One episode only (or all within an hour): its age, at least INTERVAL_SINGLE_DAYS. The lists
     put the seldom ones first within a day or week (after Kino's feed; feeds.js does the
     same for blogs). One query (the newest episodes per feed by a window), kept an hour."""
     key = f'podcasts:intervals:{user.id}'
@@ -836,8 +836,9 @@ def _intervals(user):
     out = {}
     for feed_id, ds in dates.items():
         span = (max(ds) - min(ds)).total_seconds() / 86400
-        if len(ds) == 1 or span < 1:
-            # One episode, or all of them at once (an archive, a course): no rhythm to go by.
+        if len(ds) == 1 or span < 1 / 24:
+            # One episode, or all within an hour (an archive, a course put up at once): no rhythm
+            # to go by. Hourly news still has one.
             days = max((now - max(ds)).total_seconds() / 86400, INTERVAL_SINGLE_DAYS)
         else:
             days = span / (len(ds) - 1)