diff --git a/CHANGELOG.md b/CHANGELOG.md
index 4097689..19e1ff0 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -6,6 +6,13 @@ first release is 3.0.0 because exit codes and the cookie store changed in ways a
 
 ## Unreleased
 
+### Fixed
+- MAM ranking accepted a different snatched book by the same author when the id3 title and artist were
+  already usable: search is by filename (`01.mp3`, `file.m4b`), and the author-only gate skipped the title
+  check unless the release name had to supply title or authors. `The Guest` tagged as James Patterson was
+  filed as *Along Came a Spider* (token_sort 67, above `matchrate` 60). Ranking now requires a title match
+  whenever the id3/parsed title is not junk, and applies `Config/matchrate` the same way Audible already did.
+
 ## 3.0.4 - 2026-09-22
 
 ### Fixed
diff --git a/myx_classes.py b/myx_classes.py
index f3fae2f..263bf8a 100644
--- a/myx_classes.py
+++ b/myx_classes.py
@@ -1092,9 +1092,13 @@ def getMAMBooks(self, cfg, bookFile:BookFile):
         extension = f'"{bookFile.getExtension()}"'
         #for RANKING (never for the MAM query string) use the parsed release name where the id3 tags are junk
         rankBook, parsedApplied = self.applyParsedName(self.ffprobeBook, cfg) if (self.ffprobeBook is not None and not interactive) else (None, [])
-        # when the parse supplied title or authors, the title gate must see the search title
-        # (id3 or parsed), not the file basename — otherwise `file.m4b` rejects a correct MAM hit
-        rankTitle = rankBook.title if ("title" in parsedApplied or "authors" in parsedApplied) else title
+        # MAM is searched by filename, so a usable id3/parsed title must still gate the snatched hits.
+        # Using the file basename here used to skip the title gate whenever tags were already good
+        # (author-only), and filed Along Came a Spider as The Guest.
+        if rankBook is not None and not myx_names.isJunkTitle(rankBook.title):
+            rankTitle = rankBook.title
+        else:
+            rankTitle = title
     
         # Search using book key and authors (using or search in case the metadata is bad)
         print(f"Searching MAM for\n\tTitleFilename: {title}\n\tauthors:{authors}")
@@ -1144,9 +1148,13 @@ def getMAMBooks(self, cfg, bookFile:BookFile):
 
                 else:
                     bestMatchRate=0
+                    minMatchRate = int(cfg.get("Config/matchrate"))
                     #find the best match
                     print(f"Finding the best MAM match out of {len(books)} results")
                     targetBook = '|'.join([book.title, book.getAuthors(), book.getSeriesParts()])
+                    # a usable title (id3 or parsed) must pass the title gate: MAM search is filename-based,
+                    # so author-only used to accept that author's other snatched books (token_sort 67 >= 60)
+                    requireTitle = rankBook is None or not myx_names.isJunkTitle(rankBook.title)
             
                     for abook in books:
                         #if this book is snatched, include in the match
@@ -1155,7 +1163,7 @@ def getMAMBooks(self, cfg, bookFile:BookFile):
                             #otherwise, if maybe this title is close enough
                             #print (f"{abook.title} by {abook.authors}...")
                             authorOK = bool(len(book.authors) and myx_utilities.isThisMyAuthorsBook(book.authors, abook, cfg))
-                            if authorOK and "title" not in parsedApplied and "authors" not in parsedApplied:
+                            if authorOK and not requireTitle:
                                 mamBook = '|'.join([abook.getAuthors(), abook.getCleanTitle(), abook.getSeriesParts()])
                                 if add_narrators:
                                     mamBook = '|'.join([mamBook, abook.getNarrators()])
@@ -1173,7 +1181,7 @@ def getMAMBooks(self, cfg, bookFile:BookFile):
 
                             print(f"\tMatch Rate: {matchRate}\n\tSearch: {targetBook}\n\tResult: {mamBook}\n\tBest Match Rate: {bestMatchRate}\n")
                             
-                            if (matchRate[fuzzy_match] > bestMatchRate):
+                            if (matchRate[fuzzy_match] > bestMatchRate) and (matchRate[fuzzy_match] >= minMatchRate):
                                 bestMatchRate=matchRate[fuzzy_match]
                                 self.bestMAMMatch=abook
         else:
diff --git a/tests/test_matching.py b/tests/test_matching.py
index f7bce77..91035eb 100644
--- a/tests/test_matching.py
+++ b/tests/test_matching.py
@@ -612,6 +612,25 @@ def test_different_mam_title_by_the_parsed_author_is_rejected(self):
         self.assertIsNone(best)
         self.assertIn("This book doesn't have a matching title or author", out)
 
+    def test_good_id3_author_and_title_still_reject_a_different_book_by_that_author(self):
+        # usable id3 title AND artist: parse does not apply, so the author-only gate used to accept
+        # Patterson's Along Came a Spider as The Guest (token_sort 67 >= matchrate 60). MAM search is
+        # the file basename (`file.m4b` / `01.mp3`), which is how the wrong snatched hit is returned.
+        best, out = self._run("The Guest",
+                              id3_book("The Guest", ["James Patterson"], 400 * 60),
+                              [_mam_book("Along Came a Spider", ["James Patterson"])])
+        self.assertIsNone(best)
+        self.assertIn("This book doesn't have a matching title or author", out)
+        self.assertNotIn("Using parsed release name", out)
+
+    def test_good_id3_author_and_title_still_accept_the_matching_title(self):
+        best, out = self._run("The Guest",
+                              id3_book("The Guest", ["James Patterson"], 400 * 60),
+                              [_mam_book("The Guest", ["James Patterson"])])
+        self.assertIsNotNone(best)
+        self.assertEqual(best.title, "The Guest")
+        self.assertNotIn("Using parsed release name", out)
+
 
 class NarratorInArtistTagTest(unittest.TestCase):
     def secrets(self):