diff --git a/changes/176.bugfix.rst b/changes/176.bugfix.rst
new file mode 100644
index 0000000..ab6d6e4
--- /dev/null
+++ b/changes/176.bugfix.rst
@@ -0,0 +1 @@
+Fix PlackettLuce rating adjustment for ties.
diff --git a/openskill/models/weng_lin/plackett_luce.py b/openskill/models/weng_lin/plackett_luce.py
index 44c8e4d..aba3d1d 100644
--- a/openskill/models/weng_lin/plackett_luce.py
+++ b/openskill/models/weng_lin/plackett_luce.py
@@ -848,7 +848,7 @@ def _compute(
                     delta += (
                         i_mu_over_ce_over_sum_q * (1 - i_mu_over_ce_over_sum_q) / a[q]
                     )
-                    if team_q.rank == team_i.rank:
+                    if q == i:
                         omega += (1 - i_mu_over_ce_over_sum_q) / a[q]
                     else:
                         omega -= i_mu_over_ce_over_sum_q / a[q]
diff --git a/tests/models/data/plackettluce.json b/tests/models/data/plackettluce.json
index a9a0241..36937a7 100644
--- a/tests/models/data/plackettluce.json
+++ b/tests/models/data/plackettluce.json
@@ -86,21 +86,21 @@
         ],
         "team_2": [
             {
-                "mu": 23.083700519022745,
+                "mu": 20.599656070925114,
                 "sigma": 8.222674397835641
             },
             {
-                "mu": 21.717476072569045,
+                "mu": 20.47545384852023,
                 "sigma": 8.277466291367071
             }
         ],
         "team_3": [
             {
-                "mu": 21.717476072569045,
+                "mu": 20.47545384852023,
                 "sigma": 8.277466291367071
             },
             {
-                "mu": 23.083700519022745,
+                "mu": 20.599656070925114,
                 "sigma": 8.222674397835641
             }
         ],
@@ -160,7 +160,7 @@
     "ties": {
         "team_1": [
             {
-                "mu": 23.241111866333558,
+                "mu": 21.63766806988004,
                 "sigma": 8.310709773172306
             }
         ],
@@ -176,15 +176,15 @@
         ],
         "team_3": [
             {
-                "mu": 21.479966996534408,
+                "mu": 19.87652320008089,
                 "sigma": 8.237522411103104
             },
             {
-                "mu": 21.479966996534408,
+                "mu": 19.87652320008089,
                 "sigma": 8.237522411103104
             },
             {
-                "mu": 21.479966996534408,
+                "mu": 19.87652320008089,
                 "sigma": 8.237522411103104
             }
         ]
@@ -261,4 +261,4 @@
             }
         ]
     }
-}
\ No newline at end of file
+}
diff --git a/tests/models/weng_lin/test_common.py b/tests/models/weng_lin/test_common.py
index 44e8cd7..f0f8d91 100644
--- a/tests/models/weng_lin/test_common.py
+++ b/tests/models/weng_lin/test_common.py
@@ -205,20 +205,52 @@ def test_ladder_pairs():
 
 @pytest.mark.parametrize("model", MODELS)
 @pytest.mark.parametrize("tie_score", [-1, 0, 0.1, 10, 13.4])
-def test_ties(model, tie_score):
+@pytest.mark.parametrize("num_teams", [2, 5, 10])
+@pytest.mark.parametrize("team_size", [1, 2, 5, 10])
+@pytest.mark.parametrize("tie_type", ["score", "rank"])
+def test_ties(model, tie_score, num_teams, team_size, tie_type) -> None:
     model_instance = model()
+    teams = [
+        [model_instance.rating() for _ in range(team_size)] for _ in range(num_teams)
+    ]
+    player_mu_before = [player.mu for team in teams for player in team]
+    assert all(
+        mu == player_mu_before[0] for mu in player_mu_before
+    ), f"Model {model.__name__} with score {tie_score}: All players should start with equal mu"
+
+    player_sigma_before = [player.sigma for team in teams for player in team]
+    assert all(
+        sigma == player_sigma_before[0] for sigma in player_sigma_before
+    ), f"Model {model.__name__} with score {tie_score}: All players should start with equal sigma"
+
+    if tie_type == "score":
+        scores = [tie_score for _ in range(num_teams)]
+        new_teams = model_instance.rate(teams, scores=scores)
+    else:  # rank
+        ranks = [tie_score for _ in range(num_teams)]
+        new_teams = model_instance.rate(teams, ranks=ranks)
+
+    player_mu_after = [player.mu for team in new_teams for player in team]
+    assert all(
+        mu_after == mu_before
+        for mu_after, mu_before in zip(player_mu_after, player_mu_before)
+    ), f"Model {model.__name__} with score {tie_score}: All players should end with equal mu"
+    player_sigma_after = [player.sigma for team in new_teams for player in team]
+    assert all(
+        sigma_after <= sigma_before
+        for sigma_after, sigma_before in zip(player_sigma_after, player_sigma_before)
+    ), f"Model {model.__name__} with score {tie_score}: All players should end with lower or equal sigma"
 
-    player_1 = model_instance.rating()
-    player_2 = model_instance.rating()
 
-    result = model_instance.rate(
-        [[player_1], [player_2]], scores=[tie_score, tie_score]
-    )
+@pytest.mark.parametrize("model", MODELS)
+def test_ties_with_close_ratings(model) -> None:
+    model_instance = model()
+
+    player_1 = model_instance.rating(mu=30)
+    player_2 = model_instance.rating(mu=20)
+
+    new_teams = model_instance.rate([[player_1], [player_2]], ranks=[0, 0])
 
-    # Both players should have the same rating change
-    assert (
-        result[0][0].mu == result[1][0].mu
-    ), f"Model {model.__name__} with score {tie_score}: Players should have equal mu after tie"
-    assert (
-        result[0][0].sigma == result[1][0].sigma
-    ), f"Model {model.__name__} with score {tie_score}: Players should have equal sigma after tie"
+    # ratings should converge on ties.
+    assert new_teams[0][0].mu < 30
+    assert new_teams[1][0].mu > 20