diff --git a/tests/test_arc_1d.py b/tests/test_arc_1d.py
index 6c32f79e..0190d353 100644
--- a/tests/test_arc_1d.py
+++ b/tests/test_arc_1d.py
@@ -98,7 +98,7 @@ def test_arc_1d_scoring():
     assert dataset.score_answer(entry["answer"], entry) == 1.0
 
     # Test partial match (answer contained within response)
-    assert dataset.score_answer(f"The answer is: {entry['answer']}", entry) == 0.5
+    assert dataset.score_answer(f"The answer is: {entry['answer']}", entry) > 0.5
 
     # Test incorrect answer
     assert dataset.score_answer("wrong answer", entry) == 0.01
diff --git a/tests/test_products.py b/tests/test_products.py
index 34ff1623..d794e28a 100644
--- a/tests/test_products.py
+++ b/tests/test_products.py
@@ -111,7 +111,9 @@ def test_products_scoring():
     assert dataset.score_answer("wrong", item) == 0.01, "Wrong answer should score 0.01"
 
     # Test scoring with partial match (answer contained in response)
-    assert dataset.score_answer(f"The answer is {item['answer']}", item) == 0.5, "Partial match should score 0.5"
+    assert (
+        dataset.score_answer(f"The answer is {item['answer']}", item) > 0.1
+    ), "Partial match should scored len(oracle_answer)/len(answer)"
 
     # Test scoring with None
     assert dataset.score_answer(None, item) == 0.0, "None should score 0.0"