diff --git a/artifact.1/tests/test_core.py b/artifact.1/tests/test_core.py
new file mode 100644
index 0000000..ae6cc91
--- /dev/null
+++ b/artifact.1/tests/test_core.py
@@ -0,0 +1,64 @@
+import pytest
+from typing import List, Dict
+from src.artifact_1 import core
+
+
+class TestResult:
+    def __init__(self, group: str, pinned_status: str, warn_rate: float, unknown_rate: float, delta_t_rate: float):
+        self.group = group
+        self.pinned_status = pinned_status
+        self.warn_rate = warn_rate
+        self.unknown_rate = unknown_rate
+        self.delta_t_rate = delta_t_rate
+
+
+def make_sample_data() -> List[TestResult]:
+    return [
+        TestResult('A', 'pinned', 0.1, 0.05, 0.02),
+        TestResult('A', 'pinned', 0.2, 0.1, 0.03),
+        TestResult('A', 'unpinned', 0.3, 0.15, 0.04),
+        TestResult('B', 'pinned', 0.15, 0.07, 0.01),
+        TestResult('B', 'unpinned', 0.35, 0.12, 0.08),
+        TestResult('B', 'unpinned', 0.4, 0.18, 0.06),
+    ]
+
+
+def test_analyze_ab_data_structure():
+    data = make_sample_data()
+    result = core.analyze_ab_data(data)
+    assert isinstance(result, dict)
+    # groups should contain A and B, with pinned/unpinned stratum inside
+    for group_key in ['A', 'B']:
+        assert group_key in result
+        assert 'pinned' in result[group_key] or 'unpinned' in result[group_key]
+
+
+def test_aggregation_values():
+    data = make_sample_data()
+    result = core.analyze_ab_data(data)
+    # A-pinned
+    a_pinned = result['A']['pinned']
+    assert pytest.approx(a_pinned['warn_rate_mean'], rel=1e-6) == (0.1 + 0.2) / 2
+    assert pytest.approx(a_pinned['unknown_rate_mean'], rel=1e-6) == (0.05 + 0.1) / 2
+    assert pytest.approx(a_pinned['delta_t_rate_mean'], rel=1e-6) == (0.02 + 0.03) / 2
+
+    # B-unpinned
+    b_unpinned = result['B']['unpinned']
+    assert pytest.approx(b_unpinned['warn_rate_mean'], rel=1e-6) == (0.35 + 0.4) / 2
+    assert pytest.approx(b_unpinned['unknown_rate_mean'], rel=1e-6) == (0.12 + 0.18) / 2
+    assert pytest.approx(b_unpinned['delta_t_rate_mean'], rel=1e-6) == (0.08 + 0.06) / 2
+
+
+def test_invalid_input_raises():
+    # Input validation should fail for non-list input
+    with pytest.raises((TypeError, ValueError)):
+        core.analyze_ab_data('invalid_input')
+
+    # Input validation should fail for missing attributes
+    class Bad:
+        def __init__(self):
+            self.group = 'A'
+
+    bad_data = [Bad()]
+    with pytest.raises((AttributeError, ValueError)):
+        core.analyze_ab_data(bad_data)