diff --git a/pyroengine/engine.py b/pyroengine/engine.py
index e109ed0f..55289d3f 100644
--- a/pyroengine/engine.py
+++ b/pyroengine/engine.py
@@ -80,12 +80,12 @@ class Engine:
     def __init__(
         self,
         model_path: Optional[str] = None,
-        conf_thresh: float = 0.2,
+        conf_thresh: float = 0.35,
         model_conf_thresh: float = 0.05,
         max_bbox_size: float = 0.4,
         api_url: Optional[str] = None,
         cam_creds: Optional[Dict[str, Dict[str, str]]] = None,
-        nb_consecutive_frames: int = 5,
+        nb_consecutive_frames: int = 7,
         frame_size: Optional[Tuple[int, int]] = None,
         cache_backup_period: int = 60,
         frame_saving_period: Optional[int] = None,
diff --git a/src/run.py b/src/run.py
index f022c3bb..f33ac04a 100644
--- a/src/run.py
+++ b/src/run.py
@@ -82,7 +82,7 @@ def main(args):
     )
     # Model
     parser.add_argument("--model_path", type=str, default=None, help="model path")
-    parser.add_argument("--thresh", type=float, default=0.2, help="Confidence threshold")
+    parser.add_argument("--thresh", type=float, default=0.35, help="Confidence threshold")
     parser.add_argument("--max_bbox_size", type=float, default=0.4, help="Maximum bbox size")
 
     # Camera & cache
@@ -106,7 +106,7 @@ def main(args):
     parser.add_argument(
         "--nb-consecutive_frames",
         type=int,
-        default=5,
+        default=6,
         help="Number of consecutive frames to combine for prediction",
     )
     parser.add_argument(
diff --git a/tests/test_engine.py b/tests/test_engine.py
index 7452b759..73129ca3 100644
--- a/tests/test_engine.py
+++ b/tests/test_engine.py
@@ -56,14 +56,25 @@ def test_engine_offline(tmpdir_factory, mock_wildfire_image, mock_forest_image):
     assert isinstance(out, float)
     assert 0 <= out <= 1
     assert len(engine._states["-1"]["last_predictions"]) == 3
-    assert engine._states["-1"]["ongoing"]
+    assert not engine._states["-1"]["ongoing"]
     assert isinstance(engine._states["-1"]["last_predictions"][0][0], Image.Image)
     assert engine._states["-1"]["last_predictions"][2][1].shape[0] > 0
     assert engine._states["-1"]["last_predictions"][2][1].shape[1] == 5
-    assert len(engine._states["-1"]["last_predictions"][-1][2][0]) == 5
     assert engine._states["-1"]["last_predictions"][2][3] < datetime.now().isoformat()
     assert engine._states["-1"]["last_predictions"][2][4] is False
 
+    out = engine.predict(mock_wildfire_image)
+    assert isinstance(out, float)
+    assert 0 <= out <= 1
+    assert len(engine._states["-1"]["last_predictions"]) == 4
+    assert engine._states["-1"]["ongoing"]
+    assert isinstance(engine._states["-1"]["last_predictions"][0][0], Image.Image)
+    assert engine._states["-1"]["last_predictions"][3][1].shape[0] > 0
+    assert engine._states["-1"]["last_predictions"][3][1].shape[1] == 5
+    assert len(engine._states["-1"]["last_predictions"][-1][2][0]) == 5
+    assert engine._states["-1"]["last_predictions"][3][3] < datetime.now().isoformat()
+    assert engine._states["-1"]["last_predictions"][3][4] is False
+
 
 def create_dummy_onnx_model(model_path):
     """Creates a small dummy ONNX model."""