diff --git a/openstef/tasks/train_model.py b/openstef/tasks/train_model.py
index 98e078268..855102e0e 100644
--- a/openstef/tasks/train_model.py
+++ b/openstef/tasks/train_model.py
@@ -53,6 +53,8 @@ def train_model_task(
     datetime_start: Optional[datetime] = None,
     datetime_end: Optional[datetime] = None,
     ignore_existing_models: bool = DEFAULT_IGNORE_EXISTING_MODELS,
+    train_period_days: int = TRAINING_PERIOD_DAYS,
+    maximum_model_age: int = MAXIMUM_MODEL_AGE,
 ) -> None:
     """Train model task.
 
@@ -68,6 +70,9 @@ def train_model_task(
         check_old_model_age: check if model is too young to be retrained
         datetime_start: Start
         datetime_end: End
+        ignore_existing_models: Ignore existing models when training
+        train_period_days: Number of days to fetch for training
+        maximum_model_age: Maximum model age in days to skip retraining
 
     Raises:
         SkipSaveTrainingForecasts: If old model is better or too young, you don't need to save the traing forcast.
@@ -113,9 +118,9 @@ def train_model_task(
     )
 
     # Check old model age and continue yes/no
-    if (old_model_age < MAXIMUM_MODEL_AGE) and check_old_model_age:
+    if (old_model_age < maximum_model_age) and check_old_model_age:
         context.perf_meter.checkpoint(
-            f"Old model is younger than {MAXIMUM_MODEL_AGE} days, skip training"
+            f"Old model is younger than {maximum_model_age} days, skip training"
         )
         if pj.save_train_forecasts:
             raise SkipSaveTrainingForecasts
@@ -123,9 +128,9 @@ def train_model_task(
 
     # Define start and end of the training input data
     training_period_days_to_fetch = (
-        TRAINING_PERIOD_DAYS
+        train_period_days
         if pj.data_balancing_ratio is None
-        else int(pj.data_balancing_ratio * TRAINING_PERIOD_DAYS)
+        else int(pj.data_balancing_ratio * train_period_days)
     )
 
     if datetime_end is None: