diff --git a/vizier/_src/benchmarks/experimenters/shifting_experimenter.py b/vizier/_src/benchmarks/experimenters/shifting_experimenter.py
index ab9e401a6..095297aec 100644
--- a/vizier/_src/benchmarks/experimenters/shifting_experimenter.py
+++ b/vizier/_src/benchmarks/experimenters/shifting_experimenter.py
@@ -54,8 +54,10 @@ def __init__(
     exptr_problem_statement = exptr.problem_statement()
 
     if exptr_problem_statement.search_space.is_conditional:
-      raise ValueError('Search space should not have conditional'
-                       f' parameters  {exptr_problem_statement}')
+      raise ValueError(
+          'Search space should not have conditional'
+          f' parameters  {exptr_problem_statement}'
+      )
     dimension = len(exptr_problem_statement.search_space.parameters)
     if dimension <= 0:
       raise ValueError(f'Invalid dimension: {dimension}')
@@ -64,8 +66,8 @@ def __init__(
       self._shift = np.broadcast_to(shift, (dimension,))
     except ValueError as broadcast_err:
       raise ValueError(
-          f'Shift {shift} is not broadcastable for dim: {dimension}.'
-          '\n') from broadcast_err
+          f'Shift {shift} is not broadcastable for dim: {dimension}.\n'
+      ) from broadcast_err
 
     # Converter should be in the underlying extpr space.
     self._converter = converters.TrialToArrayConverter.from_study_config(
@@ -83,26 +85,28 @@ def __init__(
       ):
         if parameter.type != pyvizier.ParameterType.DOUBLE:
           raise ValueError(f'Non-double parameters {parameter}')
-        if (bounds := parameter.bounds) is not None:
-          if abs(shift) >= bounds[1] - bounds[0]:
-            raise ValueError(
-                f'Bounds {bounds} may need to be extended'
-                f'as shift {shift} is too large '
-            )
-          # Shift the bounds to maintain valid bounds.
-          if shift >= 0:
-            new_bounds = (bounds[0] + shift, bounds[1])
-          else:
-            new_bounds = (bounds[0], bounds[1] + shift)
-          self._problem_statement.search_space.add(
-              pyvizier.ParameterConfig.factory(
-                  name=parameter.name,
-                  bounds=new_bounds,
-                  scale_type=parameter.scale_type,
-                  default_value=parameter.default_value,
-                  external_type=parameter.external_type,
-              )
+        if (bounds := parameter.bounds) is None:
+          raise ValueError(f'Parameter {parameter} has no bounds')
+
+        if abs(shift) >= bounds[1] - bounds[0]:
+          raise ValueError(
+              f'Bounds {bounds} may need to be extended'
+              f'as shift {shift} is too large '
           )
+        # Shift the bounds to maintain valid bounds.
+        if shift >= 0:
+          new_bounds = (bounds[0] + shift, bounds[1])
+        else:
+          new_bounds = (bounds[0], bounds[1] + shift)
+        self._problem_statement.search_space.add(
+            pyvizier.ParameterConfig.factory(
+                name=parameter.name,
+                bounds=new_bounds,
+                scale_type=parameter.scale_type,
+                default_value=parameter.default_value,
+                external_type=parameter.external_type,
+            ),
+        )
 
   def problem_statement(self) -> pyvizier.ProblemStatement:
     return copy.deepcopy(self._problem_statement)
@@ -116,8 +120,9 @@ def evaluate(self, suggestions: Sequence[pyvizier.Trial]) -> None:
     for parameters, suggestion in zip(previous_parameters, suggestions):
       suggestion.parameters = parameters
 
-  def _offset(self, suggestions: Sequence[pyvizier.Trial],
-              shift: np.ndarray) -> None:
+  def _offset(
+      self, suggestions: Sequence[pyvizier.Trial], shift: np.ndarray
+  ) -> None:
     """Offsets parameter values (OOB values are clipped)."""
     for suggestion in suggestions:
       features = self._converter.to_features([suggestion])