fix: the scaling of the weighted accuracy

packages/valory/skills/decision_maker_abci/policy.py

@@ -24,9 +24,14 @@
 from dataclasses import asdict, dataclass, field, is_dataclass
 from typing import Any, Dict, List, Optional, Union
 
+from packages.valory.skills.decision_maker_abci.utils.scaling import scale_value
+
 
 RandomnessType = Union[int, float, str, bytes, bytearray, None]
-WEIGHTED_ACCURACY_MAX = 101
+
+VOLUME_FACTOR_REGULARIZATION = 0.5
+UNSCALED_WEIGHTED_ACCURACY_INTERVAL = (0.5, 100.5)
+SCALED_WEIGHTED_ACCURACY_INTERVAL = (0, 1)
 
 
 class DataclassEncoder(json.JSONEncoder):
@@ -135,12 +140,14 @@ def best_tool(self) -> str:
     def update_weighted_accuracy(self) -> None:
         """Update the weighted accuracy for each tool."""
         self.weighted_accuracy = {
-            tool: (
+            tool: scale_value(
                 (
                     acc_info.accuracy
-                    + (acc_info.requests - acc_info.pending) / self.n_requests
-                )
-                / WEIGHTED_ACCURACY_MAX
+                    + ((acc_info.requests - acc_info.pending) / self.n_requests)
+                    * VOLUME_FACTOR_REGULARIZATION
+                ),
+                UNSCALED_WEIGHTED_ACCURACY_INTERVAL,
+                SCALED_WEIGHTED_ACCURACY_INTERVAL,
             )
             for tool, acc_info in self.accuracy_store.items()
         }

packages/valory/skills/decision_maker_abci/utils/scaling.py

@@ -0,0 +1,65 @@
+# -*- coding: utf-8 -*-
+# ------------------------------------------------------------------------------
+#
+#   Copyright 2024 Valory AG
+#
+#   Licensed under the Apache License, Version 2.0 (the "License");
+#   you may not use this file except in compliance with the License.
+#   You may obtain a copy of the License at
+#
+#       http://www.apache.org/licenses/LICENSE-2.0
+#
+#   Unless required by applicable law or agreed to in writing, software
+#   distributed under the License is distributed on an "AS IS" BASIS,
+#   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#   See the License for the specific language governing permissions and
+#   limitations under the License.
+#
+# ------------------------------------------------------------------------------
+
+
+"""This package contains helpers for scaling operations."""
+
+
+from typing import List, Tuple
+
+
+def min_max(li: List[float]) -> Tuple[float, float]:
+    """Get the min and max of a list."""
+    if not li:
+        raise ValueError("The list is empty.")
+
+    min_ = max_ = li[0]
+
+    for num in li[1:]:
+        if num < min_:
+            min_ = num
+        elif num > max_:
+            max_ = num
+
+    return min_, max_
+
+
+def scale_value(
+    value: float,
+    min_max_bounds: tuple[float, float],
+    scale_bounds: tuple[float, float] = (0, 1),
+) -> float:
+    """Perform min-max scaling on a value."""
+    min_, max_ = min_max_bounds
+    current_range = max_ - min_
+    # normalize between 0-1
+    std = (value - min_) / current_range
+    # scale between min_bound and max_bound
+    min_bound, max_bound = scale_bounds
+    target_range = max_bound - min_bound
+    return std * target_range + min_bound
+
+
+def min_max_scale(
+    li: List[float],
+    scale_bounds: tuple[float, float] = (0, 1),
+) -> List[float]:
+    """Perform min-max scaling on a list of values."""
+    min_max_ = min_max(li)
+    return [scale_value(value, min_max_, scale_bounds) for value in li]