perf(ReforceXY): tune default hold potential ratio to 0.03125

Signed-off-by: Jérôme Benoit <jerome.benoit@piment-noir.org>

diff --git a/ReforceXY/reward_space_analysis/README.md b/ReforceXY/reward_space_analysis/README.md
index 65450f4b43fb8527f136b180393ff28957721ab5..df8b2cfd8c90315de9c4c24b3ddac37de97db817 100644 (file)
--- a/ReforceXY/reward_space_analysis/README.md
+++ b/ReforceXY/reward_space_analysis/README.md
@@ -334,7 +334,7 @@ across samples) and does not apply any drift correction in post-processing.
 
 | Parameter                           | Default  | Description          |
 | ----------------------------------- | -------- | -------------------- |
-| `hold_potential_ratio`              | 0.015625 | Hold potential ratio |
+| `hold_potential_ratio`              | 0.03125  | Hold potential ratio |
 | `hold_potential_gain`               | 1.0      | Gain multiplier      |
 | `hold_potential_transform_pnl`      | tanh     | PnL transform        |
 | `hold_potential_transform_duration` | tanh     | Duration transform   |

diff --git a/ReforceXY/reward_space_analysis/reward_space_analysis.py b/ReforceXY/reward_space_analysis/reward_space_analysis.py
index a8ac936e0be2ff41e81e80281d079f40775362db..e6daa92b89e969baa4959e2474950eb822e5fd8d 100644 (file)
--- a/ReforceXY/reward_space_analysis/reward_space_analysis.py
+++ b/ReforceXY/reward_space_analysis/reward_space_analysis.py
@@ -164,7 +164,7 @@ DEFAULT_MODEL_REWARD_PARAMETERS: RewardParams = {
     "exit_potential_decay": 0.5,
     # Hold potential (PBRS function Φ)
     "hold_potential_enabled": True,
-    "hold_potential_ratio": 0.015625,
+    "hold_potential_ratio": 0.03125,
     "hold_potential_gain": 1.0,
     "hold_potential_transform_pnl": "tanh",
     "hold_potential_transform_duration": "tanh",

diff --git a/ReforceXY/reward_space_analysis/tests/constants.py b/ReforceXY/reward_space_analysis/tests/constants.py
index ba564e754d419b79c6f23f840f775e25fa76c7ef..887aafee5e1986374f3e9f22aef13dc7734282ae 100644 (file)
--- a/ReforceXY/reward_space_analysis/tests/constants.py
+++ b/ReforceXY/reward_space_analysis/tests/constants.py
@@ -285,7 +285,7 @@ class TestParameters:
         ADDITIVE_GAIN_DEFAULT: Default additive gain (1.0)
 
         # PBRS hold potential parameters
-        HOLD_POTENTIAL_RATIO_DEFAULT: Default hold potential ratio (0.015625)
+        HOLD_POTENTIAL_RATIO_DEFAULT: Default hold potential ratio (0.03125)
     """
 
     BASE_FACTOR: float = 90.0
@@ -315,7 +315,7 @@ class TestParameters:
     ADDITIVE_GAIN_DEFAULT: float = 1.0
 
     # PBRS hold potential parameters
-    HOLD_POTENTIAL_RATIO_DEFAULT: float = 0.015625
+    HOLD_POTENTIAL_RATIO_DEFAULT: float = 0.03125
 
 
 @dataclass(frozen=True)

diff --git a/ReforceXY/user_data/freqaimodels/ReforceXY.py b/ReforceXY/user_data/freqaimodels/ReforceXY.py
index be8e942b77d2b150ef82297e19cf469d342fe247..4bdba9ebd19941c070b41381b29254abaa643618 100644 (file)
--- a/ReforceXY/user_data/freqaimodels/ReforceXY.py
+++ b/ReforceXY/user_data/freqaimodels/ReforceXY.py
@@ -161,7 +161,7 @@ class ReforceXY(BaseReinforcementLearningModel):
     DEFAULT_ENTRY_ADDITIVE_RATIO: Final[float] = 0.125
     DEFAULT_ENTRY_ADDITIVE_GAIN: Final[float] = 1.0
     DEFAULT_HOLD_POTENTIAL_ENABLED: Final[bool] = True
-    DEFAULT_HOLD_POTENTIAL_RATIO: Final[float] = 0.015625
+    DEFAULT_HOLD_POTENTIAL_RATIO: Final[float] = 0.03125
     DEFAULT_HOLD_POTENTIAL_GAIN: Final[float] = 1.0
     DEFAULT_EXIT_ADDITIVE_ENABLED: Final[bool] = False
     DEFAULT_EXIT_ADDITIVE_RATIO: Final[float] = 0.125

diff --git a/ReforceXY/user_data/strategies/RLAgentStrategy.py b/ReforceXY/user_data/strategies/RLAgentStrategy.py
index 29487ff0e8856933fd0402780aaeef1e4f007c2d..dadcce6b4df6d42eac0cec0c4e868e679ef2a71b 100644 (file)
--- a/ReforceXY/user_data/strategies/RLAgentStrategy.py
+++ b/ReforceXY/user_data/strategies/RLAgentStrategy.py
@@ -1,6 +1,6 @@
 import datetime
 import logging
-from functools import cached_property, reduce
+from functools import reduce
 from typing import Any, Final, Literal, Optional
 
 # import talib.abstract as ta