From: Jérôme Benoit <jerome.benoit@piment-noir.org>
Date: Tue, 18 Feb 2025 22:17:50 +0000 (+0100)
Subject: fix(reforcexy): ensure the same reward is applied with force_actions
X-Git-Url: https://git.piment-noir.org/?a=commitdiff_plain;h=4fa9a7a7eb919a388e87cdec45745ace87bde7b4;p=freqai-strategies.git

fix(reforcexy): ensure the same reward is applied with force_actions

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

diff --git a/ReforceXY/user_data/freqaimodels/ReforceXY.py b/ReforceXY/user_data/freqaimodels/ReforceXY.py
index 8a7a1d8..0da09cb 100644
--- a/ReforceXY/user_data/freqaimodels/ReforceXY.py
+++ b/ReforceXY/user_data/freqaimodels/ReforceXY.py
@@ -674,6 +674,10 @@ class ReforceXY(BaseReinforcementLearningModel):
                 ForceActions.Stop_loss,
                 ForceActions.Timeout,
             ):
+                if pnl > self.profit_aim * self.rr:
+                    factor *= self.rl_config.get("model_reward_parameters", {}).get(
+                        "win_reward_factor", 2
+                    )
                 return pnl * factor
 
             # first, penalize if the action is not valid
@@ -736,7 +740,7 @@ class ReforceXY(BaseReinforcementLearningModel):
                     factor *= self.rl_config.get("model_reward_parameters", {}).get(
                         "win_reward_factor", 2
                     )
-                return float(pnl * factor)
+                return pnl * factor
 
             # close short
             if action == Actions.Short_exit.value and self._position == Positions.Short:
@@ -744,7 +748,7 @@ class ReforceXY(BaseReinforcementLearningModel):
                     factor *= self.rl_config.get("model_reward_parameters", {}).get(
                         "win_reward_factor", 2
                     )
-                return float(pnl * factor)
+                return pnl * factor
 
             return 0.0
 
@@ -850,7 +854,7 @@ class ReforceXY(BaseReinforcementLearningModel):
                 "tick": self._current_tick,
                 "position": self._position.value,
                 "action": action,
-                "force_action": self._get_force_action(),
+                "force_action": self._force_action,
                 "pnl": self.get_unrealized_profit(),
                 "reward": round(reward, 5),
                 "total_reward": round(self.total_reward, 5),