Update src/streamlit_app.py

src/streamlit_app.py

@@ -258,52 +258,45 @@ if st.button("Run Honest Backtest"):
             
             # --- STRATEGY LOGIC (Same as before) ---
             
-            test_df['Signal'] = np.where(test_df['EMA_Short'] > test_df['EMA_Long'], 1, 0)
+            test_df['Signal'] = np.where(test_df['EMA_Short'] > test_df['EMA_Long'], 1, -1)
+            
             avg_train_vol = train_df['Volatility'].mean()
             test_df['Risk_Ratio'] = test_df['Predicted_Vol'] / avg_train_vol
             
-            test_df['Position_Size'] = 1.0
+            # Initialize Position Column
+            test_df['Target_Position'] = 0.0
+
+            # --- CONDITIONS ---
+            # BULLISH Conditions
+            cond_bull_trend = (test_df['Signal'] == 1)
+            cond_safe_regime = (test_df['Regime'] == 0)   # HMM says "Calm"
+            cond_low_risk = (test_df['Risk_Ratio'] < risk_threshold) # SVR says "Low Vol"
+
+            # BEARISH Conditions (The "Destined to go down" setup)
+            cond_bear_trend = (test_df['Signal'] == -1)
+            cond_crash_regime = (test_df['Regime'] == (n_states - 1)) # HMM says "Crash"
+            # We use Risk Ratio > 1.0 to confirm high volatility momentum (Panic Selling)
+            cond_high_vol_momentum = (test_df['Risk_Ratio'] > 1.0) 
+
+            # --- LOGIC EXECUTION ---
             
-            # Logic
-            cond_safe = (test_df['Regime'] == 0)
-            cond_low_risk = (test_df['Risk_Ratio'] < risk_threshold)
-            cond_crash = (test_df['Regime'] == (n_states - 1))
+            # 1. LONG LOGIC (Sniper Longs)
+            # If Bull Trend + Safe + Low Risk -> Boost Leverage
+            test_df.loc[cond_bull_trend & cond_safe_regime & cond_low_risk, 'Target_Position'] = leverage_mult
+            # If Bull Trend but Risk is normal -> 1x Long (Base case)
+            test_df.loc[cond_bull_trend & (~cond_safe_regime | ~cond_low_risk), 'Target_Position'] = 1.0
             
-            # Boost
-            test_df['Position_Size'] = np.where(cond_safe & cond_low_risk, leverage_mult, test_df['Position_Size'])
-            # Cut
-            test_df['Position_Size'] = np.where(cond_crash, 0.0, test_df['Position_Size'])
+            # 2. SHORT LOGIC (Sniper Shorts)
+            # ONLY Short if: Trend Down + HMM Crash Detected + Volatility is Expanding
+            test_df.loc[cond_bear_trend & cond_crash_regime & cond_high_vol_momentum, 'Target_Position'] = -1.0
             
-            # Calculate Returns
-            test_df['Final_Position'] = (test_df['Signal'] * test_df['Position_Size']).shift(1)
+            # Note: If Bear Trend exists but HMM says "Calm" (State 0) or Vol is low, 
+            # we stay in CASH (0.0) because it might just be a sideways chop, not a crash.
+
+            # 3. Calculate Returns (Shift Logic)
+            # We apply the position chosen 'Today' to 'Tomorrow's' returns
+            test_df['Final_Position'] = test_df['Target_Position'].shift(1)
             test_df['Simple_Returns'] = test_df['Close'].pct_change()
-            test_df['Strategy_Returns'] = test_df['Final_Position'] * test_df['Simple_Returns']
-            
-            # Metrics & Plots
-            test_df['Strategy_Value'] = (1 + test_df['Strategy_Returns'].fillna(0)).cumprod()
-            test_df['Buy_Hold_Value'] = (1 + test_df['Simple_Returns'].fillna(0)).cumprod()
-            test_df.dropna(inplace=True)
-            
-            metrics_df = calculate_metrics(test_df)
-            st.subheader("Performance vs Benchmark")
-            st.table(metrics_df)
-            
-            st.subheader("Equity Curve")
-            fig = go.Figure()
-            fig.add_trace(go.Scatter(x=test_df.index, y=test_df['Buy_Hold_Value'], name='Buy & Hold', line=dict(color='gray', dash='dot')))
-            fig.add_trace(go.Scatter(x=test_df.index, y=test_df['Strategy_Value'], name='Smart Leverage', line=dict(color='#00CC96', width=2)))
-            st.plotly_chart(fig, use_container_width=True)
-            
-            st.subheader("Leverage Deployment")
-            fig_lev = go.Figure()
-            fig_lev.add_trace(go.Scatter(x=test_df.index, y=test_df['Position_Size'], mode='lines', fill='tozeroy', name='Lev', line=dict(color='#636EFA')))
-            st.plotly_chart(fig_lev, use_container_width=True)
             
-            trade_log = generate_trade_log(test_df)
-            st.subheader("📝 Trade Log")
-            if not trade_log.empty:
-                display_log = trade_log.copy()
-                display_log['Trade PnL'] = display_log['Trade PnL'].map('{:.2%}'.format)
-                st.dataframe(display_log, use_container_width=True)
-            else:
-                st.write("No trades generated.")
+            # Math: If Final_Position is -1.0 and Simple_Returns is -5%, Strategy makes +5%
+            test_df['Strategy_Returns'] = test_df['Final_Position'] * test_df['Simple_Returns']